THESIS

“‘5 LIBRARY 293

Michigan State
University

 
  

 

This is to certify that the
dissertation entitled

THE BIOBEHAVIORAL DYNAMICS
OF THE INFANT FEEDING PROCESS

4 a. ,7
presented by

Sara Ann Quandt

has been accepted towards fulfillment
of the requirements for o ,

Ph.D. degree in Anthropology

a jor professor

mafiafl

Date _N03Le.mhm:_.l_1,_l_9_8l

MSU is an Affirmative Action/Equal Opportunity Imtitution 0-12771

 

W- a_ | Li ”(LN ll Nil Hill in (ill 11mg! iii Iii IL H! ”ii i L

 

r .
“is OVERDUE FINES .

25¢ per 60 per 1m

RETURNING LIBRARY MATERIALS:

Place in book return to runove
charge from circulation records

I 35"=\\\\ k
\ . ‘4
\

 

 

 

 

 

THE BIOBEHAVIORAL DYNAMICS OF

THE INFANT FEEDING PROCESS

BY

Sara Ann Quandt

A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Anthropology

1981

 

ABSTRACT

THE BIOBEHAVIORAL DYNAMICS OF
THE INFANT FEEDING PROCESS
BY

Sara Ann Quandt

This research examines longitudinally the interaction
of infant growth, maternal postpartum nutritional status,
and infant feeding strategies involving breast-feeding.
Subjects were 80 normal, first-born neonates and their
mothers, drawn from a low-risk, Caucasian obstetrical
population in Michigan. Infant weight, length, and fat-
folds were measured monthly from birth to six months.
Maternal weight, upper arm circumference, and triceps fat-
fold were measured monthly during the postpartum. Infant
dietary data were reported by mothers in diary form every
eight days throughout their six months of participation.
Additional social data were obtained through interviews
and observations at monthly home visits.

The number of nursing bouts per day during the early
weeks of infancy appears to be a key variable in predicting
long term infant diets and to be part of a complex inter-
action with infant growth. Those infants nursing seven or
more times per day were far more likely to be exclusively
breast-fed for at least three months and to be weaned

after six months than were infants initially fed less

frequently. Infants larger at birth and one month
followed this pattern of frequent nursing more than did
smaller infants. These findings are used to demonstrate
that the seemingly contradictory results of previous
growth and diet research can be explained by focusing on
how infants become members of particular feeding groups.
Maternal postpartum weight loss was not associated
with breast-feeding strategies. Two patterns of change
in maternal midarm fat area were found among women ex-
clusively breast-feeding at least three months and not
weaned until after six months postpartum. Frequent
nursing was strongly correlated with progressive reduction
in fatness, while infrequent nursing was associated with
increases in fatness. It is hypothesized that the differ-
ences observed in maternal fat change patterns result
from variations in lipoprotein lipase and hormone activity,
induced by variable levels of serum prolactin resulting

from different breast-feeding styles.

ACKNOWLEDGMENTS

I gratefully acknowledge the financial support I have
received for the research reported here. The Medical
Anthropology Training Program at Michigan State University
provided predoctoral training funds from National Institute
of Mental Health grant SSS T32 RM 243 MH15132. Substantial
additional funding was contributed by Gerber Products
Company, Fremont, Michigan. Both the College of Social
Science and the College of Human Medicine, Michigan State
University, provided computer funding.

My thanks go to the members of my doctoral guidance
committee: Drs. Cheryl Ritenbaugh, Norman Sauer, Kathryn
Kolasa and Ann Millard. I appreciate the help they have
given me throughout my graduate training and their mainte-
nance of high standards for my work in both anthropology
and nutrition. I am especially grateful for their prompt
and critical reading of the earlier drafts of this disser-
tation.

Special thanks go to Dr. Cheryl Ritenbaugh who urged
me to undertake this research. Her interest in the study
and faith in the significance of its results helped to
maintain my enthusiasm during three rather long years of

data collection.

ii

iii

Perhaps my greatest debt is owed to the mothers and
infants who participated in this research. Their willing-
ness to share their personal experiences with me and to
cooperate in all phases of the study is sincerely appreci-
ated.

I thank my friends in the Department of Anthropology
at Michigan State for tolerating endless discussions of
breast-feeding and baby fat and for providing the diversions
necessary for the successful completion of graduate school.

And finally, I thank my parents who encouraged me to
return to school. During the past six years they have
supported me in many ways--and almost never lost hope that

I would someday graduate!

TABLE OF CONTENTS

Chapter
I Introduction . . . . . . . . . .
II Review of the Literature . . . .

Infant Diet

U.S. Breast-Feeding and Artificial

Practices since 1900 . . . .
Current Research Directions
Infant Nutrition . . . . . . .

Milk Consumption and Growth

Growth and Consumption of Non-Milk

Underlying Assumptions . . .

Maternal Nutritional Status .

Breast-Feeding and Weight Loss

Breast-Feeding and Fatness .
Summary . . . . . . . . . . .
III Methods . . . . . . . . . . . .
Research Design . . . . . . .

The Sample
Description of the Sample .
Sample Recruitment . . . . .
Data Collection . . . . . . .

Maternal Data Collection . .

Infant Data Collection - Anthropometric

13
17
18
21
23
29
30
32
37
39
39

41
44
46
46

49

Infant Data Collection - Anthropometric .

Statistical Considerations and Reporting

Reliability

Levels of Statistical Significance

Indicators of Accuracy: Infant Diet Diaries

Combining Subjects from Study I and Study II

IV Results . . . . . . . . . . . . . .

Infant Diet

General Description of the Infant Diets .

Association of Infant Diet with
Education and Household Income

Length and Frequency of Nursing
Transitions to a Mixed Diet . .
Major Dietary Patterns . . . .
Infant Growth . . . . . . . . . .
Description of the Sample . . .
Preliminary Data Transformations

weaning 0 O O o o o o o o o o o

Maternal

Bouts

Introduction of First Non-Breast Milk

Common Infant Dietary Patterns
Details of Breast Feeding . . .

Interview Data - - - . . . . . .

Prenatal Plans for Infant Feeding .

Infant Feeding in the Parental Generation

Spatial Arrangements of Families
Decision Making Data . . . . .

Maternal Nutritional Status . . .

Descriptive Statistics for the Total Sample

49

54
55
56

S9

59

64
73
84
91
97
98
101
102
107
108
124
127
130
130
131
134
138

140

vi

Maternal Measures by Common Infant Dietary
Patterns . . . . . . . . . . . . . . . . . . . 146

Maternal Measures by Months of Breast- .
Feeding O O O O O O O C I O O O O 0 O O O O O O 149

Maternal Measures and Details of Breast-
Feeding . . . . . . . . . . . . . . . . . . . . 152

V Discussion
Emergence of Infant Feeding Strategies . . . . . 164
Interaction of Infant Diet and Infant Growth . . 173

Influence of Growth on the Dietary Pattern
of Infants . . . . . . . . . . . . . . . . . . 174

Influence of Dietary Patterns on the Growth
of Infants . . . . . . . . . . . . . . . . . . 180

Supporting Evidence from the Infant Growth
and Diet Literature . . . . . . . . . . . . . . 186

Conflicting Evidence for the Impact of
Diet on Growth . . . . . . . . . . . . . . . . 188

Interaction of Infant Diet and Maternal
Nutritional Status . . . . . . . . . . . . . . . 190

Breast-Feeding and Weight Loss . . . . . . . . 190

Breast-Feeding and Changes in Maternal
Body Fat . . . . . . . . . . . . . . . . . . . 194

VI Synthesis and Conclusions - . . . . . . . . . . . . 199

A General Model of Infant Feeding Systems 199
‘/Lactation Amenorrhea and Maternal Body Fat . . . 205
The Adequacy of Breast Milk . . . . . . . . . . . 209
Conclusions . . . . . . . . . . . . . . . . . . . 213

References Cited . . . . . . . . . . . . . . . . . 216

Appendix A: Letters of Introduction and Consent
Forms 0 o o o o c o o o o o o o . . . 227

Appendix B: Distribution of Mothers' Returning

Infant Diet Diaries . . . . . . 233

Appendix C:

Appendix D:

vii
Infant Outlines Used in Data

Collection . . . . . . . . . . . .

Reasons for Introducing New Food to
a Breast Milk Diet . . . . . . . .

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

III.1

III.2

III.3

III.4

IV.l

IV.2

IV.3

IV.4

IV.5

IV.6

IV.7

IV.8

IV.9

LIST OF TABLES

Number of Infants Measured at Birth and
at each Month . . . . . . . . . . . . .

Gross Household Income by Study and for
Total Sample . . . . . . . . . . . . .

Highest Level of Maternal Formal
Education by Study and for Total Sample

Comparison of Biological Variables,
Analysis of Variance: Study I versus
study II C C O C O O O O O C O I C O 0

Number and Percentage of Infants
Consuming Major Non-Breast Milk Food
Types during each Month of the Study

First Solid Consumed; Number of Infants
by Month of Introduction . . . . . . .

Percentage and Number of Infants
Consuming Combined Dietary Components
during each Month of the Study . . . .

Highest Level of Maternal Formal
Education and Month of Introduction
of First Non-Breast Milk Food . . . . .

Highest Level of Maternal Formal
Education and Month of Weaning . . .

Month of Introduction of First Non-
Breast Milk Food by Household Income
Category . . . . . . . . . . . . . . .

Month of Weaning by Household Income
Category . . .

Percentage and Number of Infants
Consuming Formula by Sex and Month . .

Percentage and Number of Infants
Consuming Solid Food by Sex and Month .

viii

42

43

43

58

62

62

63

65

66

67

68

69

71

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

IV.10

IV.11

IV.12

IV.13

IV.14

IV.15

IV.16

IV.17

IV.18

IV.19

IV.20

IV.21

IV.22

Number and Percentage of Infants
Consuming First Solid Food by Month,
Sex, and Study

Number of Breast-Feedings per Day in
Month 1 by Length of Exclusive Breast-
Feeding .

Number of Breast-Feedings per Day in
Month 2 by Length of Exclusive Breast-

Feeding .

Change in Number of Breast-Feedings per
Day between Month 1 and Month 2, by
Months of Exclusive Breast-Feeding

Change in Minutes per Breast-Feeding
between Month 1 and Month 2, by Months

ix

of Exclusive Breast-Feeding .

Role of Newly Introduced Foods in Diet
Formerly Consisting Exclusively of
Breast Milk

Test for the Presence of a Population-
Wide Decline in Number of Breast-
Feeding Bouts per Day during Months

2 and

Introduction of Formula and Solids
Relative to Median Date for Introduction
of First Non-Breast Milk Food

Role of New Food in Infant Diet, by Sex
and Month of Introduction

Month of Weaning for Infants Given
Formula or Solids in Month 1, by

3

Addition and Replacement Categories .

Nutrient Values for Average Daily Intake
of Non-Breast Milk Foods at their First

Introduction; Mean : Standard Deviation .

Nutrient Values for Average Daily Intake
of Non-Breast Milk Foods at their First
Introduction for Feeding Group 3; Mean

‘1 Standard Deviation

Descriptive Statistics for Infants at
Birth and at End of each Month

71

78

79

80

81

86

88

89

90

91

96

97

99

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

IV.23

IV.24

IV.25

IV.26

IV.27

IV.28

IV.29

IV.30

IV.31

IV.32

IV.33

x
Mean NCHS Percentile : Standard

Deviation for Infants at Birth and at

End of each Month . . . . . . . . . . . .

Means : Standard Deviations of Four
Infant Growth Measures at Birth and End
of Months 1 to 6, by Month of Weaning . .

Means : Standard Deviations of Four
Infant Growth Measures at Birth and End
of Months 1 to 6, by Month of First
Introduction of Non-Breast Milk Foods . .

Means : Standard Deviations of Four
Infant Growth Measures at Birth and

End of Months 1 to 6, by Infant Feeding
Groups . . . . . . . . . . . . . . . . . .

Regression Equations: Dependent Variable
is Weight for Length and Independent
Variable is Month of Age (1 - 6 months). .

Comparisons of Number of Infants in
Lowest and Highest Quartiles of Weight
for Age, at Birth and End of each Month,
by Feeding Group . . . . . . . . . . . . .

Comparisons of Number of Infants in
Lowest and Highest Quartiles of Weight
for Length, at Birth and End of each
Month, by Feeding Group . . . . . . . . .

Fisher's Exact Test of Difference in
Number of Infants in Highest and Lowest
Weight for Length Quartiles between
Groups "Replacing" Breast Milk with
Formula (lb and 2a) and those NOT
"Replacing" with Formula (la and 2b) . . .

Relationship between Growth and Details
of Breast-Feeding among ; Infants of
Mothers in Groups A and B . . . . . . . .

Study II Responses to the Question:

When will you first give your baby
anything but breast milk (or an
occasional "convenience bottle") to eat? .

Parental Generation Feeding as Infants
by Feeding Group of Infants in Present
Study (Study II only). . . . . . . . . . .

100

103

109

114

122

123

'125

126

128

131

132

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

Table

IV.34

IV.35

IV.36

IV.37

IV.38

IV.39

IV.4O

IV.41

IV.42

IV.43

IV.44

IV.45

IV.46

xi

Distribution of Infant Feeding Groups
by Spatial Arrangement of Family
Dwelling O O C O O O C O O O I O C O O O 0

Distribution of Spatial Arrangement
of Dwellings by Maternal Formal
Education . . . . . . . . . . . . . . . .

Categories of Maternal Decisions to
Introduce First Non-Breast Milk Food . . .

Addition or Replacement of First Non-
Breast Milk Food by Decision Categories .

Cumulative Measures of Maternal Weight
Loss during the Postpartum . . . . . . . .

Cumulative Changes in Midarm Muscle
and Fat Areas during the Postpartum . . .

Cumulative Maternal Weight Loss from
Last Prepartum Weighing by Postpartum
Month and Infant Feeding Group . . . . . .

Cumulative Change in Midarm Fat Area

from First Postpartum Measurement by
Postpartum Month and Infant Feeding

Group . . . . . . . . . . . . . . . . . .

Cumulative Weight Loss from Last
Prepartum Weighing by Postpartum Month
and Months of Breast-Feeding . . . . . . .

Cumulative Change in Midarm Fat Area

from First Postpartum Measurement by
Postpartum Month and Months of Breast-
Feeding. . . . . . . . . . . . . . . . . .

Number of Mothers with Increasing and
Decreasing Fat Measures in the Month
Following Weaning, by Month of Weaning . .

Numbers of Mothers with Increasing and
Decreasing Fat Measures in the Month
Following First Introduction of Non-

Breast Milk Food, by Month of Intrdduction.

Spearman Rank-Order Correlation
Coefficients: Change in Maternal Weight
and Fat with Length and Number of Breast—
Feedings . . . . . . . . . . . . . . . . .

133

134

135

139

141

144

147

148

150

153

154

154

Table

Table

Table

Table

Table

Table

Table

IV.47

IV.48

IV.49

IV.50

IV.51

IV.52

VI.1

xii

Maternal Anthropometric and Birth
Outcome Variables for Groups A and
B, Mean 1 Standard Deviation . . . . . .

Number of Breast-Feedings per Day at
Four Weeks for Infants of Mothers in
Groups A and B . . . . . . . . . . . . .

Number of Breast-Feedings per Day at
Eight Weeks for Infants of Mothers in
Groups A and B . . . . . . . . . . . . .

Number of Breast-Feedings per Day at
Twelve Weeks for Infants of Mothers in
Groups A and B . . . . . . . . . . . . .

Minutes per Feeding for Groups A and B.
Mean 1 Standard Deviation . . . . . . .

Spearman Rank—Order Correlation
Coefficient: Minutes per Breast-Feeding
with Numbers of Feedings per Day . . . .

Hypothesized Maternal Effects of
Decreased Breast-Feeding Frequency with
Varying Levels of Caloric Intake . . . .

160

161

161

162

163

163

208

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

IV.1

IV.2

IV.3

IV.4

IV.5

IV.6

IV.7

IV.8

IV.9

IV.10

IV.11

IV.12

LIST OF FIGURES

Number and Percent of Infants Breast-

Fed (Consuming any Breast Milk) and
Weaned (Consuming no Breast Milk)

during First through Sixth Months of

Life 0 O O O O O I O O O O O O O O O I O I 60
Sex Differences in Breast-Feeding Status
by Month of Study . . . . . . . . . . . . 70
Percent of Infants First Consuming Solids
for Male and Female Infants by Month of
Age 0 O O C O O C O C C O O O O C O I O O 72
Average Number of Breast—Feedings per

Day for Infants Exclusively Breast-Fed
during Months 1 and 2 . . . . . . . . . . 75
Number of Breast-Feedings per Day in
Month 1 by Length of Exclusive Breast-
Feeding . . . . . . . . . . . . . . . . . 76
Number of Breast-Feedings per Day in
Month 2 by Length of Exclusive Breast-
Feeding . . . . . . . . . . . . . . . . . 77
Mean Number of Minutes per Feeding

during Month 1 . . . . . . . . . . . . . . 82
Mean Number of Minutes per Feeding

during Month 2 . . . . . . . . . . . . . . 83
Length of Exclusive Breast-Feeding and
Infant Feeding Groups . . . . . . . . . . 93
Mean Weight for Age (Percent of NCHS
Median) by Month of Introduction of First
Non-Breast Milk Food . . . . . . . . . . .

Mean Weight for Length (Percent of NCHS
Median) by Month of Introduction of First

Non-Breast Milk Food . . . . . . . . . . . 113

Mean Weight for Age (Percent of NCHS
Median) for Infant Feeding Groups la

through 3 . . . . . . . . . . . . . . . . 118

xiii

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure

Figure
Figure

Figure

Figure

IV.13

IV.14

IV.15

IV.16

IV.17

IV.18

V.1

VI.1

V1.2

V1.3

3.1

C.1

xiv

Mean Weight for Length (Percent of NCHS
Median) for Infant Feeding Groups la
and lb 0 I O O O O O O O O O O I O O D 0

Mean Weight for Length (Percent of NCHS
Median) for Infant Feeding Groups 2a,
2b, and 3 . . . . . . . . . . . . . . .

Maternal Postpartum Weight Status;
Mean i Standard Deviation . . . . . . .

Cumulative Postpartum Change in Maternal
Midarm Muscle and Fat Areas; Mean :
Standard Deviation . . . . . . . . . . .

Cumulative Postpartum Weight Loss by
Month of Weaning . . . . . . . . . . . .

Distribution of Significance Levels for
Regression of Postpartum Maternal Fat
Area on Days since Parturition . . . . .

Hypothesized Pathways for the Inter-
action of Nursing Frequency and Maternal
BOdy Fat 0 C O C C C O O O O O O O O O 0
Model of the Interaction of Infant
Feeding Behavior, Infant Growth, and
Parameters of Maternal Physiology . . .
Components of the "Social Environment“ .

Components of "Parameters of Maternal
Physiology" . . . . . . . . . . . . . .

Distribution of Mothers Returning Diet
Diaries 0 O O O O O O C O O O O O O O 0

Infant Outlines . . . . . . . . . . . .

121

121

142

143

151

158

196

201

202

203

233

235

CHAPTER I

INTRODUCTION

In order to survive, an animal must eat; if it is a
young and helpless mammal, it must be fed. All successful
mammalian species have evolved ways of nourishing their
young and, from an intra-specific perspective, these
behaviors are remarkably uniform. However, one such species
does not conform to this generalization. When one compares
human populations, a high degree of variability emerges in
what is fed to infants, when it is fed, and by what modality.
Even within specific populations, infant feeding behaviors
vary. It is a basic premise of this study that such behaviors
are not random, but are patterned by the interactions of both
cultural and biological characteristics of the group which
the infant belongs. It is a second premise that the detailed
examination of the feeding behaviors for individual children
can help to map these biocultural interactions.

In EurOpe and the United States, infant feeding behaviors
are increasingly cited as determinants of lifelong growth and
health. As a result, the medical professions prescribe what
they consider to be desirable patterns for feeding young
children. Yet there is an enormous amount of variability in

1

2

the diets actually fed to infants--and the observable
results are not easily classified as harmful. Obviously,
factors other than medical advice are affecting infant
diet. While research has been done on the factors in-
fluencing certain isolated choices which mothers make
concerning infant diet, little study has been done of the
manner in which these choices interact to result in
observed, long term infant dietary patterns. Further,
although the effect of specific foods on growth has been
studied, little is known about the way in which feeding .
strategies as a whole affect both infant and--in the case
of breast-feeding--materna1 nutritional status.

This research follows a trend toward the increased
use of interaction and feedback models to describe and
analyze human diets. On a general, theoretical level Katz
(1980) and Jerome et a1. (1980) have published models of
the adaptability of foodways for human populations. Also
at the theoretical level, Bass et a1. (1979) have presented
a feedback model for understanding individual food habits.
All of these models consider the interaction of a variety
of aspects of the physical and social environments as well
as aspects of human biology. Related more directly to
infant feeding are two models which deal with the biosocial
components of lactation. Jelliffe and Jelliffe (1979b)
outline the basic reflexes necessary for successful breast-
feeding in individual women. Gussler and Briesemeister

(1980) have expanded on this to explore specific attributes

3
of social interactions which can lead to lactation failure
both in individual cases and as a population-wide pheno-
menon. However, their model remains untested on a single
body of data.

This dissertation presents a prospective study in
which infant growth, maternal postpartum nutritional status,
and infant feeding strategies involving breast-feeding
were documented in a free-living population from central
Michigan. The dissertation has three primary goals. The
first is to identify the principle infant feeding strategies
which exist in this population. A processual approach is
taken in the analysis of these feeding strategies. This
approach-~in contrast to that of many previous studies--
serves to emphasize that an infant feeding strategy is not
a set of independent behaviors. Rather, it is the product
of a series of decisions, each informed and limited by
previous ones and, in turn, influencing those subsequent
to it. In addition to examining the interaction of dietary
components with each other, the study examines the influence
on the decision points of factors external to the infant
feeding process. These include socioeconomic variables
and food beliefs as they are shaped by parental interaction
with the medical community, family, and friends.

Having identified the major infant feeding strategies
and associated influences, the second goal of the dissertation
will be to examine the interaction of these strategies with

infant growth. Specifically, it will determine whether

4
different patterns of growth as well as differences in
attained size can be linked to feeding strategies.
Weight, length and measures of body fat will be examined.
The final goal is to study the interaction of those
feeding strategies involving breast-feeding with maternal
nutritional status assessed anthropometrically. Varia-
bility in times of weaning and supplementation, as well
as differences in lactation "styles" (e.g., length and
frequency of nursing bouts) will be examined in attempts
to explain postpartum changes in weight and body composition.
The results of this investigation may be valuable in
a number of ways. First, they will provide a comparison
of the explanatory power of processual descriptions of diet
for infant growth with more traditional infant diet cate-
gories. The results should provide mother/infant "profiles"
of the most common growth and diet patterns. Furthermore,
the results will identify the decision points in the infant
feeding process for particular subsets of the mother/infant
population, a finding of potential significance for those
interested in diet and growth intervention. And finally,
this investigation will provide basic data on the impact of
different kinds of breast-feeding practices on maternal
nutritional status. Currently, such data are virtually
nonexistent in both the medical and nutritional literature.
They are potentially relevant to policy decisions regarding
promotion of various infant feeding practices among women

of varying levels of nutritional status, and are of interest

S
in the formulation of statements on the biologically
adaptive value of breast-feeding as a human food habit.

The chapters of the dissertation are arranged in the
following sequence: literature review, methods, results,
discussion, and a final synthesis with conclusions.

Chapter II reviews previous studies of infant diet, infant
growth, and maternal nutritional status. It presents
findings relevant to the present study and assesses the
limitations of current knowledge in the area of infant
feeding.

Chapter III describes and justifies the research
design. It also discusses the sample and the data
collection techniques. The rationale for choosing a
particular infant dietary data collection strategy is pre-
sented and its effectiveness is assessed.

Chapter IV contains the results of the data analysis.
In the first section, the most common infant dietary
patterns are defined in terms of sample demographics, timing
of major feeding events, and details of the breast-feeding
process. A second section summarizes data obtained through
interviews with mothers. These focus on the process by
which infant feeding decisions are made. Analyses of
infant growth are presented in the third section, and the
fourth describes analyses designed to identify patterns of
change in maternal nutritional status.

These somewhat discrete categories of results are each

summarized and discussed in Chapter V. They are then merged

6
to form the basis for a synthetic biobehavioral model of
the infant feeding process in Chapter VI. The potential
application of this model to two current scientific contro-
versies--the critical body composition hypothesis and the

nutritional adequacy of breast milk-~conclude the chapter.

CHAPTER II
REVIEW OF THE LITERATURE

This review of the literature will discuss infant
feeding studies from the last several decades in terms of
three general and not necessarily discrete categories.

The first is infant diet, the behavioral study of what is
fed to infants and the rationale for such behavior. The
second is infant nutrition, the interaction of infant diet
and physiology. Infant growth studies fall into this
category. The third involves the relationship of lactation
to the nutritional status of nursing mothers.

Investigations will be reviewed with respect to their
findings and their assumptions in order to discover why the
current picture of infant feeding is generally fragmentary
and inconsistent. It will be shown that these inconsistencies
can often be understood through considering the nature of the
researcher and of the research design. Finally this chapter
will attempt to demonstrate that by changing or at least
making more explicit the assumptions about infant feeding
inherent in these studies, it may be possible to design new
and more productive research which can present a more complete
and coherent picture of the infant feeding process.

7

8

Infant Diet

 

U.S. Breast-Feeding and Artificial Feeding Practices since
1900

 

An historical overview of infant feeding practices
and recommendations provides two types of information:
what has been fed to infants and the sources of influence
on physicians and others which have caused the observable
trends in feeding patterns.

Although a survey of medical texts throughout the
19003 shows a consistent recommendation of breast milk as
the best source of nutrients for infants, physicians have
apparently been less than consistent in their enthusiasm
for or effectiveness in promoting the practice. Reported
figures vary, but in general, the incidence of breast-
feeding in the United States and Western Europe decreased
from high levels at the beginning of the 20th century to a
very low percentage in the 19503 and 19603 (Fomon 1974).
Meyer collected data on infant feeding modality at hospital
discharge at ten year intervals in a representative group
of U.S. hospitals and found a steady decline from 38% to
21% to 18% in 1946, 1956, and 1966, respectively (1958,
1968). The decrease was not uniform across all social
strata. Breast-feeding remained the most common form of
infant feeding in the lower social classes until the 19603,
while the upper classes adopted the more "modern" practice
of artificial feeding. Since the 19603 there has been a

resurgence of interest and incidence of breast-feeding in

9
the U.S. and Europe. This has been particularly apparent
in the upper social classes, with the lower classes lagging
behind due to what one author has called the "osmotic
trickling down" action of social attitudes and practices
(Meyer 1968:713).

Medical recommendations of appropriate ages for the
feeding of solid foods have shown marked changes through
time. (Most U.S. authors include cereal, juice and all
non-milk foods in the "solids" category.) In the early
19003 solids in general were forbidden until after one year
of age, with vegetables, for example, considered inappro-
priate until three years of age (Fomon 1974). By 1937,
pediatricians were suggesting such foods by four to six
months of age (A.M.A. Council on Foods 1937). The extreme
was reached by the 19503 when Sackett published his "new
concept in baby feeding" (1953), which prescribed solids
on the second day of life. Since that time, surveys of
medical textbooks and practicing physicians show a return
to the 19303 recommendation of four to six months as the
appropriate age for solids (Jelliffe and Jelliffe 1979b).
This is also the most recent recommendation of the American
Academy of Pediatrics (1980).

In reviewing changing patterns of infant feeding, it
is evident that trends in the relative proportions of
infants consuming breast milk or formula and in ages for
introduction of solids have tended to coincide (American

Academy of Pediatrics 1978). This suggests that infant

10
feeding has been influenced by changes in medical wisdom
through time. It also suggests that the entire process
may be responsive to general changes in lifestyles external
to the health professions.

The amount of information available to physicians
relevant to infant feeding increased rapidly at the
beginning of this century. New (although not always
accurate) data on thecomposition of human milk led to
what has been called the "chemical period" (Meyer 1952).
Rotch, for example, advocated."percentage feeding," in
which minute proportions of different nutrients in infant
milks were calculated (1907). Milk prescriptions were in
vogue, and the intricacies of these--some were over a page
in length--were responsible for the term "formula" which
is still in use today. Knowledge about rates of gastric
emptying and curd tension of different kinds of milk
contributed to the scientification of infant feeding.

While the official medical policy on issues such as breast-
feeding versus artificial milks remained unchanged, the
increasing research emphasis on artificial feeding and its
inclusion in medical texts probably helped make physicians
more amenable to the use of such methods when advocated by
the public.

Research on lactation or any aspect of infant feeding
dealing with the mother was conspicuously absent in the first
half of the century. When writers did start to mention

mothers, it was usually in the context of easing their

ll

workload through reducing the number of infant feedings
per day, often by lowering the age at which solid foods
were introduced (cf. Clein 1943, Sackett 1953). However,
this one-sided view of infant feeding was replaced to some
extent in the 19703 by an interest in the interactions
inherent in the total infant feeding process (e.g.,
maternal-infant interactions, food interactions).

Paralleling the changes in infant feeding through the
19003 have been broader trends in life styles and popular
conceptions about infants. The increased use of infant
formulas, first mixed at home and later commercially pre—
pared, was part of a general increased reliance on things
scientific and technologically modern. Time and labor
saving devices were sought in all areas of women's work
in the home, and the use of formula and canned baby foods
fed on a rigid schedule was consistent with this. At the
same time, the conceptualization of the infant changed.
Writers of the early 19003 (cf. Rotch 1907) referred to
infants as totally helpless and vulnerable creatures. By
the 19403 and 19503 this view had changed. Child rearing
practices were seen as the time in which lifelong habits
were taught. The emphasis was on conformity, on orderly
patterns of behavior, and on molding the willful infant
into his role as a member first of a family and later of
society (Aldrich 1942; Clein 1943; Sackett 1953).

However, since the 19603, reaction against this trend

has become widespread, as an emphasis on naturalism has

12
become a pervasive force in certain segments of Western
society. For diet in general, this has included advocating
a reduced consumption of processed and fortified foods and
for infants and mothers in particular, a return to breast-
feeding. Role models for this naturalism in food consump-
tion come from a variety of sources, including pOpular,
romanticized notions about the ways in which technologically
simple societies function. Part of these ideas focus on
the child as a free, self-determining spirit. This has
resulted in a widespread belief that breast-feeding should
be carried out on demand from immediately after birth for
as long as several years. Other foods should be restricted
until at least the second six months of life, a time when
the infant has enough muscular control to regulate its own
intake. Infant diet research has tended to follow these
popular influences. Articles on the long term adequacy of
breast milk are now increasingly common (Ahn and MacLean
1980; Saarinen and Siimes 1979).

Yet despite the current medical recommendations for
extended breast-feeding and delayed non-breast milk foods
and despite the vocal nature of the support this receives
from many mothers, data from a number of sources (Fomon
et a1. 1979; Kemberling 1979; American Academy of Pediatrics
1978) indicate that 40 to 50% of all newborns are formula-
fed rather than breast-fed at hospital discharge and that
most are fed some non-milk solid foods by two months of age.

This contradiction has led to a new dimension of research--

13

the study of why infants are fed in different ways.

Current Research Directions

Recently there have begun to appear reports of research
designed to discover the rationale for different patterns
of infant feeding from the side of those directly re-
sponsible for implementing such action: parents, particu-
1arly mothers. These include studies of the following
decisions: to breast- or bottle-feed from birth, when to
cease breast-feeding, and when to introduce solid foods
into the infant diet.

The first decision--to breast-feed or use formula--
appears to be made well before delivery. Guthrie and Kan
(1977) report that among unmarried and non-pregnant females
up to 21 years of age in Pennsylvania, most had first
thought about their future choice in infant feeding methods
between the ages of 12 and 17. Furthermore, all of the
oldest women could articulate reasons for their choice
beyond simple previous exposure to the infant feeding method.
This indication that the bottle/breast-feeding decision is
made long prior to delivery is supported by Brimblecombe
and Cullen (1977). .They report that a concerted effort by
midwives, public health visitors, and physicians to increase
the rate of mothers choosing to breast-feed in 1975 was
unsuccessful. Rates of breast-feeding were not significantly
different in before and after surveys. The most significant
factors in these mothers' feeding decisions were the methods

by which they had fed previous infants and the way they

l4

themselves had been fed as children.

Postpartum decisions on the length of breast-feeding
have been investigated in New Zealand by Hood et a1. (1978)
and in Sweden by Sjolin et a1. (1977, 1979). In a retro-
spective survey of 574 New Zealand mothers, Hood et a1.
(1978) found that 26% of the weanings (defined by the
authors as the complete cessation of breast-feeding) were
said by the mother to be voluntary. Either she had breast-
fed as long as she wanted to, or the infant had "weaned
itself." The latter usually occurred after six months of
age. Thirty-seven percent reported weaning because of
"concerns with the milk supply," and nine percent because
of "maternal stress." The categories "maternal illness,"
"nipple problems," and "other breast problems" accounted
for 10%, and several other miscellaneous reasons accounted
for the remaining mothers. In another retrospective survey
of 298 Swedish mothers, Sjolin et a1. (1977) found that the
following accounted for over 40% of the decisions to wean:
1) concern about the child's getting sufficient milk,
2) anxiety and insecurity, 3) lack of motivation, and
4) stress. There is no indication whether these are actual
reasons given by the mothers, categories presented to them
from which to choose, or whether the authors grouped the
mothers' responses as they saw fit.

A prospective study of individual breast-feeding ex-
periences among 75 Swedish mothers was designed to explore

in detail the factors leading to weaning (Sjolin et al.1979).

15
These factors were categorized as emotional, rational, and
medical reasons connected with the mother; environmental
reasons; and the interplay of maternal and environmental
reasons. However, the distinctions among these categories
are not clear, nor are sufficient statistical computations
done to show whether or not each of these categories is
specific to any group of children or age of weaning.

There have been very few studies of the decisiOn to
add solids (and other non-milk foods) to the infant diet.
Auerbach (1978), in a study of 102 American mothers, notes
that infants on mixed breast milk and formula diets tend
to be fed solids much earlier than those on a diet con-
sisting exclusively of breast milk. The results of
Wilkinson and Davies' (1978) study of British mothers
support those findings and indicate that maternal perceived
hunger is the most common reason for starting solids,
regardless of composition of the milk portion of the diet.

While the emphasis on maternal decision making and
behavior in infant feeding is recent and thus the studies
cited above do much to open up this area of research,
several important shortcomings must be mentioned. First,
the research generally focuses on a single food or food
category. For example, Brimblecombe and Cullen concentrate
only on the effect of nutrition education on the breast/bottle
decision. Yet, delayed feeding of solids is normally part
of the breast-feeding education message and may in fact have

been effectively transmitted. Similarly, Sjolin et a1. (1979)

16
purport to examine causes of lactation crises and weaning,
without ascertaining the composition of the infant diet
beyond breast milk.

A second and related problem with many of the studies
is that foods are viewed in presence/absence terms. Women
in the Hood et a1. (1978) study are listed as breast-feeding
at up to three years postpartum. Obviously other foods are
being consumed by the child, and having more information
about them would make it possible to interpret reasons
for weaning in the appropriate context, as would such
details as daily frequency of breast-feeding sessions.

A final problem is that while factors such as "maternal
stress," "insufficient milk supply," and "poor infant growth"
constitute a major proportion of the reasons given for wean-
ing or adding solids to the diet, these are rarely explored
in any detail. Even though some studies did collect infant
anthropometric data (Wilkinson and Davies 1978; Brimblecombe
and Cullen 1977), no attempts were made to associate specific
decision making factors with growth parameters. One reason
for the failure to examine these categories in any depth
may be the orientation of those conducting the research.

The majority are either physicians or public health workers
coming out of a medical training background. It may be that
while such training sensitizes them to the subtle distinctions
in some categories (e.g., Hood et a1. [1978] have three
separate categories related to maternal health), possibly

meaningful distinctions in other categories are ignored.

17

Although the studies cited above fail to integrate
the factors cited by mothers as reasons for modifying
the course of breast-feeding, the interaction of these
and their connection via the physiology of lactation have
been discussed by others. Jelliffe and Jelliffe (1978)
have called attention to the importance of alleviating
maternal anxiety in order to insure an adequate oxytocin
induced let-down reflex. This, in turn, promotes suckling,
which stimulates the prolactin secretion necessary for
continued milk production. Gussler and Briesemeister
(1980) have developed this in greater detail, concentrating
on the specific components of the social and physical
environments which, when combined, result in either success-
ful lactation or the "insufficient milk syndrome." This
Gussler and Briesemeister see as a phenomenon specific to
Western society or to those non—Western societies undergoing
a transition from traditional to more modern lifestyles.
The model they propose for this "syndrome" contains many
of the factors cited by mothers in studies such as Hood et
al. (1978) and Sjolin et a1. (1977, 1979), but it remains
to be verified with a single complete body of behavioral

and biological data.

Infant Nutrition

 

Research interest in the interaction of infant diet
and growth has centered on two main questions. First, are
there growth differences between infants breast-fed and those

fed artificial formulas? Second, does the age at which infants

18
start to consume solid foods influence their later growth?
The way in which these questions have been phrased has
dictated the design of the research based on them and has
probably contributed to the confusion and controversy

surrounding diet and growth of young infants.

Milk Consumption and Growth

 

A recent review of this research area is provided by
Himes (1979), who notes the conflicting results obtained by
investigators studying infants in industrialized countries.
Some of the most recent or most frequently cited studies are
discussed below.

The strongest support for the notion that growth
differences exist between breast milk- and formula-fed
infants comes from Fomon et a1. (1970, 1971). They studied
the growth from 8 to 112 days of 149 infants breast-fed and
154 formula-fed from birth. Breast-fed infants were per-
mitted one supplementary formula feeding daily. After 28
days, both groups of infants were permitted strained cereal
and fruit on a prescribed schedule of introduction. However,
feeding these foods was left up to parental discretion.
Although there were no significant differences between groups
as a whole in terms of mean weight and length parameters,
-differences did appear between groups by 112 days in the
distribution of body weight. While the distributions had
been the same for breast- and bottle-fed infants at 8 days,
at 112 days the 90th percentile for bottleefed infants was

"substantially" higher. The 10th percentiles remained equal

19

for the two groups. These results have been supported by
results obtained in England by Taitz (1971). In this
study, 21 breast-fed and 240 bottle-fed infants were com-
pared, and significantly greater weight gain velocity was
noted from birth to six weeks among those infants bottle-
fed. In contrast to data in the Fomon et a1. studies,
Taitz's came entirely from clinic records. There were no
exclusions for feeding solid foods in any form or amount
or at any age.

A number of other studies, however, have found no
growth differences between breast milk and formula-fed
infants. Thomson (1955) found no significant differences
in weight gain to 16 weeks between wholly bottle-fed and
wholly breast-fed members of 40 pairs of female infants in
Scotland matched for birthweight. Evans (1978) studied 94
normal British infants from birth to six months of age.
Each infant consumed either breast milk or one of three
commercial formulas. No significant differences in weight,
length, and skinfold velocities were found. It was pointed
out that the breast-fed infants made the most rapid gains
during the first month. No data were given on samp1e~
attrition. In another frequently cited study, deSwiet et
a1. (1977) found no effect on body weight of the type of
milk fed to 758 infants in Great Britain. Their comparisons
were made at six weeks and six months, among groups identi-
fied at six weeks as breast-feeding, bottle-feeding, or

mixed breast- and bottle-feeding. Some infants in each of

20
these three groups consumed solid food as well. No-
significant differences were found at six weeks or at
six months. The same three feeding types were studied to
nine weeks of age in another sample of 457 British infants
(Holly and Cullen 1977). Again, no significant differences
in weight could be found between feeding groups. Despite
the lack of significant results, an examination of the
data shows differences in the patterns of growth. The
mixed feeding group is most notable. Although it has a
lower growth rate than either of the other two feeding
groups in the first month, these infants recover and by
nine weeks are over-represented in the highest decile of
weight for age.

Saarinen and Siimes (1979) followed the growth of 238
Finnish infants during the first 12 months of life. One
group of these infants was breast-fed for at least six
months, with no solids until three and a half months. These
infants were compared with others who were weaned from breast
milk to formula by one month of age. These infants also
consumed solids after three and a half months. No signifi-
cant differences in weight or length were found between
these two groups at any point during the study. However,
an examination of the data Saarinen and Siimes present shows-
differences in the average patterns of weight gain. The
extended breast-feeding group gained weight more rapidly
for the first two months. After that, its growth velocity

slowed relative to the early weaning group, which was on the

21
average heavier from six to twelve months of age.

One recent study of 92 female infants has concentrated
on the relationship of infant fatfold thickness to feeding
modality (Ferris et a1. 1979). Four fatfolds were measured
monthly from one to six menths. There were no significant
differences among four groups of infants identified at
two months as breast-fed or formula-fed, exclusively or
with solids.

Thus, a review of the literature presents no definitive
answers to questions concerning the differential impacts of
breast milk and formula on growth. After examination of
studies involving growth and solid food, reasons for the
inconclusive nature of the breast milk/formula investigations

will be discussed.

Growth and Consumption of Non-Milk Foods

Results of studies designed to answer the question of
i the impact of solids on growth are as conflicting as those
described above from milk types. Weil (1977) provides a
review and critique of some of the older research in this
category. Other investigations will be reviewed below.

One study which is repeatedly cited as evidence for an
impact of solid food on infant weight is that of Shukla et
a1. (1972). In it, weight, length, fatfolds, and types of
foods consumed were monitored periodically from birth to one
year of age in 300 British infants. The investigators found
that at all ages, these infants consistently exceeded accepted

weight, weight for length, and fatfold standards. Furthermore,

22
they found that 40% had been given solids by four weeks and
93% by 13 weeks. From this, a casual link between the con—
sumption of solids and overweight was inferred. Almost all
these infants also consumed formula, but the investigators
failed to determine whether or not the solid food was
replacing or truly supplementing this intake.

Of the more recent studies of growth and solids, only
a few confirm the 1972 findings of Shukla et al. One of
these is Hodgson (1978). In this, 301 British infants were
weighed at birth, six weeks, six months, and one year of age.
Dietary histories were collected at unspecified times.
Hodgson divided the infants into the following groups by
date of first solid food: earlier than six weeks, six
weeks to three months, and later than three months- At
six months, infant weight of the first group was found to
be significantly greater than that of the last. Hodgson
thus concluded that early introduction of solid food leads
to increased weight gain in infancy.

Davies et al. (1977) studied 821 British infants and
used the same age divisions for introduction of solids as
those used by Hodgson (1978). In this study, however, no
significant differences in weight or length were found.

In a later study of only 50 infants, Wilkinson and Davies
(1978) again found no statistically Significant relation-
ship of growth parameters to date of introduction of solids.
Thus, studies exist to support either side of the question

of the impact of solids on infant growth.'

23

Underlying Assumptions

 

The confusing and conflicting results of the studies
discussed above appear to be derived, at least in part,
from the nature of the assumption on which the research
questions are based. There are at least five such assump-
tions, not all of equal impact and significance, and each
will be discussed in turn.

The first assumption implicit in this research is that
infants in dietary studies can be treated as though they are
randomly assigned to feeding groups. If this assumption
were valid, then there should be no differences, for
example, in demographic variables between infants breast-
or bottle-fed. However, Sauls (1979), in a review of the
infant feeding and morbidity literature, has shown that
consistent differences do exist. At present in the U.S.,
compared with bottle-feeders, breast-feeding mothers are
likely to be older, married mothers, who are healthy, non-
smokers, not employed outside the home, and not receiving
any type of governmental assistance. Furthermore, the
educational level of breast-feeding mothers is generally
higher than that of bottleefeeders (Cunningham 1977). .Such
differences make it likely that basic differences in pre-
natal influences, postnatal health care, and other environ-
mental factors exist. These may, in turn, be confounding
factors in studies which do not control for them.

A corollary of the first assumption of random assign-

ment of infants to feeding groups is the assumption that

24
"breastéfed" and "bottle-fed" are the same sort of
categories. Since, as Himes (1979) has stated, there is
a one-way passage of infants from the breast-fed to the
bottle-fed category, the former is a highly selected group.
This is particularly obvious in Ahn and MacLean's (1980)
study of growth in exclusively breast-fed infants. Their
sample size decreases rapidly after five months of age; at
eight months only 28 of 96 infants remain. There is no way
of knowing whether or not actual or perceived growth crises
were responsible for the attrition. Thirty percent of the
breast-fed infants studied by Fomon et al. (1970) dropped
out of the project by day 112, compared to an attrition
rate of only 8% among formula-fed infants (1971). A
comparison of those leaving the breast-feeding study with
those continuing to participate showed the former to be on
the average lower in weight in the early weeks of the study
than the latter. Fomon et a1. (1970) report that at least
some attrition was due to failure to thrive on breast milk.
Bottle-feeding, in contrast with breast-feeding, is more of
a cumulative group than a selected one. In studies such as
Saarinen and Siimes (1979), many of the infants in the bottle-
fed category had originally been breast-fed, but for one month
or less. The proportion of these weaned due to growth in-
adequacies is unstated. However, based on data of Hood et
a1. (1978) and Sjolin et a1. (1979), it may have been quite
high. This may, in turn, have confounded the results of growth

data analyses.

25

The second implicit assumption in these studies is
that a onefway, casual relationship exists between infant
diet and growth. That is, dietary manipulations lead to
growth differences. This orientation is contrary to what
most clinicians would acknowledge to really be the case:
that dietary changes are made when growth has deviated
from an expected norm. While a growth correction following
the diet change is expected, it was a growth, rather than
a dietary, action which originally set the chain of events
in motion. Only attending to the one-way effect of diet
on growth may obscure the interrelationships of infant
diet and growth.

The issue can be further confused by the way in which
infant dietary intake data are collected. Twenty-four hour
diet recalls and diaries such as those used by Shukla et a1.
(1972) and Evans (1978) provide information for only that
short period of time. There is no way of knowing whether
the diets reported represent long term patterns of feeding--
and thus "causes" of the ob3erved growth--or recent feeding
changes which might be "effects" of previous growth. This
problem can be overcome to some extent by increasing the
frequency of the diet recalls or, as Ferris et a1. (1979)
have done, by taking a diet history at the same time.
However, the investigator must still decide which information
is to be given priority in assigning infants to feeding groups.
And his or her assumptions about the interrelationships of

infant diet and growth will help to inform that decision.

26
This can be a fairly serious confounding factor if feeding
changes tend to coincide with times chosen for dietary data
collection. This problem has been described by Ritenbaugh
and Harrison (1980) for a study of Arizona infants in which
dietary and growth data were obtained at six month intervals,
often at approximately the same time as a regular medical
examination at which a diet change was prescribed. In
these cases several days difference in the researchers'
appointment scheduling made little difference in anthropo-
metric data, but resulted in major shifts in dietary intake.
Thus, the fact that most of the British infant studies are
conducted by physicians collecting data at regular well-
baby checkups, while most of the American studies are not,
may explain some of the conflicting results obtained. At
least it indicates another set of problems to be considered
in interpreting such results.

A third assumption about diet and growth made by
several researchers is that excessive weight indicates fat-
ness. Although it is widely recognized that the contribution
of fat to body weight is less than that of the same volume
of lean body tissue, the confusion of body size with body
composition persists. Taitz (1971), for example, uses no
measure of fat, yet talks about the prevalence of over-
nutrition and obesity in the study population. It is, in
fact, quite possible for individuals of the same body weight
to have significantly different percentages of body fat

(Brozek 1963).

27

A fourth assumption is that when solids are introduced
into the diet, they are nutritional additions to the exist-
ing dietary base. It is this assumption which has been
behind research designed to identify the suspected con-
nection between early solids and obesity. However, Marlin
et al. (1980) have shown that, at least until the age of
12 weeks, solids actually replace either breast milk or
formula in caloric terms. They conducted a longitudinal
study of 39 infants fed a diet selected by the mother.
Mothers kept three-day diet diaries for their infants at
two, four, eight and twelve weeks.’ Breast milk volume was
obtained by pre- and post-feeding infant weighings, and its
nutrient composition was estimated using values published
by Fomon (1974). When caloric intake of the entire diet
was calculated, there were no significant differences at
any age betWeen those consuming solids and those consuming
only breast milk and/or formula. The research practice of
attending only to single foods-or food categories has led
to this ignorance of the impact of dietary change on the
existing diet. It is not known whether formula also acts
as a nutritional addition or replacement in a breast milk
diet and whether solids continue to be replacements at ages
of greater than three months.

A final assumption inherent in much of the research on
infant nutrition conditions the tone in which results are
discussed and evaluated. It is the-assumption that "normal

growth" exists as a constant standard against which growth

28
or research subjects and samples can be evaluated. Shukla
et a1. (1972) and others discovering high rates of "over-
nutrition" in Great Britain in the late 19603 and early
19703 assumed this, and never investigated what their
subjects' idea of normal growth was or whether their own
medical notion of normal growth had changed through time.
Normal growth does, of course, follow certain curves at
different ages or stages of maturation. However, both the
magnitude and amplitude of the curves can vary. The
boundaries between "normal" and "abnormal" variability are
culturally defined. In the past in almoSt all cultures
and continuing to the present in many, "bigness" in infants
had been considered healthy and desirable. Bryant (1978)
found Anglo, Puerto Rican, and Cuban mothers in Dade County,
Florida still using evaporated milk formula because it
"fattens" babies better. Weil (1977) describes a group
of mothers he labels "more nutritionally;sophisticated" who
fed diets designed to produce fatter infants due to their
concerns about the possibility of brain damage resulting
from underfeeding. In contrast, among the medical community
and other segments of the public, such healthy looking
(by one standard) babies have recently come to be labeled
obese and considered at risk for adult obesity, heart
disease, and reduced lifespan (of. Charney et a1. 1976;
Kannel and Dawber 1972). Adair (1980) found white, middle-
class Philadelphia mothers manipulating their infants'

growth patterns through sometimes extreme dietary means to

29
produce the thin infants now considered "normal" by some
physicians. Thus the findings of Fomon et a1. (1971), -
Hodgson (1978), and others cited above that different
patterns of infant feeding produce statistically significant
differences in infant growth should not necessarily lead to
the conclusions that these differences: 1) are culturally
meaningful, or 2) are indicative of abnormalities in the
infants' growth.

These five assumptions, then, have shaped the research
conducted on infant growth and diet interactions in recent
years. Like any research base, they have in some ways
limited and selected for the research results which have
been derived, and they have contributed to the ambiguity

of those results.

Maternal Nutritional Status

 

Several long term studies exist of lactation and
maternal nutritional status of women living under con-
ditions of nutritional stress. A British study in the
Gambia, West Africa, for example, has identified annual
patterns of maternal weight change, birthweight, and lac-
tation performance and has correlated these with seasonal
fluctuations in disease, rainfall, and food supply para-
meters (Rowland 1978; Rowland and Paul 1981; Whitehead et
a1. 1978). In an experimental study in Guatemala, the
effects of different nutritional supplementation plans on
reproductive performance have been studied (Lechtig et a1.

1975; Newman et al. 1980). In addition to these, a large

30
body of knowledge exists specifically on the composition
of human milk under conditions of nutritional stress.
Jelliffe and Jelliffe (1978) have published a review of
this literature.

Together, the studies cited above and others like
them have provided a general picture of the average pat-
terns of maternal nutritional status and milk production
in populations existing under sub-optimal conditions. What
are missing, however, are comparable baseline studies of
lactation and the postpartum period in well-nourished
women in which the dynamics of maternal nutritional status,
including milk production, are documented. This section
will review the few studies of well-nourished mothers
which can be found in the literature. It will then con-
sider evidence from a variety of other types of studies
which point to possible patterns of change during the
postpartum period among lactating women and the ways in
which these may interact with milk production and, hence,

infant growth.

Breast-Feeding and Postpartum Weight Loss

 

It is widely assumed that women store nutrients during
pregnancy which are later used to meet the demands of lac-
tation and that, as these stores are mobilized, weight loss
occurs. Twenty years ago Hytten and Thomson commented that
there was "very little evidence in the literature on this
rather elementary point" (1961:27). The evidence that did

exist at the time of their review was inconclusive, showing

31
that women both gained and lost weight during lactation.
Hytten and Thomson also presented their own data on mean
weight change of 37 primiparae who lactated more than 10
weeks and 35 who lactated less than three weeks (1966).
Between the second and thirteenth weeks postpartum most
women lost weight. Those breast-feeding longer tended to
lose more weight than those breast-feeding for a limited
time. However, the distribution of weight change is wide
in both groups ranging from losses of over twelve pounds
to gains of more than eight. Furthermore, no data on
pregnancy weight gain or supplementation of breast milk
are given to help interpret these results.

Research since 1961 to substantiate the alleged
association of rapid weight loss and breast-feeding is
equally meager. Dennis and Bytheway (1965) combined post-
partum weight loss data gathered for other reasons in a
number of different studies. They show that between ten
days and ten weeks postpartum lactating women lose on the
average 13 pounds more than non-lactating women. While
this study is widely cited (e.g., Jelliffe and Jelliffe
1979), it is based on a highly selected sample of 17
subjects. All women with any pregnancy edema, heavy blood
loss at parturition, or cesarean delivery were excluded.
While such exclusions are reasonable in a research context,
they may invalidate or greatly limit any promises of weight

loss to individual women based on this research.

32

Naismith and Ritchie (1976) report postpartum weight
changes in 22 lactating and 20 non-lactating women. Group
weight losses from hospital discharge to three months
postpartum average 2.6 kg and 2.9 kg, respectively. By
six months, the nonflactating women had a mean cumulative
weight loss of 4.4 kg, compared to only 2.7 kg for those
lactating. These findings run counter to the assumed
relationship of postpartum weight change and breast-feeding.
However, the investigators note that dieting was reported
more frequently among non-lactating than lactating women.
This points to one of the major factors confounding such
studies. In non-laboratory samples, dietary intake cannot
be standardized. Blackburn and Galloway (1976) studied
energy intake and expenditures in 12 mothers at various
stages of lactation and five controls who were not lac-
tating. They found that breast-feeding women consumed an
average of 30 kcal/kg and expended the same amount on non-
lactation activities. In contrast, the non-lactating women
consumed an average of only 19 kcal/kg, while expending
34 kcal/kg. If these energy consumption patterns are
typical of postpartum women in general, they may explain
the failure of postpartum weight 1033 studies to provide
conclusive support for the assumed relationship of breastf

feeding to maternal body weight.

Breast-Feeding and Changes in Maternal Body Fat
The primary component of weight loss expected during

lactation is fat. Taggart et a1. (1967), in a longitudinal

33
study of primagravidae in Scotland, documented the
regular growth of adipose tissue during pregnancy. It
has been reasoned that this is an adaptive mechanism to
support breast feeding, and the automatic assumption of
fat loss during lactation follows. Popular publications
(e.g., LaLeche International 1963) promise women that the
fat will be "burned" during lactation and health pro-
fessionals predict a "slimming effect" (Jelliffe and
Jelliffe l979b:1l3). However, studies of infant feeding
and postpartum changes in maternal fatness are virtually
non-existent. A review of the literature yielded a single
research report. In the study by Naismith and Ritchie
(1976) described above, biceps, triceps, subscapular, and
superiliac fatfolds were measured and used to compute
total body fat. No distinct patterns of change in total
body fat were found which could be correlated with lac—
tation status.

Despite the lack of supporting evidence, this assumed
loss of fat during lactation has been used uncritically as
the basis for other proposed interactions, most notably
the critical body composition hypothesis of Frisch and
colleagues (cf. Frisch et al. 1973; Frisch and McArthur
1974). According to this hypothesis, a certain minimum
level of body fat is necessary for onset and maintenance of
ovulatory cycles. This is alleged to be a determining factor
in menarche as well as in the length of amenorrhea during

lactation.

34

Although what is commonly believed about the relation-
ship of infant feeding to maternal fatness is based more
on logic than on solid evidence, there are a series of
interactions documented in the literature which suggest
relationships between maternal body fat and a number of
components of the infant feeding process. The following
paragraphs review this evidence.

One breast-feeding variable--the daily frequency of
nursing-fhas been linked to levels of prolactin and gonadal
hormones in mothers. These in turn have been correlated
with maternal fatness and with the quality and quantity of
milk produced.

In both animal and human studies, the frequency of
nursing periods has been found to be directly related to
the level of serum prolactin. Grosvenor et a1. (1967)
found that frequent suckling was necessary to maintain
pituitary prolactin release in laboratory rats. The length
of the suckling stimulus had no effect on prolactin level.
Delvoye et a1. (1977) studied a group of 97 nursing mothers
in Zaire and obtained similar findings. Serum prolactin
levels rose with frequency of daily nursing. Moreover,
nursing frequency was also related to the length of time
mean prolactin levels remained elevated after parturition.
Women nursing seven or more times per day maintained high
prolactin levels for a year. Those nursing four to six
times per day returned to non-lactating levels by about

ten months, and those nursing less than four times per

35
day, by six months.

The relationship of nursing bout frequency to serum
concentrations of the gonadal hormones estrogen and pro-
gesterone has also been documented. Among lactating
lKung mothers, Konner and Worthman (1980) found that high
nursing frequency was associated with low levels of
estrogen and progesterone. Less frequent nursing correlated
with higher hormone levels. They hypothesize that this is
a result of the suppression by prolactin of gonadal hormone
production. (Tyson [1977] reviews the details of this
interaction.) Among the EKung, total nursing time and
length of feedings were uncorrelated with hormone levels.
The time interval between the initiation of nursing bouts
was the crucial variable.

Variations both in the gonadal hormone progesterone
and in prolactin have been linked to body fat. An inverse
relationship exists between serum prolactin and progesterone:
the higher the prolactin, the lower the progesterone pro-
duction. Studies in both humans (Hytten and Leitch 1971)
and rats (Galletti and Klopper 1964) show that increasing
the level of progesterone induces fat deposition. Thus,
the variations in progesterone levels mediated during
lactation by prolactin would be expected to have an impact
on fat levels. Those women nursing frequently should be
less disposed to increase fat stores than those women

nursing infrequently.

36

Prolactin is associated with body fat through the
activity of lipoprotein lipase, the enzyme responsible
for the uptake of dietary lipids into adipocytes. In
studies of laboratory animals, prolactin has been found to
induce and maintain shifts in lipoprotein lipase activity
level in different body sites (Zinder et a1. 1974; Scow
et a1. 1973). When prolactin is at normal low levels,
lipOprotein lipase activity is high in the peripheral
layers of adipose tissue. However, when prolactin levels
are elevated, there is a suppression of lipoprotein lipase
activity in adipose tissue and a concomitant increase
in activity in the mammary glands. Thus, when nursing
frequency is high, serum prolactin concentration should
be high also, causing dietary lipids to be diverted from
body stores to the mammary glands. When nursing frequency
is low, uptake of lipids by peripheral adipose tissue
should be closer to normal levels.

It is well known that the frequency of nursing bouts
has an impact on the quantity of milk produced. Egli et
a1. (1961) in a frequently cited study varied the feedings
per day of a woman who was successfully lactating. They
noted reduced milk output following a reduction in feedings
and an increase to original levels of milk production three
days after the number of feedings returned to normal. The
cause of such variations in milk production is assumed to
be the corresponding variation in prolactin, which causes

milk secretion in the alveoli.

37

In addition to affecting breast milk quantity, the
frequency of feedings affects milk quality. Tyson et a1.
(1972) found that artificially stimulated prolactin release
in lactating women was followed by increased fat content
of their breast milk. The mechanism for this may be the
prolactin-induced increases in mammary lipoprotein lipase
activity. Therefore, it appears that infant growth should
also be affected by the ongoing interactions of feeding
frequency and maternal physiological parameters during the
postpartum period.

Thus, the literature suggests that the relationship
between breast-feeding and maternal nutritional status is
considerably more complex than the simple loss of weight
and fat often assumed to occur during lactation. It
appears that attention to specific behavioral aspects of
breast-feeding may help to reveal patterns of maternal
biological response. These responses, in turn, may affect

the way in which infants grow.

Summary

This review of the literature has indicated that a
complex interaction exists among the components of the
infant feeding process: infant diet, infant nutrition,
and maternal nutritional status. It has further shown that
different aspects of this interaction are better understood
than others. This chapter has attempted to indicate what
those aspects are and why the remaining ones have either

not been studied or, having been studied, have produced

38
contradictory results.

Studies of maternal decision making in infant diet
indicate that perceptions of infant growth patterns are a.
major factor in mothers' choice of feeding strategies.
However, infant growth data are generally analyzed as
though only a one-way relationship between diet and growth
exists. Because of this and other assumptions about the
relationship of infant diet to growth, few conclusions can
be drawn at present about this interaction.

The aspect of the relationship of infant diet to
maternal nutritional status about which most is known is
the average composition and volume of milk produced by
malnourished mothers. However, the dynamics of milk
production in individual women-~particularly in well-
nourished women--are relatively unknown. The effect of
infant feeding on maternal nutritional status is widely
assumed to be a gradual draining of body stores, but
research on the hormones involved in lactation indicates
that this may be overly simplistic. This research hints
that different patterns of change in maternal nutritional
status may accompany different styles of breast-feeding.
To investigate these requires attention to behavioral as
well as biological data, and to the ways in which these

two types of phenomenon interact in individual women.

CHAPTER I I I
METHODS

This chapter describes the research undertaken to
meet the objectives identified in Chapter I. After a
discussion of the overall research plan, it details the
sample and the methods by which it was obtained. Data
collection procedures are described, and the rationales
for their use discussed. The final section of the chapter
presents the justification for combining various types of

data.

Research Design

 

Because the objectives of this research are to examine
the interaction of infant diet and growth and maternal
postpartum nutritional status through time, a longitudinal
design was selected. Such a research strategy has a dis-
tinct advantage over the less time-consuming cross-sectional
design in that it enables one to document individual courses
of growth or behavior, thus clarifying the nature of the
relationship between events at one time and those at another.

The research reported here actually comes from longi-
tudinal studies of two samples from a single population.

The first, an exploratory study referred to hereafter as

39

40
Study I, was conducted in central Michigan from January
1978 through January 1979. Its purpose was to examine
general relationships between infant diet and growth. The
data collected in Study I were subjected to only minimal
analysis (Quandt and Ritenbaugh 1979). Further analysis
was postponed because, in the course of the study, it
became apparent that a number of potentially interesting
problems in the data could be explored only with the
collection of additional behavioral data.

Thus, the second study (Study II) was conducted from
August 1979 through February 1981. It continued the same
data collection strategy used in Study I, but added other
items to the research protocol which were designed to ex-
plore the problem areas which had emerged during Study I.-
It should be stressed that although it was hoped the two
samples could be combined into one larger one, Study II
was not in any way intended to increase sample size to
improve levels of statistical significance of some already
recognized associations.

It was planned that, if no differences were found
between the two studies in basic data (e.g., pregnancy
characteristics, parental size), it could be assumed that
they were samples from the same population. Thus, while
different frequencies of behaviors might exist at the times
of the two studies--and indeed would be expected given the
rapid changes which take place in infant feeding norms--

this would not preclude their being combined into a single

41
sample. Biologically similar individuals should react
in similar ways to the same diet or behavior at different
times (Bass et a1. 1979). On the other hand, it was
assumed that for any analysis examining cultural factors
and their changes through time, the two studies could

legitimately be separated into two independent groups.

The Sample

 

Description of the Sample

 

The total sample consisted of 79 mothers and 80
neonates. All infants weighed at least 2,500 grams at
birth and were products of at least 36 week pregnancies.
Forty-five of these mothers and 46 infants (44 singleton
infants and one set of fraternal twins) participated in
Study I. The remaining 34 mothers and 34 infants were
in Study II. Table III.1 shows the number of infants by
sex, Study, and month of participation. The diminishing
sample size is due to mothers moving from the area. This
group of subjects was taken from a larger sample, drawn
to study other pregnancy and postpartum issues, which
included infants never breast-fed, infants born prematurely,
and infants whose mothers left the study area before two

months postpartum.

42

TABLE III.l—-Number of Infants Measured at Birth and at
each Month.

 

 

 

 

 

Month‘
Sample Birth 1 2 3 4 5 6
TOTAL SAMPLE
Total N 80 80 80 79 76 75 72
Females 43 43 43 42 41 40 40
Males 37 37 37 37 35 35 32
STUDY I
Total N 46 46 46 45 44 43 41
Females 27 27 27 26 25 24 24
Males 19 l9 l9 l9 l9 19 17
STUDY II
Total N 34 34 34 34 32 32 31
Females 16 16 16 16 16 16 16
Males 18 18 18 l8 l6 16 15

The sample used in this study represents a predominantly
middle class population. The annual household incomes ranged
from $5,000 to over $40,000; the median falls in the $20,000
to $24,999 category (Table III.2). The median income was
slightly higher in Study II than in Study I. However, con-
sidering the inflation rate during the two years between
studies, the effective income can be considered equal.
Maternal education ranged from high school graduate to Ph.D.
The modal category was composed of women who had completed

a bachelor's degree program (Table III.3).

43

TABLE III.2--Gross Household Income by Study and for Total

 

 

 

Sample. ~
- ---- Study-II-- Study II Total
Income Category N N N
(1) $5,000 - 9,999 3 2 5
(2) $10,000 - 14,999 9 4 l3
(3) $15,000 — 19,999 11 7 .18
(4) $20,000 — 24,999 10 12 22
(5) $25,000 - 29,999 7 5 12
(6) $30,000 - 39,999 6 2 8
(7) $40,000 0 2 2

TABLE III.3--Highest Level of Maternal Formal Education by
Study and for Total Sample.

 

 

Study I Study II Total
Education Category N N N
(1) High School or 10 9 . 19
Vocational School
(2) Some College (Less 10 ' 8 18
than Bachelor's
Degree)
(3) Bachelor's Degree 16 ll 27

(4) Any Graduate School 10 6 l6

44

Sample Recruitment

 

Throughout the sample recruitment, care was taken not
to influence potential participants' choices of feeding
modality or to differentially attract women with different
feeding plans. Women asked to participate in the study
were told that the researcher had a number of interests
that involved new mothers and their infants. They were
told that she was aware that there were many different
ways in which mothers fed their infants and that she was
trying to find out what these feeding methods were. She
was also interested in the way infants grew and how that
might be related to the way they were fed. Finally, they
were told that some mothers might be breast-feeding their
infants and, in those cases, maternal postpartum anthropo-
metric data might help to explain infant growth. Since
comparisons between those breast-feeding and those not
breast-feeding were useful, they were told, the researcher
would like to monitor postpartum changes in all mothers.

Slightly different procedures for enrolling mothers
were followed in Studies I and II. Women participating in
Study II were identified during pregnancy with the assistance
of obstetricians in private practice in the Lansing area.
The physicians agreed to give letters explaining the study
(Appendix A) to patients who met the criteria for the study.
These criteria are discussed below. These women were then
contacted by telephone by the investigator, who further

explained the research and asked their participation. The

45
rate of refusal was approximately 20%.

In Study I, women already enrolled in a study of
pregnancy weight gain were given a letter by the 36th week
of pregnancy requesting their participation in a study of
infant growth and dietary practices (Appendix A). These
women had originally been recruited through most of the
same physicians and in the same way as described for
Study II. The rate of refusal to participate in the preg—
nancy weight gain study was approximately 10%; that for
the follow-up infant study was 8%.

In order to be included in either study, women had
to meet certain criteria. First, only Caucasian women
were contacted in order to minimize possible genetic
differences in growth patterns. Second, each woman had
to be either a primipara or, if a multipara, expecting
her first liveborn child. This requirement was an attempt
to, minimally at least, standardize experience in child-
rearing. Finally, the women were expected by their
physicians to have no pregnancy complications which might
adversely affect the health or development of the infant.
There were no criteria related to mode of infant feeding,
and the fact that breast-feeding was of primary interest
was not mentioned during any of the study recruitment or
during the course of the research.

Mothers read and signed statements of informed consent

(Appendix A) at their first meeting with the investigator.

46

Data Collection

 

All data were collected in the subjects' homes at
their convenience. S. Quandt collected all Study II data
and about 80% of that in Study I. C. Ritenbaugh collected

the remaining 20%.

Maternal Data Collection

 

The nutritional status of mothers was assessed anthro-
pometrically eight times: once during the third trimester
of pregnancy (a maximum of two weeks before delivery),
early in the postpartum, and at monthly intervals from
one to six months postpartum. The early postpartum data
were collected at one week after delivery in Study I and
at two weeks, in Study II. This change was made to
facilitate the collection of infant data described below.
Measurements on each of the occasions included weight,
mid-arm circumference, and triceps fatfold. Weight was
measured to the nearest tenth of a kilogram on an Accu-
Weigh portable beam balance. Women were weighed without
shoes; they were requested to wear approximately the same
weight indoor clothing for each weighing. A steel tape and
Lange skinfold calipers were used for the remaining anthro-
pometrics. Standard methods for mid-arm circumferences and
triceps fatfold described by Weiner and Lourie (1969) were
used.

Stature was measured using a free-standing anthropometer
once during the study. Prepregnant weight was obtained from

the mother when the prenatal measurements were made. She

47
was asked to give her best recollection of her weight at
the time of conception, rather than her "normal" weight,
which might have been different.

Although it is possible to measure fatfolds at a variety
of sites on the body, triceps was chosen for two reasons.
First, unlike sites such as superiliac or midabdomen,
triceps is unlikely to be distorted shortly before or after
parturition by gross anatomical adjustments related to
delivery. Second, triceps, along with mid-arm circumfer-
ence enables one to compute upper arm fat area (Gurney and
Jelliffe 1973). This is an expression of adiposity which
can be used to assess changes in adiposity independent of
changes in underlying lean body tissue. It may be parti-
cularly important in monitoring changes following pregnancy,
which is--to a variable degree in all women--a super-
hydrated physiological state.

No biochemical or dietary intake data were collected
as additional assessments of maternal nutritional status.
Although such information would possibly have been valuable,
it was decided that the high degree of long term cooperation
required for the study as a whole might be jeopardized
with the inclusion of these types of data collection. The
collection of dietary data is not usually considered invasive.
However, it was observed that many of these women were ex:
tremely sensitive about their food consumption during the
postpartum. Although most had expected to weigh more after

pregnancy than before, many were quite unprepared for the

48
proportion of weight left to lose. Furthermore, they
experienced conflicting pressures regarding diet. On the
one hand, they were told to eat more than usual for breast-
feeding. On the other, they were told to eat less than
usual to return to their pre-pregnant size and shape.
Thus, because of the feelings attached to food intake, it
was decided that the collection of such data might endanger
the rapport between mothers and researcher necessary for
other portions of the study.

Mothers were asked their expected delivery dates at
the prenatal measuring session, as well as the method by
which the dates were calculated.

Parental ages at the child's birth as well as socio-
economic information were solicited during an interview
at the six month measuring session. Total gross household
income for the previous year was obtained, as well as
highest level of maternal education. The decision to use
maternal education was based on research by Green that
indicates that in American society, most health related
decisions are made by mothers; and thus their education
rather than that of fathers correlates most closely with
health behaviors (1970). Observations were also made of
the style of dwelling in which the mother and infant lived.
Since it has been suggested by others that differences in
physical environment affect maternal child interaction
including breast-feeding in the neonatal period (Klaus and

Kennell 1970), the possibility that gross differences in

49

environment might continue to be influential was considered.

Infant Data Collection-—Anthrgpometric

 

Infant growth was assessed seven times during the
study: in the hospital within 24 hours of birth and at
home, monthly from ages one to six months. Weight, length,
and eight fatfolds were measured on each occasion. Birth
weight was obtained from the hospital chart. Subsequent.
weights were calculated to the nearest 0.1 kg by subtracting
the weight of the mother from that of the mother holding
the child on the beam balance. Infants were weighed naked
except for a diaper. Length was measured to the nearest
0.1 cm as the distance on a flat surface from the crown of
the head to the soles of the feet. The researcher positioned
the child and applied pressure to the infant's knees while
another person (usually the mother) held the head still
and in the sagittal plane. The triceps fatfold was measured
on the posterior of the arm midway between olecranon and
acromion. Biceps fatfold was taken at the same level on the
anterior surface of the arm over the biceps muscle. The
midabdominal fatfold was measured in a vertical line at a
point 2 cm lateral to the umbilicus, and the subscapular,
below the angle of the scapula on the back. All of these
fatfolds were taken bilaterally with the child in a relaxed,
recumbant position. The calipers were applied to the fold
and read as soon as the needle had stabilized, after the
method of Brans et a1. (1974). A total fatfolds variable

was calculated by summing each set of fatfolds. This was

50
done to control for assymetry and trunk-extremity variability

in fat distribution.

Infant Data Collection--Dieta£y

 

Data pertaining to infant dietary intake were obtained
through written records kept by the mother and through
information elicited during interviews at monthly measure-
ment sessions. The emphasis on mothers was not meant to
deny the role others--particu1arly fathers--play in infant
feeding. Whenever other family members were present during
a measuring session, their views were solicited as well as
those of the mother. However, since among families such as
those in the pOpulation represented by this sample, mothers
have the major responsibility for care of newborn infants,
most of the data collection strategy concentrated on mothers.

Mothers completed 24-hour diet diaries for the infants
approximately every eight days from the second week post-
partum to the end of the study. These forms were mailed to
the mother at the appropriate interval with a stamped,
addressed envelOpe in which to return the completed diary.
Mothers were asked to complete the form as soon as possible,
recording everything the child consumed during a 24-hour
period, the times of consumption, and the amount consumed.
For breast-feeding, they were asked to note the minutes of
active suckling for each nursing bout. Time and length of
sleep episodes were also noted to provide information on

feeding frequency.

51

Of the numerous methods used by researchers to collect
dietary intake data, the diet diary system was chosen
because it seemed best able to provide the data on infant
feeding behaviors desired for this investigation while
minimizing problems inherent in many recall reporting
techniques. Data were reported frequently enough to allow
reasonable verification of dates of dietary changes such
as first introduction of solids, breast milk to formula
transitions, and elimination of night feedings. Had
information on specific nutrient intake been desired, more
or less frequent reporting might have been required. With
the eight day interval, food behaviors specific to a parti-
cular day of the week were not overrepresented. Because
the data were recorded by the mother or caretaker as the
day progressed, the problems of accurately remembering
numerous similar events were avoided.

The level of cooperation of mothers in completing and
returning the diet diaries was generally high (Appendix B).
The median number of diaries returned was 15. Mothers
tended to be more conscientious about the forms during the
first months of their participation. It became apparent
within the first few weeks postpartum that certain mothers--
often despite their professed good intentions--were never
going to return many of the diet diaries. In these cases
other means of obtaining the data were intensified. These

methods are described below.

52

The second source of infant diet data is information
obtained during both unstructured and formal interviews
and through observations conducted whenever the investi-
gator met with subjects. In Study I, observations and
unstructured interviews took place at all postpartum
measuring sessions. Questions were asked to verify changes
in infant feeding which had appeared on the diet diaries
since the previous meeting. In cases where no diet diaries
had been completed, specific questions about the infant's
usual diet-were asked. At the sixth month measuring
session a more formal interview took place during which
mothers were asked to give as detailed a diet history
from birth to six months for the infant as they could
remember. They were also questioned about their own re-
covery from pregnancy and childbirth (e.g., when did they
resume a normal schedule of activities; when did they feel
like they were really "back on their feet?) Responses
to questions at six months were written during the inter-
view. On all occasions prior to that, notes were written
after leaving the subject's home.

In Study II the content of the postpartum interviews
was expanded. In addition to using measuring sessions to
verify content of the diet diaries, information was elicited
on the decision making process related to each dietary
change. While the sequence and wording of questions varied,
in general the investigator tried to establish why the

change had been made, whether or not there had been input

53
from family members or others, and whether or not any
behavioral or physical change observable in the child
preceded or followed the dietary change. Much of this
discussion took place while the child was being measured
30, as in Study I, responses were recorded after leaving
the home.

At six months a formal final interview was conducted
with Study II mothers. They were asked to give a diet
history for their infants and describe the decision making
process involved at the various transition points. The
same questions on maternal recovery were asked as had been
asked in Study I. Each mother was also asked to describe
her infant's body shape by choosing from a set of six
infant outlines varying from thin to fat (Appendix C) the
one she thought looked most like her own child.

In any case where a mother did not complete a diet
diary for a particular time period for which data is
analyzed in later chapters, one of two possible actions
was taken during the analysis. If data were available
from interviews during that time, such data were used and
the mother or infant included in the analysis. If none
were available, the subjects were excluded from that parti-
cular analysis. Data from interviews on types of foods
consumed and dates of introducing new foods or weaning
were considered of acceptable accuracy. Data on numbers
of feedings per day and length of feedings were not. Hence,

the numbers of cases varies somewhat in those data analyses

54

involving any of the infant feeding data.

Statistical Considerations and Rgporting Reliability

Levels of Statistical Significance

 

The statistics used throughout this dissertation vary
with data being analyzed. Anthropometric data meet the
conditions of statistics which assume a normal distribution
and an interval or ratio level of measurement. However,
many of the other data--e.g., socioeconomic measures,
numbers of breast-feedings--do not meet such assumptions.
Therefore, nonparametric tests are used in their analysis.
At various points throughout the description of the results
(Chapter IV), specific statistical treatments will be done,
based on subgroupings of data derived in the course of the
data analysis. These statistical treatments will be de-
scribed in the apprOpriate section of Chapter IV.

When the results of statistical tests are reported in
this dissertation, the actual significance level associated
with each test will be given. Any results at or below the
0.05 level of probability that the relationship is due to
chance will be considered significant. This level was
chosen because many of the tests used involve nonparametric
statistics. The weakness of such tests is their relatively
low power. That is, they fail to show statistical signifi-
cance even when samples being compared are different. Hence,
a 0.05 level of probability is normally used in such tests.

In some of the description of the data analysis,

results significant at the level of up to 0.10 will be

55
discussed, though not described as statistically signifi-
cant. This will be done where groups with small sample
size are being compared because lower sample sizes, in
general, reduce the power of a test. That is, they increase
the likelihood of failing to show statistically significant

associations where such associations do exist (Siegel 1956).

Indicators of Accuracy: Infant Diet Diaries

 

Because the 24-hour diet diary has not been widely
reported as a means of collecting infant dietary data and
because verification could only be made by further question-
ing of the same person who had filled out the diary, several
factors were examined which were assumed to indicate
accuracy in completing the forms.

One indicator was the presence or absence of night
feedings. In all cases when a child was reported to be
regularly sleeping through the night no night feedings were
noted in the diary. More importantly, when many infants were
reported to be awakening for night feedings again at about
five months of age, the night feedings reappeared on the
diet diary forms.

A second indicator of reporting accuracy was evidence
that the diet diaries were completed as ongoing records
rather than as a mother's recollections at the end of a day.
In many cases, differences in writing implements and in hand
writing styles indicated that feeding events were jotted
down as they occurred throughout the day, instead of all at

one time. Many mothers reported carrying the form around

56
in a pocket during the day they chose to record data, and
the crumpled condition of the forms supported this. Also,
in cases where a mother had returned to work, notes by the
babysitter were evident because of differences in handwriting
and because of the types of foods fed the child—-e.g.,
"frozen mother's milk."

A third possible indicator was variation in reported
length of breast-feeding bouts. It might be expected that
feedings reported in five minute increments were estimates
or evidence of rounding off the actual length of a feeding.
While such times appeared fairly commonly, closer checking
revealed that for many infants, this was accurate reporting.
Mothers had read or been told that they should limit nursing
to a certain number of minutes to avoid nipple irritation,
and they tended to stay with such a pattern for a long time.

The final indicator was the correlation of infant growth
extremes with diets reported. As will be described below,
such correlations exist. Moreover, significant relation-
ships exist between details of reported breast-feeding be-
havior and maternal body composition.

Thus, despite the lack of an independent means of
verifying the accuracy of the diet diaries, the indicators
discussed above suggest that the majority of mothers were

honest and accurate in their reporting.

Combining Subjects from Study I and Study II

 

In order to combine data from Studies I and II for some

of the analytical procedures which follow, it was necessary

57
to establish whether or not the two samples differed signi-
ficantly. Several categories of variables were examined:
parental biological characteristics, pregnancy, birth
outcome, and socioeconomic status.

Analysis of variance showed no statistically signifi-
cant differences between the studies in parental ages,
heights, or weights (p <.05; Table III.4). There also
were no differences in length of pregnancy, maternal weight
gain during pregnancy, and infant birth weight. Chi square
tests of independence were used to compare income and edu-
cation categories for the two studies.8 Again, no statisti-
cally significant differences were found.

Thus, for all variables it was impossible to reject
the null hypothesis that the two samples do not differ
biologically and demographically. For this reason, the
samples are combined for analysis where appropriate in

Chapter IV.

58

 

N8. 03.3 0.0 H 0.0 8.0 H 8.0 1033 32038333333
00. 30.3 8.33 H 0.003 3.03 H 3.003 383883 230003 300800833
00. 30.3 0.0 H 0.03 0.0 H 0.83 1033 e388 3:0383 300800833
80. 30.3 0.0 H 3.83 0.0 H 8.03 1838031 804 38:38383
88. 83.3 0.0 H 0.83 «.0 H 0.33 1838030 803 38038382
88. 30.3 0.8 H 8.383 3.3 H 0.033 1301 300303 38038383
03. 08.3 8.0 H 8.03 0.33 H 8.00 1033 330383 38038383
80. 30.3 0.3 H 0.883 0.0 3 0.083 1301 300383 38:38383
88. 33.3 8.3 3 8.38 8.0 3 0.08 303v 830383 3:800833833
33 30038 3 30030
038 3 8338338>

>00 838 H 0802

 

 

.HH acsum msmuo>

H xcsum "0038338> mo mHmaamsm .moanmwum> HmofimoHoflm mo conflnomfioollv.HHH m3m<e

CHAPTER IV

RESULTS

The purpose of this chapter is to present the results
of data analyses. First, the infant dietary data are
discussed, both in terms of specific foods and feeding
behaviors and in terms of general patterns of individual
courses of feeding. The second section deals with infant
growth. Data on factors related to maternal decision
making regarding infant feeding are presented in the third
section. The last describes analysis of postpartum changes

in maternal nutritional status.

Infant Diet

 

General Description of Infant Diets

 

Figure IV.l shows the number and percentage of infants
who consumed breast milk and were weaned during the first
through sixth months of life. Different authors use the
term weaning for different changes in a breast milk diet.
It is used throughout this study as the total cessation of
breast feeding. "Breast-feeding" refers to consumption of
any breast milk, whether in combination with other foods or
alone. Breast-feeding decreased in frequency over time,

with the greatest decrement during the fourth month, when
59

.oMHA mo mcucoz
cuxwm nmsounu umHHm @3333c Axawz umcoum o: mC3Esmcoov cocmmz can
AxHMZ ummoum >30 mC3EDmcoov comlummoum muccmcH mo ucoouom can HonEDZII3.>H mmDon

 

 

 

60

 

 

 

 

 

 

 

 

 

 

£8.00 03 I L 28.08 80 8
338.08 03 .. (M 308.08 38 8
303.08 33 l L 80.08 88 0
383.03 83 I p. 28.8 88 8
300.08 0 ! L 30.08 80 3
300.00 0 m C 100.33 00 3

c: z 2: z

mmmmmm 803-388838 aucoz

61
12% of the infants were weaned. Sixty-three percent of
the infants were still being breastefed at six months
of age.

By the end of the first month, 20% of the infants
received commercial formula of some kind (Table IV.1).

This prOportion increased to a peak of 41% in the fifth
month. The largest incremental change after the first
month was a change of 11% in the second month. Cows milk
replaced or supplemented formula or breast milk in only a
small percentage of cases; in no case was cows milk con-
sumption reported until the fourth month. There were no
reports of home prepared evaporated milk formula being
consumed by any infant.

The term "solids" is used throughout this study to
refer to all foods except breast milk, formula, or cows
milk. It includes fruit juices. The addition of solids
to the infant diet in general started later than that of
formula, but the rate of increase over the six month period
was greater, with the cumulative total rising smoothly from
9% in the first month to 93% in the sixth. An increase in
29% in the fifth month was the greatest monthly change.

Cereal was the first solid food given to most of these
infants (Table IV.2), regardless of timing of introduction
of solids. Seventy percent of those ever consuming non-milk
foods were fed cereal first. An additional 14% consumed it
as a first food in combination with fruit, vegetables, or

meat.

62

TABLE IV.1--Number and Percentage of Infants Consuming
Major Non-Breast Milk Food Types during each
Month of the Study.

 

Month'

 

 

Type of Food 1 2 3 4 5 6

Formula N 16 25 28 28 31 26

(%) (.20) (.31) (.35) (.37) (.41) (.36)

Solids N 7 18 29 42 63 67

(%) (.09) (.23) (.37) (.55) (.84) (.93)

Cows Milk N - - - 2 9 l4

(%) - - - (.03) (.12) (.19)

TABLE IV.2--First Solid Consumed; Number of Infants by
Month of Introduction.

 

 

Month
Food 1 2 3 4 5 6 Total
Cereal 3 12 8 10 16 3 52
Fruit/Juice 2 l l 2 4 l 11
Meat - - - - - 1 1
Cereal + (other) 1 - 3 3 2 l 10
Totals 6 13 12 15 22 6 741

1Six infants consumed no solids.

63

Table IV.3 shows the frequency of combined consumption
of major dietary components. Concurrent consumption of
breast milk and formula was most frequent during the first
month and disappeared entirely by six months. In contrast,
frequency of breast milk and solids combinations increased
from lows of 3% and 4% during the first two months to a
high of 74% by six months. This increase was greatest
(26%) during the fifth month. Consumption of both solids
and formula with breast milk ranged from 5% to 17% and

followed no distinct pattern during the six month period.

TABLE IV.3--Percentage and Number of Infants Consuming
Combined Dietary Components during each
Month of the Study.

 

 

 

Month
Diets l 2 3 4 5 6
Breast Milk N 12 9 4 2 l -
& Formula (%) (.15) (.12) (.06) (.04) (.02) -
Breast Milk N 2 5 7 17 29 34
& Solids (%) (.03) (.04) (.11) (.31) (.57) (.74)
Breast Milk, N 4 10 11 “7 10 8
Formula (%) (.05) (.14) (.17) (.13) (.20) (.17)

& Solids

64

Association of Infant Diet with Maternal Education and
HousehoId Income

 

 

Highest level of maternal formal education and first
introduction of non-breast milk food are not independent
(Table IV.4). Mothers with less than a four year college
education tended to supplement at or before the median month
of supplementation (the fourth month), while those with more
education supplemented later (x2 = 5.55, p < .02). Mater-
nal education is also related to weaning (Table IV.5).

Women with fewer years of education weaned before six months
more frequently and, within that time, earlier than women
with more education. The relationship between the occurrence
of weaning before or after six months and maternal education
is significant (x2 = 10.90; p < .001). However, total
household income is not significantly related to infant diet.
Timing of first supplementation and incidence and date of
weaning are distributed evenly across the range of incomes
found in the sample (Tables IV.6, IV.7).

A review of sex differences in consumption of single
food items or dietary combinations reveals a weak associ-
ation of longer breast-feeding with female infants (Figure
IV.2). Although a higher percentage of females were weaned
initially, by four months of age males had a higher rate of
weaning. At the end of six months, the sex differences in
numbers weaned and still breast-fed are not statistically
significant. The lower incidence of breast—feeding among
males is reflected in their higher rate of formula consump-

tion (Table IV.8).

65

.Ammma 30m03mv 303008 0:3 A m0sa0> mo 03300

00333800 053 £333 E023 m030QEoo cc0 c03c0E 0s» m. 00330> mo 03300 m0cwheoo
3003 03:9 .0003 30 m 03 >030sv03m c0300mx0 0:3 50353 30% 303x0 03300 0030003
0oz0cc0m0c33 303 3003 0303001350 023 30:3 303303 030: U003 03 3003 :03c0E 0c9m

.0m0 mo mnucos 0 030303 mcsum 033 3303 0330303 0031300033 >H0>3msaox0 039

 

~0.v_ d .88.8 n ~x .0 03002 n 003002. "33003 003002
300 03 03 m3 0 33 03 030303
8.2.... .88. {2.88
88 m 8 8 0 8 33 003000 0.30300000
:mmm ...... w ....... ....3 ................ Human...
03003 I:

 

 

 

.coom £332 30003mlcoz 30333 no
303305c03333 30 £330: cc0 30330oscm 308303 30230302 mo H0>0A um0£m3milv.>H mqmfla

66

Hoo.v m «om.oa u mx 3000 30 033308 c 30330 30

030303 @232003 30 002003033 cc0 303300200 30 0030cc0m0003 30 3003 030sv013no

.0233003 30

 

38 0 m m 0 8 0 030303
--m-----m-----... ..... .... 8.... 9.9.8..
83 0 m N 8 8 0 003000 0.30300000
....... ............... 33.33.33.300
03002

 

00002 000 003300003 300303 30030302 30 30>03 300003mun8.>3 33000

67

.03300 >3m80
030238330 03 .3002032 033 023 030 00 .00233800 030 0033000300 300303 033 0:93

 

 

8
00 03 03 83 8 33 03 030300
8 u u 3 n 3 . 000.000.3
0 3 3 3 m 3 8 000.08 n 000.080
03 3 0 0 3 3 3 000.08 1 000.880
88 8 8 8 3 8 8 000.00 s 000.000
03 0 0 8 3 3 8 000.03 I 000.830
83 3 8 n 8 8 8 000.03 s 000.030
8 I 8 3 n i . 000.0 u 000.8 0
030303 8 8 0 8 8 3 000003
03002

 

.23000300 080023
030200202 >3 0002 x332 30003mi2oz 30333 30 203302003323 30 232021Im.>3 33223

68

0033000300 00030800 0333 00300800 000 00030800 000.0mw.w 0033000300 080003

.ooo.m~w A

30003330030 300 N00.3 0x N000.300.000.002 n 003002

000 30 003002 X30

30330 000 030300 0030003 30 000003003 000 080003 030000000 30 3009 3003002

 

00 38 u 8 8 0 0 0 030303
0 0 u u 1 n u 1 000.000 M
0 8 u - 0 3 u n 000.08 I 000.080
03 03 u 3 - u I 3 000.00 s 000.800
00 03 u u u 8 8 0 000.00 n 000.000
03 03 u 0 u 3 0 3 000.03 . 000.830
83 0 u n 3 8 3 3 000.03 . 000.030
8 8 n u n u u u 000.0 I 000.8 0
30303 0 0 8 0 8 0 3 000003
0300:

 

 

 

.33000300 080003 030000000 00 0030003 30 030021I>.>3 03009

69

TABLE IV.8--Percentage and Number of Infants Consuming
Formula by Sex and Month.

 

 

 

 

Month

1 2 3 4 5 6

Males N 7 14 15 15 18 14
% (.19) (.38) (.41) (.43) (.42) (.44)

Females N 9 ll 13 '13 13 12
% (.21) (.26) (.31) (.32) (.33) (.30)

Total N 16 25 28 28 31 26
% (.20) (.31) (.35) (.37) (.41) (.36)

A higher proportion of males were fed solids than fe-
males in any month (Table IV.9). By six months all males
were consuming solid food. Sex differences in introduction
of first solids are more apparent in Study I than in Study
II (Table IV.10). The distribution of Study I males is
bimodal, with the highest percentage of male infants re-
ceiving first solids at three months of age or less or at
five months or more. Study I females were introduced to
solids at a less variable rate, with a peak in the fourth
month. In contrast, Study II males and females had almost
identical distributions, with median age of introduction

of solids at four months (Figure IV.3).

70

000.00
000.ov

Ahm.ov
008.00

000.00
338.00

303.00
303.00

303.00
300.00

Aoo.ov

000.00
300

.30030 30 03002 30 003030 003000m|30003m 03 00000303330 x0mllm.>3 000030

 

33

 

m3

 

 

 

33

 

m3

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

83_L_.

 

20

000003

380.00
Amm.ov

0m0.ov
0mm.ov

300.00
Amm.ov

Avm.ov
300.00

300.00
0m0.ov

000.30
Aoo.3v
300

0m
03

mm
mm

3m
0N

mm
3m

mm
mm

mv
0m
2

 

0001300030

0030800
00302

0030800
00302

0030800
00302

0030800
00302

0030800
00302

0030800
00302

 

71

TABLE IV.9--Percentage and Number of Infants Consuming Solid
"Food by Sex and Month.

 

 

 

Month

 

Males N 4 10 17 22 32 6
% (.ll) (.27) (.46) (.63) (.91) (.100)

Females N 3 8 12 20 31 3S
% (.07) (.19) (.29) (.49) (.78) (.88)

Total N 7 18 19 22 63 67

% (.09) (.23) (.24) (.29) (.84) (.93)

.-~-~_--‘*n-‘-‘----“*--‘---_-‘------n~--_-~*~—---*-‘~“~---~

TABLE IV.10—-Number and Percentage of Infants Consuming First
Solid Food by Month, Sex, and Study.

 

 

 

 

Month
1 2 3 4 5 6 Total
STUDY I Males N 3 5 4 l 3 3 19
% (.16) (.26) (.21) (.05) (.16) (.16) (1.00)

Females N 2 5 3 6 5 2 23
%

STUDY II Males N O 2 3 5 6 l 17
%

Females N 1 l 2 3 8 0 15
%

PERCENT OF EACH SEX INFANTS

72

.50

.40 STUDY I

.30

.20 ‘

.10

 

 

.50 1

 

STUDY II
.40

.30

.20 .

.10

 

MONTHS

FIGURE IV.3--Percent of Infants First Consuming Solids
for Male (O—-——O) and Female (l——-l)
Infants by Month of Age.

73

Length and Frequency of Nursing Bouts

 

The only infants included in analyses in this section
are those for whom at least one initial diet diary of
exclusive breastffeeding is available. This excludes
those 18 infants whose diets were supplemented at or
before 30 days. Also, as was discussed in Chapter III,
infants whose diet diaries are incomplete or missing for
a particular period are excluded from that section of the
analysis.

Mean number of feedings per day in the first month
ranged from 5 to 13, with a median of 7 feedings per day
(Figure IV.4). In the second month, the median of mean
feedings per day of those still exclusively breast-fed
remains at 7, although the range shifts slightly lower.

Infants receiving first non-breast milk foods during
Months 1 and 2 tended to have fewer feedings per day in
the first month than did those infants exclusively breast-
fed for longer periods of time (Figure IV.5). The median
number of feedings per day ranged from six among infants
supplemented between one and two months to eight among
those supplemented during the third month or later. A
median test produces a x2 of 13.64, showing a significant
(p '<.01) relationship between month of supplementation
and proportion of infants less than and equal to ( y: )
or exceeding ( > ) the overall sample median of seven
feedings (Table IV.11). Similarly, infants exclusively

breast-fed longer had more feedings per day in Month 2

74
(Figure IV.6). A median test demonstrates the statistical
significance of the association between length of exclu-
sive breastefeeding and the number of feedings per day in
Month 2 (Table IV.12); p < .05).

The duration of exclusive breast feeding also varied
with the change in the number of feedings per day for
individual infants between the first and second months
(Table IV.13). Although the overall tendency was for the
number of feedings to decrease with time, infants who
consumed their first non-breast milk foods later appeared
to be more likely to have the same or greater number of
feedings in the second as in the first month. Conversely,
the fewer the months of exclusive breast-feeding, the
more likely was the infant to have had a stable or de-
creasing number of feedings from one to two months of
age.

Mean minutes per feeding ranged from 7 to 53 in the
first month and from 7 to 42 in the second. Both had
medians of 18 minutes. There was no correlation of
minutes per feeding with length of exclusive breast-
feeding similar to that observed in number of feedings per
day (Table IV.14). In both months, the distribution of
minutes per feeding of those feeding non-breast milk foods
early is not significantly different from that of those

doing So later (Figure IV.7 and IV.8).

75

N 29202

.N 000 3
003002 003300 0001300030 030>3003oxm 0300003
3om 000 300 0m030000l30003m 00 300802 0m030>0llv.>3 000030

mBZ¢0ZH 00 mmmZDZ

 

 

 

 

 

 

 

 

 

 

 

 

 

 

HEM.
“.2 U 3 5.202
u
U

N3

m3

L—

 

 

NdQ 000 mUZHDmmm 00 KWMZDZ mw¢mm>¢

76

 

 

 

 

 

 

 

 

 

 

 

 

 

1 F‘—
11
1
¢ *
g 2
H
a 1
gr: 1:1
9' 1
U)
n: 1
E 1
m 3
E 1—
§ ‘ {—1__ I747 NF]
2
m 1
U) *
a 4
z 4 1
O
' , , [_1
(
>4 1 *
UHHH n
5 6 7 8 91011

12 13
FEEDINGS PER DAY (*Monthly Medianl)

FIGURE IV.5--Number of Breast-Feedings per Day in Month 1
' by Length of Exclusive Breast-Feeding.

1Since there are an even number of observations, median
for Month 3 is the mean of the two observations in the
central position (Dixon and Massey 1969).

77

 

 

 

 

 

 

 

 

 

 

 

 

 

 

T
'k

2 4
0 fl 4
2
H
o
0
It}
‘7 ‘ "—1:
a 3 "
U)
fi .
E H [1 mm
0
>
a 4
3 4 Fh'kr
(>3 41
[:J
E? 1 H
a {—1 e
g *
2 PW

>4 :

 

 

 

 

UHF—l
4567891011

FEEDINGS PER DAY (*Monthly Median)

FIGURE IV.6--Number of Breast-Feedings per Day in Month 2
by Length of Exclusive Breast-Feeding.

78

H
O
v

CL
V
\O
0;
H

II

N

:4

>1

(0
Q
\

U)

01

C
"-1
'O

(D

(D
In
l‘

H

C.

10
"-1
‘U

(D
0
4..)

U)

Q)
E4

0

10
-H
'U

Q)
2

mm wm mm 030309

0 m v 003002 3000 A

m3 m3 0 . 003800 3:00

0 m 0 003002 00309

m3 3 «3 003:0: 033 M
030309 000\00000 0 000\00000 0 . 00300001300030
0300003 30 300802 0300303.30 300802 .0>3003000 003002

.00300001300030

0>3003000 30 030003 00 3 03002 03 000 300 000300001300030 30 3008021133.>3 00009

79

«w 03

0 0

03 03

03 m
030309 000\00000 0

0300303 00 300802

0>3003000 30 030000 00 N 03002 03 >00

 

N
mm 030309
m 003002 3000 A
0 003002 3000
m3 083:0: 00303 w
000\00000 9 0030000 300030
0300303,30 300802 0>3003000 003002
.00300001300030

300 000300001300030 30 30080211N3.>3 03049

80

m m N 003002 3000 A
m 0 v 003002 3000
m 3 3 003002 00309
N m c 003002 039
00003000 000000 02 00003003 00300001300030 0>3003000 003002

 

 

.00300001300030 0>3003000 30 003002 00 .N 03002
000 3 03002 0003300 000 300 000300001300030 30 300802 03 00000011m3.>3 03009

 

81

m 3 m 003002 3000 A

03 m 0 003002 3000

m o 0 003002 00309

m N 0 003002 039
00003000 000000 02 00003003 00300001300030 0>3003ox0 003002

'll'll"ll'lllII'|'||l|'|""'ll"I'I"l'lll|||
ll'llllllIII-I'llIllIll!IIll'lllIIIlllllll'lllll'l'lll'lll'l

 

.00300001300030 0>30030x0 30 003002 00 .N 03002
000 3 03002 0003300 00300001300030 300 0030032 03 00000011V3.>3 03049

82

.3 03002 003300 0030000 300 0030032 30 300802 0002110.>3 000030

om m0 00 mm om mN cm 03 03 m o

 

1! n1:

 

Si

vA

 

 

 

 

 

 

 

 

 

 

DNIGHBJ-LSVEHH HAISHTDXE JO SHLNOW

83

.N 03002 003300 0030000 300 0030032 00 300802 000211m.>3 000030

mv 00 mm om mN 0N m3 03 m 0

ii I‘I II

 

vA

 

 

 

i

 

I1
I.
J
DNIQHHJ-LSVEHH EAISH'IDXE JO SHCLNOW

84

Transitions to a Mixed Diet

 

The feeding of formula, juice, or solid food in an
infant diet previously consisting solely of breast milk is
usually termed "supplementation" (e.g., Fomon 1970). The
use of this term, however, can create confusion. Although
it implies the addition of more food to an existing level
of intake, this is not necessarily the case. While--in
terms of absolute numbers of foods consumed--the supplement
is always a dietary addition, it is not necessarily a

nutritional addition. The new food can replace part or

 

all of the previous diet; or it can, in fact, be a true
addition, increasing overall nutrient intake.

To avoid confusion, the terms "addition" and "replace-
ment" are used in the following analysis rather than "supple-
ment." If the number of breast feedings stays the same or
increases when a new food enters the diet, the food is
considered an addition. If the number of breast-feedings
decreases, that food is considered a replacement. Since
prior feeding data do not exist for Month 1, criteria for
differentiating additions from replacements have been
modified slightly. The consumption of formula as one
isolated feeding is labeled replacement. "Addition" is
considered to be the use of small amounts of formula
immediately after breastefeeding bouts.

Of the two breastefeeding attributes readily observable
in a study such as this one-~number of feedings per day and

length of feedings--the former is used here as the criterion

85
for assessing the impact of dietary change. It is used
with the recognition that it is hardly an exact estimate
of breast milk intake. The other variable, length of
feeding, varies with the enthusiasm of the infant. It
has been shown that an infant can empty a breast in 10
minutes and that the length of feeding is more an indi-
cation of the vigor and attentiveness of the infant than
the amount of milk consumed. Number of feedings per day,
on the other hand, can indicate actual differences in
amount consumed (Egli et al. 1961). Only that amount of
milk which can be stored in the mammary gland ducts will
be available, regardless of how long the infant sucks or
whether or not more time is allowed to elapse between
feedings. In fact, a reduction in milk production may
occur with decreased sucking frequency. Thus, the "replace-
ment" of a breastefeed should reflect lower total milk
intake from both loss of one or more feedings per day and
from lower milk output at the remaining feedings.

In this study, the effect of the introduction of the
first non-breast milk food on the diet appears to depend
on the infant's age at introduction. Although there is
no real distinction between the frequency of addition and
that of replacement in Month 1, replacement clearly pre-
dominates in Months 2 and 3. After that time, the majority

of new foods are additions (Table IV.15).

86

TABLE IV.15--Role of Newly Introduced Foods in Diet
Formerly Consisting Exclusively of Breast

 

 

 

 

 

Milk.
—_——--—===—_-Month _
Role of Food 1 2 3 4 5 6 Totals
Addition 7 l 2 8 15 5 38
Replacement ll 10 8 2 4 l 36
Totals 18 ll 10 10 19 6 74

A x2 test of additions and replacements at or before the
median of three months with those after is highly signifi-
cant (x2 = 21.70; p < .001).

The predominance of replacements in Months 2 and 3
is not simply a function of some overall decrease in
(feedings per day. If such a decrease were the rule, any
new food item would automatically appear to be replacing
the missing breast-feed(s). To test for the existence of
a general decrease in feedings, the number of feedings per
day in Months 1 and 2 were examined in infants who did not
experience the first introduction of non-breast milk foods
until after Month 3. Among those infants, 23 (51%) had a
decrease in number of feeds per day. Thus, about half
would have been classified as "replacing" breast milk if a
dietary change had been made. Almost the same number--22

(49%)--either increased their rate of feedings (N=9) or

87
maintained the same level (N=l3), either of which would
have made the introduction of a new food appear to be an
“addition." The difference between this distribution
and that observed among those actually experiencing
introduction of a new food in Months 2 or 3 is signifi-
cant (Table IV.16, p < .01). Thus, the predominance
of "replacement" in Months 2 and 3 is not simply an
artifact of a general reduction in feedings per day
during those months.

Whether the non-breast milk item consumed was formu-
la or solid food appeared to be a function of the infant's
age at introduction of the food rather than its action
as an addition to or replacement of existing intake.
Formula consumption predominated before four months and
solids, after (Table IV.17, p < .001).

Sex of the infant was unrelated to whether a food was
an addition or a replacement. Table IV.18 shows that
the percent of male and female infants was almost equal
for the two categories at every month.

As mentioned above, there was little difference in
the frequencies of addition and replacement of non-
breast milk food in Month 1. These categories, however,
are predictive of future time of weaning for the 18
infants involved (Table IV.19). In those 11 cases where
the new food was a replacement, lO weaned by the end of
the second month. In contrast, those 7 for whom the

new food was an addition continued breast-feeding longer.

88

.N 000 3 003002 03 000 300

00030003 30 300800 30 00300000033 0>330300800 003 00 00000 03 0303
.3003 003 03 000303000 00 .0000003303 0000 000 000 .000 0>00 03003
0003 0 0000 0303 0303030 003 30 3008000000 00 300003003 030003 00003
.30303 30 v 03002 03 0338 3000301000 30333 003 00300003303 00003 300

3
30. v 0 .30.3 u «x “3003 030000 300
00 03 33 30303
33 mm 33 30000003008 3
3 000 N 003000
30 0030 03
0003 3332
3000351202 NO
0300 30030000
mm mm 3 300333000
030303 30303 30 3 0300: m 0:0 N 003000

 

0000 0332 3000301002 30330 30 003300003303

l'll'l'-|||l|llll'l'"l"lI'|l|'l'IIIIIIIII'II'll|'l'll'lll'|llll'l"llll
i]I'lllll'iillllllIlllllll'llllllllllll'l'lll'lliillllll"lll I'll-Ill

.m 000 N 003002 003300 000 300 03000 00300001300030 30
300802 03 0033000 003310033030000 0 30 00000030 003 303 30091103.>3 03009

89

300 0300

300. v 0 “3.33 n «x 03 03002 n 003002 "3003 003002
33 03 030303
N mN MHDEHOM
mm 33 003300
003002 mA 003002 mm... 0000

.0000 x332 30003mL0oz 30330 30 003300003303
00300: 03 0>33030m 003300 000 0308300 30 0033000033031103.>3 00009

90

Adv fay AHW Amv Amy Any 2

 

mm. HH. om mm. oo.H on. w mmamemm
on Amv Adv Amy Amv Avv z

oo. cm. «a. mu. mm. om. w mmamz ucmsmomammm
Amy Amv Avv AHV on Amv 2

no. mm. om. 5H. 00. om. w mmameom

Amy Amy Amy AHV AHV Avv z
oo.a om. gm. mm. 5H. cm. W mmamz coflufinc<
m m e m m H

nucoz

.COwuosoouucH mo nucoz can xmm ha .umfln unamcH :a moon 3mz mo mHOMLImH.>H mqmda

91

TABLE IV.19--Month of Weaning for Infants Given Formula
"or Solids in Month 1, by Addition and
Replacement Categories.

 

' MOnth of Weaning‘

 

l 2 3 4 5 6 Totals
Addition - l 4 l - l 7
Replacement 7 3 - - - 1 ll
Totals 7 4 4 l 0 2 18

Median Test: Median = Month 3; Fisher's Exact Probability
= .002'

None weaned until the second month; the majority weaned inf

Month 3. Using Fisher's Exact Test to compare frequency

of weaning at or before the median of 2 months with that

after 2 months, the difference is highly significant

(p = .002).

Major Dietary Patterns

 

The interaction of dietary components, maternal infant
feeding behaviors, and certain demographics as discussed
above facilitates the identification of major dietary
patterns for these infants. These patterns permit the
breakdown of this heterogeneous sample into relatively
homogeneous groups to be used in a number of analyses pree
sented later in this chapter. This section will present
the rationale for group differentiation and describe the

groups so formed.

92

The primary basis for defining the most common
dietary patterns is length of exclusive breastffeeding.
Whether or not causal relationships exist (and regardless
of the direction of any causality), length of exclusive
breast-feeding has been found in this study to be con-
cordant with variables from a wide variety of domains.
These include parental socio-economic status, details of
breast-feeding behaviors, and the impact of introduction
of non-breast milk foods on the overall diet.

Using length of exclusive breast-feeding as the
criterion, there are three main groups of infants (Figure
IV.9):

Group 1: 4 weeks or less exclusive breast-
feeding. N=18

Group 2: more than 35 days but less than 94
days exclusive breast-feeding. N=21

Group 3: more than 101 days exclusive breast-
feeding. N=4l

Group l--Most of the infants in Group 1 were exclu-
sively breast-fed for a very limited time--shorter than
the four weeks used to define the group. However, because
of differences in data collection between Studies I and II
and because of the necessity of drawing on various sources
of information to reconstruct the perinatal period, four
weeks is used here as a conservative estimate of the end
of exclusive breast-feeding. Mothers and infants in
Study I were first seen at home only one week after

parturition. At that time most were newly discharged

93

Group 1 Group 2 Group 3
N=19* l

O 20 40 60 80 100 120 140 160 180 >180

 

 

Days Postpartum

FIGURE IV.9--Length of Exclusive Breast-Feeding and
Infant Feeding Groups.

*Because the length of exclusive breast-feeding in this
group was determined by a combination of methods, its
duration as reported on diet diaries is only approximate
and reflects the timing of the early diet diary rather
than actual feeding behavior.

94

from the hospital and had not had time to settle into
any routine. Some cases of mixed formula and breast milk
feeding were identified at that point, but most were still
determined to establish exclusive breast-feeding. Although
diet diaries were sent to these women at approximately
ten days postpartum, some were not completed and returned
until two weeks later. When mixed feeding was reported
at that point, additional details on the events of that
first month were obtained retrospectively at the one month
measuring session and even later. The latest date for
modification of the breast milk diet for any case is in
the fourth week. Had data collection been more precise,
the four week cut off could probably have been made
earlier. Thus the fact that these infants had been fed
breast milk exclusively for a very short time came from
the first diet diary. The date of this diary was used
unless interview data could pinpoint the date more accur-
ately.

Because of the difficulties in obtaining early feed-
ing data in Study I, the first at-home contact in Study II
was at two weeks postpartum. In all cases but one, the
infants in Study II in "Group 1" were receiving mixed
diets or had already been weaned by the second week visit.
In the remaining case, a diet diary at 20 days listed
formula and included a note explaining that exclusive
breast-feeding had ended two days earlier. In addition

to limited exclusive breast-feeding, Group 1 is

95
characterized by a high frequency (89%) of the use of
formula rather than solids within the first month. As
discussed above, there is no clear preference for
addition over replacement or vice versa. However, since
each of these categories correlates with later differences
in time of weaning, two subgroups will be identified for
some of the later growth analyses. These subgroups are:
la less than four weeks exclusive breast-
§:$ding. Introduced food is an "addition."
lb less than four weeks exclusive breast-

feeding. Introduced food is a "replace-

ment." N=ll
In Group la the average amount of formula consumed per
feeding in Month 1 was 2.2 i 1.1 ounces. The total daily
consumption ranged from 2 to 20 ounces, with a mean of
10.8 i 7.6 ounces. In comparison, the average formula
consumed per feeding by infants in Group lb was 4.0 i 1.4
ounces. The daily total ranged from 13.5 to 39 ounces,
with a mean of 24.5 i 8.0 ounces.

Group 2--This group consists of infants who had at
least 35 days of exclusive breast-feeding. This is in
contrast to Group 1 in which exclusive breast-feeding
never became established as a feeding pattern. The upper
limit of 94 days reflects both the end of this relatively
long period in which exclusive breast-feeding was termif
nated at scattered ages and the upper limit of the

clustering of "replacements" for breast milk by formula

or solid food. Within this group two subgroups will be

96
distinguished in later analyses, based on the first food
introduced. These subgroups are:
2a exclusive breast-feeding of more than 35
and less than 94 days with formula the
first nonfbreast milk food introduced. N=9
2b exclusive breast-feeding of more than 35
and less than 94 days with solids the
first non-breast milk food introduced. N=12
In Group 2a the quantity of formula first introduced per
day varied from 6 to 15 ounces, with a mean of 10.9 i 3.2
ounces. In Group 2b, the most commonly consumed first
solid was cereal mixed with formula, cows milk, apple
juice, or water. When the nutrient content of each mix-
ture is computed and compared with that of the average
intake of formula in Group 2a, the nutritional contribution
of solids or formula to the diet is markedly different
(Table IV.20).
TABLE IV.20--Nutrient Values for Average Daily Intake of

’Non-Breast Milk Foods at their First
Introduction; Mean 1 Standard Deviation.

 

Infant Feeding Group (N)

 

Nutrient 2a (9) 2b (12)

Energy (kcal) 213.9 : 63.5 66.9 i 51.8
Protein (gm) 4.8 i 1.4 2.0 i 2.3
Fat (gm) 11.5 i 3.4 1.5 i 1.7
Carbohydrate (gm) 22.8 + 6.8 11.2 i 7.8

Source: Nutrient values used in Tables IV.21 and IV.22 for
commercial formulas come from Michigan Dept. of
Public Health (1980). Values for other foods
come from Gerber Products Company (1977).

97

Group 3f-These infants were exclusively breastefed
for at least 102 days. After that time the first none
breast milk food was introduced to 35 infants. One infant
first consumed formula; 34 first consumed solids or fruit
juice. In 28 cases, the new food was a dietary addition.
In only 7 cases was it a replacement. Estimated nutrient
values per day are shown in Table IV.21.
TABLE IV.21--Nutrient Values for Average Daily Intake of

Non-Breast Milk Foods at their First

Introduction for Feeding Group 3; Mean
:_Standard Deviation.

 

 

Role of Food in Diet

Replacement (N=7) Addition (N=28)
Energy (kcal) 101.6 : 77.0 47.8 i 71.1
Protein (gm) 3.3 i 1.8 1.2 i 1.6
Fat (gm) 3.8 i 4.1 1.9 i 4.2
Carbohydrate (gm) 13.7 i 8.3 6.7 i 8.4

Four infants did not consume any non-breast milk
food before the age of six months. Two others who had
consumed only breast milk left the study during the fourth
month. They are also considered part of Group 3, bringing

the total number of infants in the group to 41.

Infant Growth
This section presents the results of analysis of
infant growth data. After describing the growth of the

sample as a whole and the preliminary transformations of

98
the data, infant growth will be analyzed in terms of four
growth-related dietary parameters: length of breast-
feeding, age at first introduction of nonfbreast milk
food, the common dietary patterns identified above, and

details of the breast-feeding process.

Description of the Sample

 

At birth the 80 infants studied all weighed more
than 2500 grams, the traditional anthropometric boundary
between high risk and normality. From birth to six months
of age, males were heavier and longer than females (Table
IV.22). Females had higher summed fatfold measures at
birth and one, five, and six months. Males, in contrast,
were leaner at birth, had a somewhat faster increase in
adiposity during the first months of life, and then
actually decreased in fatness during the later months of
the study (Table IV.22).

When compared with the National Center for Health
Statistics percentiles for infant growth (National Center
for Health Statistics 1976), the sample was consistently
heavier and longer for age than the NCHS norms (Table
IV.23). Males had higher mean percentile rankings than
did females, though these differences were not statisti-
cally significant. In weight for height, the sample more
closely approximated the NCHS standards, though infants
were, on the average, considerably thinner than the norm

at one and six months.

m .HHHO 23 m .NHHo .Hw bJHHm .mm

99

 

o.m H>.mm o.m Hm.Hm k.> Hh.~v H.m Ha.m~ mmamsmm
m.mHHm.mm c.4HHa.mm m.mHHo.om H.HHHm.am a.oaun.mm 5.5 Hm.av 0.0 Hm.m~ mmamz
Assv wagon
neat qaaoa
a.~ Hm.am S.~ Hm.mm o.~ Hm.mm v.~ HH.HG e.~ HH.mm m.~ Hm.vm o.~ Hm.om mwamsmm
m.~ H~.os G.~ Hm.mm m.~ Ho.mo ~.~ H~.mm o.m Hm.wm o.~ Ha.om v.~ HH.Nm mmamz
have meozmq
92mmzsomm
e.o Hq.k e.o Ho.a e.o Hm.o o.o Hm.m m.o Hm.m m.o Hm.q v.0 Hv.m mmamsmm
o.a Hm.m m.o Hm.e m.o Hm.a m.o Ho.m A.o Hm.m A.o “v.4 m.o Hm.m mmamz
xmxv emonz
o m a m N H auufim wanmfium>
sumo:

 

 

.coflumwbmo onmpcmum H com:
.cucoz comm mo can an tam nuuflm um mucmmcH now mowumwumum o>flumfluommollmm.>H Manda

 

100

.HsmmCAcmmE on on
A.o.c .mnmpaouv Emumoum mmuz an pmumasoamo on panoo mmHHucmouom camcmq MOM unmwmz 30m 008«

e.m~HH.mv

m.a~Hm.m¢

 

¢.omHH.ma m.m~Hm.mm m.m~H«.mm .nun .uuu mmamemm
H.4NHv.km m.m~Ho.qa m.m~HH.mv ”.mwfim.~m a.a~HH.mm k.m~H~.mq .-u- mmamz
:aozmq
mom amonz
m.a~H~.HG o.mmuo.mm ¢.mmuo.mo a.m~Hm.~o m.mmum.¢o «.mmfia.~m m.¢~Hv.HG mmamamm
m.omum.ma o.m~Hv.ok m.~mum.oe «.mmflm.mm m.-Hm.wo o.m~Hm.vm o.o~Hm.mo mmamz
mwfi mom
meozmq
m.¢~Hm.vm ~.4~Hm.om A.H~Hm.mo a.HNHH.mo m.-Hm.mm m.qNHe.mo m.m~Ha.km mmamemm
m.m~H>.om m.¢mum.mm «.mmua.mm a.vmum.mm A.H~HH.oa m.mmHm.Hm k.amuk.om mmamz
moa mom
amonz
G m a m N H zuuflm manmflum>
sumo:

 

III|I|I|"III|I||Il
I‘IIIIIIIIIIII

 

I.au:oz comm mo cam am
can cuuflm um mucmmcH How coflumw>mo oumocmum + maflucooumm mmoz cmmzllmm.>H mam¢9

lOl

Preliminary Data Transformations

 

In order to combine males and females into a single
sample, several data adjustments were made preliminary
to actual data analysis. Weights and lengths for each
infant were transformed into three new variables based on
their percentage of the NCHS sex-specific medians for
weight for age, length for age, and weight for length.
These transformations were performed by the computer
center of Fels Research Institute under the supervision
of Dr. Alex Roche. The growth data analysis program of
Goldsby (n.d.) was used. Since the NCHS growth curves
for infants from birth to 36 months are themselves com-
puted from the Fels data base, this analysis provided
exact and accurate transformations of the data. More
readily available representations of the NCHS norms--
such as those published by Ross Laboratories (1976) and
widely used in clinical settings--were not used. For
these, the extreme percentiles have been collapsed,
giving the impression of a narrower distribution of in-
fant size than is actually the case.

All results presented below are based on these
transformations. "Percent of the median" is used rather
than percentiles because the latter cannot be calculated
for many of the low infant weights and lengths common
at birth and one month of age.

National standards were not available for infant

fatfolds. Therefore, within-sample adjustments were made

102

to take account of sex differences. First, all eight
fatfolds for each infant were totaled for each measure-
ment session. Then male and female sex-specific least
squares regression lines were computed for the entire
sample, using total fatfolds in milimeters as the depen-
dent variable. Residuals were then computed for each
infant measurement entered in the equation. These re-
siduals represent the difference between the actual
infant measurement and that predicted by the regression
equation. They are used as the new data points in all
analyses in this chapter. Total fatfolds were used rather
than any single fatfold for two reasons. First, since
infants are frequently somewhat assymetrical at birth
and for variable periods of time thereafter, using both
left and right side measures prevents the confounding of
results by biases resulting from the arbitrary choice of
either left or right side. Second, the combination of
trunk and extremity measures controls for difference in
fat patterning.
Weaning

Length of breastefeeding was most significantly
related to infant growth in the first several months of
life. Table IV.24 presents means and standard deviations
for the percent of median for all growth variables. There
were no significant differences between groups assigned
by month of weaning in length for age or weight for length

at any time from birth to six months of age.

103

m.m Hm.m0H

m.aHHo.mOH

m.aHH~.mm

 

Aka.v om. o.OHH¢.HOH $.4HHm.GOH o

1mm.v mm.H H.OHHm.o0H m.mHHm.mOH H.m Hm.mm m.m Hm.NCH m.aHHm.AOH m

xa~.v os.H H.HHHo.mOH o.mHHm.moH m.m “n.00H m.m Hm.MOH m.~HHm.OHH a

Hmo.v mH.~ A.OHHq.OHH m.mHHH.OHH G.OHHm.mm a.m Hm.mOH m.~HHm.mOH m

xao.o km.~ m.HHHq.HHH q.MHHv.~HH o.m Ho.mm m.a Hq.aOH a.mHHN.HHH m

Hao.v ms.~ m.MHHq.mOH H.HHH~.mOH v.HHHm.mm ~.oHHH.MOH «.mHHa.moH H
AHG.V mm. 8.4HHB.GOH q.HHHm.VOH N.¢HHm.oOH H.~HHo.mm ~.~Hwo.mOH nuuHm
32...... ........... m----mmwma
Awm<v

.mGHGMDZ m0 sucoz an :w ou H mnucoz mo mam can
cuuflm um mousmmoz :uBouw ucmmcH usom mo mCOHDMH>mQ oumpcmum + mammzulvm.>H mqmda

h,”

 

104

 

Go.H e.m+m.NOH m.q+m.mOH H.m+m.HoH m.~+o.NOH H.m+m.NOH m

am.H m.v+m.NOH a.m+m.m0H o.m+m.OOH m.~+m.ooH m.q+¢.HOH m

mm. m.a+o.MOH m.m+o.vOH H.m+q.HOH m.H+m.HOH m.v+m.NOH 4

mm. ~.¢+m.~0H a.m+m.mOH h.~+m.oOH H.~+¢.~¢H m.a+m.HoH m

hk.H m.v+m.NOH m.v+m.vOH v.m+o.mm m.m+m.NOH m.¢+v.NOH m

kH.H a.a+m.NOH m.v+q.~0H 4.4+H.ma m.~+a.HOH m.¢+m.HOH H
mm. 8.4+S.NOH m.m+q.aOH m.m+H.ocH o.m+~.HoH ~.~+~.~0H suuHm
.......... ..Z..~.Z.a
09m

Acchmz mmoz we we mm< How a

 

umcmHJnmcwcmmz mo cucoz

 

 

cmscHuaoou+«~.>H mHmaa

105

 

 

 

Am¢.v mm. q.a Hu.>m e.m Hm.vm m.vHo.mm H.mHa.mm ~.m HA.OOH w

Hmm.v no. ~.m Hm.mm m.m HH.mm «.mum.mm ~.¢Ho.mm m.v HA.NOH m

Aom.c am. «.mHHm.o0H a.OHHo.ma e.mHm.km m.mHm.mm m.m Hm.NOH v

Am~.1 Gm.H e.m H~.NOH H.owm.mm m.mHm.mm m.mHv.mm H.m Hm.vOH m

Aom.v mm. b.4HHm.mOH m.mHm.mm o.mH~.mOH m.mHH.mm m.¢HH~.voH N

Ane.v m4. o.HHHH.HOH v.an.mm «.mum.mm «.mHm.mm o.o Ho.OOH H
Aom.v m~.H ~.m Hm.mm m.>HH.mm «.mflo.hm H.5Hm.Hm m.OHH~.va nuHHm
4mm ...... .....m... ........... m----mmmma
mad

.Acmwcmz mmoz mo wv sumcmq How unmwmz1tmcficmm3 mo sucoz..

 

 

 

 

 

omscHucoouu¢~.>H mHmae

106

 

av. m.mHHH.vu o.m~Hu.m- o.OHHG.0Hu m.mHHm.mn m.vHHm.ku m
mm. m.hHHm.H 0.5Huo.au m.a Hm.~- G.HmHm. G.AHHG.- m
Hm. ~.mHHa.a m.GHHH.m H.mHG.H a.o~Hm.m e.mHHm.m a
on. a.OHHm.a v.4HHH.m o.m HH.m H.~HHm.m >.mHH¢.HH m
on. m.m Hm.m m.GHHH.m ~.m Hv.H m.m HH.m m.HHHH.m m
ea.~ m.a Hm. H.m H~.vn a.“ He.mu m.v H~.H s.m Ho.o H
mo.H m.m Hm.au «.8 Hm.HH- e.v Hm.nu H.m Hm.mu ~.m H~.mu nuuHm

........... ..:.HNH2§5
). mod

Ass memschmmv moHomunm Hmuoasnmchmmz mo chaos

 

 

 

 

noscflucOOIlvm.>H mumma

107

Significant differences relating to future weaning did
appear in weight for age at one and two months (p < .05).
Reduced weight for age at one and two months was found
among infants who would be weaned in the third and, to a
lesser degree, the second months. In Month 1, the highest
weight for age was of infants who would not be weaned
until after six months of age, a continuation of the trend
apparent at birth. By two months of age this group had
been equalled in size by other groups of infants.

Analysis of total fatfolds for age suggests that the
differences between groups in weight do not necessarily
reflect similar differences in fatness. At one month of
age infants who would be weaned in the third month were
the leanest infants (p < .05) and, of all weaning groups,
had experienced the least fat increase from birth.
Although those who would be weaned after six months did
show a marked initial increase in fatness, the fattest
infants were those who had been weaned during the first
month. Furthermore, this group of infants had experienced

the largest increment in total fatfolds from birth.

Introduction of First NoneBreast Milk Food

Age at introduction of first non-breast milk food
was more closely associated with weight for age than with
any other growth variable (Table IV.25). Figure IV.10
shows that those infants exclusively breast-fed for four
or more months were on the average heavier at birth than

those exclusively breast-fed for shorter periods of time.

 

108
Furthermore, this higher mean weight continued through
five months of age. Those infants eXclusively breast-
fed the shortest period of time were lightest at birth
and, in fact, experienced an initial decrease in percent
of the median weight for age. Differences between groups
in weight for age approach statistical significance at
one month (p < .06) and two months of age (p < .08)
(Table IV.25).

Infants exclusively breast-fed longest had a greater
mean weight for length from birth to four months (Figure
IV.11) and were slightly fatter than other infants in
the early months of life, but these relationships are

not statistically significant.

Common Infant Dietary Patterns

Infant growth variables were also analyzed by the
major dietary patterns identified earlier in this chapter
during analysis of the infant feeding data. Table IV.26
shows the average monthly percent of median for each growth
measure by dietary pattern group. Although these distri-
butions overlap, trends in weight for age (see Figure
IV.12) are consistent with those observed among groups
defined by timing of weaning and introduction of first
non-breast milk food. Infants in dietary Group 3e-those
with longest exclusive breast-feeding-—were heavier at
birth, and maintained this size superiority until five
months. Group la-fexperiencing the "addition" of formula

by one month of age-~had a marked decrease in weight for

109

 

 

 

 

1N4.c NN. m.m N.NOH a.oH H N.OOH m.mH H o.mOH o

ANN.V mN.H N.m H N.NoH H.HH H N.NOH N.NH H N.GOH m

ANN.V em.H m.OH H o.mpH N.oH H H.40H o.NH H m.NOH a

smo.c om.N H.¢H H N.HHH m.NH H N.mOH N.HH H m.GOH m

ANH.V NN.H N.HH H o.NHH N.4H H N.NOH H.mH H m.ooH N

soo.v Nm.N N.NH H N.NOH o.qH H m.moH a.mH H N.¢¢H H
Hoo.v Nm. v.4H H v.ooH o.mH H o.VOH m.mH H o.NOH suuHm
..:--.....1mmw ...... Mi---..---w-m..-..--::.::..-:m ............ .H.-------------mmw..mm..s
mod

HmmHoo: mmoz mo NC was How usoHoZnooom sHH: ummoumnmoz umuHm_uo sumo:

.mooom :HH: umomum
Isoz mo sofluoscouucH umuwm mo nusoz an .m on a mcucoz mo cam can

suuflm um mousmmmz suaouw acumcH Hsom no mcofluMH>oo oumocmum + msmozllmm.>H mqm<a

1mm.v mo. m.m a.NOH N.m H N.N¢H N.v H o.NOH o
sov.o Na. N.¢ H H.NOH N.a H o.NoH o.a H o.HOH m
m smo.c o4. m.q H H.NOH m.m_H o.NOH o.m.u N.N¢H a
smm.v so. N.¢ H m.NOH m.m H N.N¢H N.m H N.HoH m
sqm.o mo. N.v H o.m0H o.¢ H N.HoH o.v H o.NOH N
sam.v oH.H m.v H m.NOH N.q H.N.HOH m.¢ H H.HOH H
sHm.v OH. v.v H m.NOH m.q H H.N¢H N.v H o.NOH suuHm
---..-Mmm ...... shimulusa
09m

 

smoHoozhmmoz uo NV mom sou summoHnmoom.sHH:.ummoumumoz umuHm Ho sumo:

 

 

 

GDDGHHCOUIIm N .>H mquwB

111

ANH.V HN.H o.m H o.Nm N.N H m.mm m.o H m.mm o
soH.V om.H N.N H N.Nm N.m H «.mm o.o H m.OOH m
smo.o mm. q.NH H m.OOH N.m H N.Nm o.m H H.ooH a
ANN.C mm.H o.OH H H.m0H o.N H m.mm N.N H m.OOH N
:HN.V mm. a.mH H o.moH N.OH H H.N¢H o.HH H m.HoH N
sam.v om. o.oH H ¢.HOH o.OH H H.¢OH 4.N H o.mm H
AoN.V om. m.m H o.mm N.m H N.¢m o.N H N.mm suuHm
sonc m a.“ N H Amsumo:c
mod

 

 

 

ooosflucoounm~.>H mamde

112

AHN.V HN. m.mH H N.vu o.NH H o.Nu N.vH H H.ou o

Hom.v oH. m.NH H o.H H.HH H N.H- m.oH H o. m

HNM.V Ho.H m.mH H m.¢ o.OH H m. N.NH H m.N a

.mH.v ON.H N.m H_o.m m.OH H N.H m.mH H o.m m

10H.c NN.N m.N H m.o m.m H m.H m.HH H H.o N

sHm.v NH.H H.N H H.H a.» H N.Nu N.m H m. H
HNG.V Hm. o.m H m.Nn N.m H o.N- N.v H ¢.ou sumHm
-------.1mu.m ...... m...:~mfimss
mad

 

 

 

 

oommHumoouumN.>H mHmme

113

 

 

110 1
c; 108 -4
ma:
H m
B4106 ,
21H
«.8
O 104 4
JJ
4):
5.9102 J
o m
3'43
o
m 100 ‘
Birth 1 2 3 4 5 6 months

FIGURE IV.10-~Mean Weight for Age (Percent of NCHS Median)
by Month of Introduction of First Non-Breast
Milk Food. Month 1 (0—0) , Months 2 and
3 (l———-—l), Month 4 or Later.(A—————A).

Percent of Median
Weight for Length
H +4
\D \D \O O O
A ox (I) O N

\O
N

 

 

I T v I U

Birth 1. 2 3 4 5 6 months

FIGURE IV.ll--Mean Weight for Length (Percent of NCHS Median)
by Month of Introduction of First Non-Breast
Milk Food. Month 1 (o—————O), Months 2 and
3 (Ik————I), Month 4 or Later (Ar————A).

114

cm.
m5.
mm.
Hm.a
VH.H
mh.H

Hv.

m.m Hm.NOH
m.m HN.N0H
m.OHHo.m0H
H.OHHN.HHH
N.HHHo.NHH
m.NHHN.m0H

H.HHHH.GOH

H.NHHN.OOH
o.NHHH.HOH
m.m HN.NOH
H.HHHo.mOH
N.mHHm.NOH
H.HHHm.o0H

m.HHHh.moa

h.m H«.Hoa
m.oaflw.moa
o.NHHm.mOH
v.maHo.moa
m.wHHm.boa
H.mHHo.voH

h.vHHm.HOH

o.maflw.moa
m.NHHm.¢oa
m.HHHm.boa
m.oaflm.hoa
h.NHHm.moa
m.NHHm.MOH

m.HHHm.HOH

o.mHHH.hoa
m.maflfi.moa
m.maHo.moH
h.mafim.voa
h.maHH.moa
m.mHHm.mm

¢.oHH¢.HOH

 

 

mad

 

. ACMHomS mmuz mo wv mad HON unmamzllmmsonw.mcHomoh ucmmcH.

 

 

.mmsouo msHoomh “camcH an hm OH H

mSUCOE MO UCW UCM Cub—”Hm Hm mmHflmmmz SH3OHU HEMHCH .HSOh HO .Q.m + WCMQZIImN.>H ”NJ—”g9

115

 

1H4.c Ho.H m.mHv.NoH «.mHm.mOH H.NHm.HOH o.¢Hm.HOH o.¢HH.HOH o
14H.V om.H N.qHH.NOH o.¢Hm.HOH a.mHN.o0H m.mHN.OOH o.vHN.NOH m
ism-V mo.H m.aHH.NOH N.NHH.MOH N.NH¢.OOH m.mHo.NOH m.NHm.HOH .
Ham-o HN. N.vHN.NOH m.mHH.mOH N.mHo.HOH N.NHmHHOH o.mHN.NOH m
sea-v Ne. N.4Ho.moH o.qu.NoH N.¢HN.OOH m.¢HH.NoH m.mHm.HoH N
mks-c mm. H.HHN.NOH N.mHm.NOH N.HH4.HOH m.mHN.HoH «.mHo.HOH H
ANN.V m.- ¢.va.NOH o.mHH.MOH H.mHN.OOH o.mHo.HOH m.mHm.NOH sumHm
3.2 :.-------. ............ m ........... m---:------m ----------- m----mm.-.....a
mod

 

 

concHunoollmm.>H mam<a

116

Hh.H
5m.m
mm.H
ww.H
mv.
mm.

mm.

o.m Hm.Nm
N.m HN.Nm
v.NHHm.OOH
o.OHHH.mOH
o.NHHH.moH
o.OHHH.HOH

m.m Ho.mm

m.N Hm.om
o.o Ho.om
N.N Hm.mm
N.o Ha.om
H.HHH¢.HoH
N.¢HHN.OOH

H.N Hm.mm

m.mHo.Nm
o.NHo.NOH
o.NHo.voH
o.oHo.NOH
m.NH¢.mOH
N.mHN.mm

m.mHm.mm

N.N Hm.mm
m.o Hm.HOH
m.N Hv.HOH
o.N HN.NOH
N.NHHo.NoH
o.N.HH.OOH

N.N Hm.mm

N.¢Ho.oa o
H.NHm.mm m
m.mHN.Nm v
N.NHN.NN m
N.mHN.Na N
N.oHo.mm H

o.oHo.mm suuHm

054

 

 

 

 

U¢5CHU£OOII¢N.>H mnmflfi

117

smm.v am. N.MHHN.HI N.NNHo.qu N.NHHN.OH- H.oHHo.Nn H.mHHm.m- o

mom-V SH. m.NHHH.H N.NHHH.N- o.OHHo.H- o.oNHm.H- H.NHHo.m m
ANN-V Nm. m.mHHm.v o.mHHN. m.N HN. m.NHHo.N N.mHH¢.m a
svm.c «H.H N.m Ho.m m.HHHv.H- ¢.oHHN.m o.mHHN.m o.HHHH.o m
ANN.V NH.H m.N Hm.o H.HHHm.u N.N HN.m m.m HN.N H.HHHH.4 N
Ame-V HN.N H.N HH.H m.N HN.N- H.m Hm.- m.N HN.N N.N HN.v- H
AHN.V mm. N.m Hm.Nu N.o HN.N: N.v Ho.N- o.m Ho.m- v.m HN.N- suuHm
...m----.m-::-:-m ............ mm ........... m .......... m ------------ m---m.mmm..s-
omd

 

 

 

oommHumoonnmN.>H mHmae

112
110

o 108

(31

d

H

O

(4.4

4, 106

I:

U‘

H

0.)

3

5 104

H

"U

(D

S

5H

O

u 102

C:

(D

U

M

0)

°‘ 100
98
96

118

Group 1a

~ Group 3
Group 1b

Group 2a

‘I Group 2b

 

 

 

r t f r Y j ‘

Birth 1 2 3 4 5 6 months

FIGURE IV.12-—Mean Weight for Age (Percent of NCHS Median)

for Infant Feeding Groups la through 3.

119
age at one month, but recovered steadily and, in fact,
was the heaviest group of infants at Months 5 and 6.
Group 2b, which had breast-feeding partially replaced by
solids during Months 2 and 3, is notable for its rapid
decrease in weight for age after two months. Although
these infants were, on the average, heavier at birth
than all but Group 3, after three months they were the
lightest.

A comparison of mean total fatfolds in the different
dietary groups reveals a difference approaching signifi-
cance at one month (p < .08). This reflects the extreme
leanness of Group la as opposed to the relatively extreme
adiposity of Group 1b.

At birth there were no significant differences
among the dietary pattern groups in percent of the median
weight for length (Table IV.26). Changes in later months
indicate possible dietary effects. Group la was virtually
unchanged until two months, while Group lb increased
substantially in weight for length during the first month
(Figure IV.13). Groups 2a, 2b, and 3 were similar to
each other in weight for length growth pattern from
birth to two months of age (Figure IV.14). After that
time all three groups decreased in weight for length.
H0weVer, the decrease in Group 2b was particularly rapid,
resulting in a significant difference among feeding groups

at Month 5 (Table IV.26; p ‘<.01).

120

Since Groups 2a and 2b are in fact parts of a larger
dietary group subdivided on the basis of first nonébreast
milk food introduced, the marked difference in pattern
of change was further explored. A least squares re-
‘gression line was calculated for each dietary group,
using weight for length measures at Months 1 through 6
as the dependent, and age in months as the independent
variable (Table IV.27). The slope--rate of decrease in
percent of median weight for length--was over twice as
great among those for whom breast milk was replaced by
solids (Group 2b) as among those for whom breast milk
was replaced by formula (Group 2a).

Shifting the analysis to concentrate on the growth
extremes, both weight for age and weight for length were
in some way significantly related to the dietary groups.
For these analyses, infants were ranked from largest
to smallest for each growth measure at each month.
Rankings were then divided into quartiles.

Although the heaviest and lightest infants were
randomly distributed among the feeding groups in the first
few months of life, a significant pattern emerges by
Month 3 (Table IV.28). The frequency of lowest weight
for age quartile infants in Group 3 declined and that in
the upper quartile increased somewhat. At the same time
the weight for age of infants in the other groups was
skewed toward lighter infants in the later months. These

differences are significant (p < .05) at three, four, and

121

 

104 ‘
/.~ “‘
I '.‘
102 ~ / “mu-«4.
100 -
98 5
96 ‘
94 1

 

 

 

Birth 1. 2 3 4 5 6 months

FIGURE IV.13--Mean Weight for Length (Percent of NCHS
Median) for Infant Feeding Groups la
(0—0) and lb (0 ----- O).

106 4

1.04 1
102 5

100 ‘

96 <
94 4

92 1

 

9O ‘

 

W i l ' V T r

Birth 1 2 3 4 5 6 months

FIGURE IV.14--Mean Weight for Length (Percent of NCHS
Median) for Infant Feeding Groups 2a
(l-——-|), 2b (I ----- I), and 3 (L———A).

122

NG Nv.u mmHoo: uo N u Hsumo:c Hmm.Nc - No.mOH sN omouo

om NN.- mmHoo: uo u u :sumo:c Ham-o1 - oq.mOH mN omomo

mandamusmmmz H coHumsvm conmonom macho mchmom
mo Honssz

 

 

 

 

.H.OE m I HV 004 no nusoz mH poMHHm> ucmocomoocH new
numeog How uanmz mH mHQMHHm> ucoocommn umcoHumsvm :onmmHmmmlnn~.>H mqmde

123

.nm :msounu MH mmsouw
posHano Scum m macho mo mcoHuanHuch CH oozmquMHo How umou mumswm H£U««
.mHHuHmsq nomad .uanH «oHHuHmsv umo3oH CH Hogan: moHMOHocH puma so msam>m

 

 

mo. mo. mo. DOCMOHMHcmHm
HH.H. oq.m Hm.¢ «.ouomom Hso
HH m NH m HH 4 OH H m m NH N NH N m
N a N m H m m o m m m N N N sN
H m N N H a N H m H H N N N mN
a m H a N m N N N m N v N m sH
v N m N N N H N H m H m N N oH
G m H m N H sumHm omomo
mcHommm
.mazo:
.msouw

omHooom :s .sumo: somo uo omm omm sumHm um .omm uou usoHoz Ho
moHHuHmso ummsmHm new umm3oq CH mucmmsH mo HmnESZ mo mCOmHHMQEOUIImm.>H mqm¢9

124
five months of age.

In weight for length (Table IV.29), it is also
<3bvious that distribution of infants into quartiles was
non-random with respect to diet groups. However, the
jpattern of variation between groups was different from
that for weight for age discussed above. Groups lb and
2a, both based on dietary patterns "replacing" breast
Inilk with formula, tended to have a high proportion of
infants in the highest weight for length quartile, while
Groups la and 2b, which did not replace breast milk with
formula, had a high proportion of infants in the lowest
quartile. These differences were significant in Months

3 through 5 (Table IV.30, p < .05).

.Qetails of Breast Feeding

Because early growth appeared to be predictive of

lboth subsequent growth and diet patterns, two aspects

<3f the association of early growth and the breast-feeding
1?rocess were explored in greater detail. These were
llength of individual breast-feedings and number of feed-
iJags per day. To minimize the confounding effects of
crther food in the diet, only those infants who were ex-
c=llusively breast-fed for at least two months were included
(bh=47). Diet diaries at four weeks were examined and the
rnxmber and mean length of feedings on this data noted
for each infant. Infants were then assigned to two

groups according to whether they had long or short and

numerous or few feedings on this date. The medians were

12S

.mHHuHmsw Moan: .quHH «mHHuHMCU umo3oH CH HonECC moumoHUCH uon Co mCHm>u

 

 

HH m HH N NH 0H mH m CH m NH 0H NH m m
H m o m o v o m m v N N m v QN
N N H H m o N H m o N m N N mN
m H m o m H m H v N N N N m QH
H H N o o N H m H m o N H H MH
0 m H m N H suuHm omouo
mCHpomm
HSUCOZ
.msouw

omHooom Ns .sumo: sumo uo mmm omm suuHm um .summoH uoH usoHoz Ho
mmHHuumso umoanm oCm umoSOH CH mqumCH mo HmQECz mo mComHHmmEoollmN.>H mHmme

126

hvo. voo- mNo. muHHHnmnoum
uomxm m.HmCmHm

 

 

 

H H G H m N mN omm sH monouo
N G o G H N sN mmm oH momoHo
ummCmHC umm3oH HmoanC ummon umoCmHC HmoBOH
masono mCHoomm
mHHuHqu mHHuHMCU oHHuHqu
m CHCoz v CHCoS m CuCoz

.AQN UCm MHV MHCEHom CuH3

:mCHUMHmmmm Boz mmosu UCo AmN UCm nHv.oHCEHom Cqu xHHz ummmum

mmCHomHmmms mmCOHO Cmm3umn mmHHuHMCO CumCmH How quHoB umGBOH on
umosmHm CH mqumCH mo HonECz CH mOCmHomeo mo umos uomxm m.HmCmHmllom.>H mHmda

127
used as the upper limits of the "few" feedings and "short"
feedings categories.

Table IV.31 shows the results of t-tests for differ—
ences between the groups based on breast-feeding details
in size at birth and one month of age. An inverse re-
lationship existed between size at birth and length of
nursing bouts. Infants who had short breast-feeds had
greater weight, length, and total fatfolds for age than
did infants with lengthy breast-feeds. At one month of
age, infant size was more strongly associated with the
number of feedings per day, than their length. This
relationship of size and number of feedings was direct.
Infants who had a high number of feedings per day had
greater weight for length and were significantly fatter
(p .01) than infants nursing a relatively few times per
day. Growth velocity in the first month also appeared to
be associated with number of feedings per day. Those
infants with a high number of feedings had substantially
greater increases in weight for age, weight for length,
and total fatfolds than did those infants nursing relatively

fewer times per day.

Interview Data

 

A variety of data obtained during pre- and postnatal
interviews is compiled here. Portions of it come only
from Study II in which the interview portion of the study

was expanded and intensified.

128

smm.i HG. HGH.G Gm.H ANN.G mN. HGH.G Gm.H soommoHHHmon. u

G.G H N.Nu o.G H m.HoH m.m H m.Gm N.mH H H.NGH momHmoom mmouosmz

H.m H m-G- o.m H G.moH m.N H H.mm G.HH H o.moH mmmHooomuummoum zom

AGO-G Go.N- HNo.G GG.H- ism-G NG.- sGo.G oo.N- roommoHHHmmHmG u

G.H H N.oH- G.m H G.HoH H.m H H.NN N.mH H G.HoH momHooomuumoomm omoH

G.G H H.G- N.G H H.HoH H.m H G.mm a.mH H H.NHH momHooomnumoon uuosm
moHouuom summoH summoH\usmHoz usoHoz

 

suuHm um oNHm

o 'l
I’lll'll"l'I'llI!I'lllllll'l'lllllllll"|'l'l|lllllll'I'I'I'I'l'll'l'l'l'lll"'lll||
I'l'lll'lllllI'lIll'lllll'lllll'll'lllll'II"IIIIIIII'|l'l'Ill'Illlll'lllll'lll'll

.m CCm 4 mmsouw CH mHoCuoz Ho mUCmmCH
mCOEm mCHCommnummon mo mHHmuma oCm Cu3ouw Cmm3umn mHCmCoHuonMIIHm.>H mqmma

129

AHo.v mm.NI Hov.v mm. Aho.v hm.HI

mm.l

+1

H.HHH

+|

h.m0H

HwOCmonHCmHmV u
mmCHUomm mCouoECz

mmCHpomm ammoum 30m

AmOCoonHCmHmV u
mmCHomomnummmHm mCOH

mmCHUmmqummon unonm

o.N H H.H m.N H H.NoH m.G H o.GOH
o.G H G.H- G.m H N.moH o.HH H m.mm
sGG.G HG.- HHN.G NN.H- ANN.V HN.
N.m H H.H- G.m H G.HoH m.oH H G.ooH
G.N H v.5 m.m H G.NOH H.N H m.mm
mpH0mumm CymCoH CumCmH\quH03

 

CHCoz 0:0 as oNHm.

cmDCHHCOOIIHm.>H MHmda

130

Prenatal Plans for Infant Feeding

 

All mothers in both Studies I and II planned to breast-
feed their infants from birth. The earliest age planned
for weaning was one month in the case of a mother return-
ing to work at that time. No other mother had a specific
length of time for which she intended to breast-feed.
Most replied in general terms, with six to nine months
being a common answer.

Plans for first feeding of non-breast milk foods
were elicited in Study II. Many mothers reported they
had not given it much thought. Those mothers were in the
highest frequency in feeding Group 1, those infants
exclusively breast-fed the shortest time. In comparison,
mothers of Group 3, those infants exclusively breast-fed
the longest time, were most likely to give a specific age

range (Table IV.32).

Infant Feeding in the Parental Generation

Thirty-two women knew whether or not they had been
breast-fed as infants. Of these, only nine had been
breast-fed. Three knew that, although they had been bottle-
fed, their husbands had been breast-fed. Six reported
that either their mother or mother-in-law had tried to
breast-feed, but had given up after a maximum of two weeks.
They reported "insufficient milk" as the main cause of
weaning. The remaining fourteen had been bottle-fed, as

had their husbands.

131

TABLE IV.32—-Study II Responses to the Question: When
will you first give your baby anything but
breast milk (or an occasional "convenience
bottle")to eat?

 

Mother's Responses 1 2 3 Total
Have not thought about this 4 l l 6
When doctor advises 2 l - 3
Have thought about; not yet decided - - 4 4
When returning to work - 1 l 2
Between 3 months and 6 months - 2 - 2
After 3 months - 3 13 16
Missing data - l - 1
Totals 6 9 19 34

Table IV.33 breaks these categories down by feeding
Groups 1 to 3. There is a distinct trend toward feeding
Group 3 in cases where one or both parents had a history
of being breast-fed. Bottle-feeding predominated in

parents of Group 1 and 2 infants.

Spatial Arrangement of Families

 

The possibility that spatial arrangements of the
family might be linked to infant feeding was assessed.
It was assumed that intensity of mother-infant interaction
might vary with spatial distributions. In order to ex-
amine this, two categories of family dwelling types were

devised. The first consisted of those dwellings judged

132

TABLE IV.33-~Parental Generation Feeding as Infants by
Feeding Group of Infants in Present Study
(Study II only).

 

Infant Feeding Group

 

Neither Parent Breast-Fed 5 5 4 14

One Parent Unsuccessfully
Breast-Fed; Other not

Breast-Fed 1 l 4 6

Father Breast-Fed;

Mother not Breast-Fed - - 3 3

Mother or both Parents 1

Breast-Fed l 2 7 9

Feeding Modalities Unknown - l l 2
Totals 6 9 19 34

1

Both of these mothers cut short their infants' exclusive
breast-feeding to return to work.

to be compact: one story apartments and small houses with
a maximum of four rooms, all on one floor and in a 2 x 2
arrangement. The second category consisted of dwelling
arrangements judged to be dispersed: two story townhouses
and houses, and one story or split level ranch style homes.
Table IV.34 shows the distribution of dwelling types
by infant feeding group. Infants raised in dispersed type
dwellings were far more likely to be exclusively breast-
fed for a short period of time (Groups 1 and 2) than were

those raised in compact dwellings (x2 = 10.30; p < .01).

133
This does not appear to be a spurious association result-
ing from a more direct association of socio-economic
status with spatial arrangement. Table IV.35 shows that
maternal formal education, the best socio-economic pre-
dictor of breastffeeding behavior,is randomly distributed

with respect to dwelling type.

TABLE IV.34--Distribution of Infant Feeding Groups by
’Spatial Arrangement of Family Dwelling.

 

 

 

Spatial Arrangement

 

Infant Feeding Group Compact Dispersed Totals
Group 1 5 13 18
Group 2 5 16 21
Group 3 25 16 41
Totals 35 45 80

134

TABLE IV.35--Distribution of Spatial Arrangement of
Dwellings by Maternal Formal Education.

 

 

Spatial Arrangement

 

Education Level Compact Dispersed Totals

High School or

Vocational School 9 9 18

Some College (Less 6 14 20

than Bachelor's

Degree)

Bachelor's Degree 12 14 26

Any Graduate School 8 8 16
Totals 35 45 80

Chi Square Test: x2 — 2.16; Not Significant

Decision Makinngata

 

Reasons given by mothers of Study II infants for
adding the first non-breast milk foods to their infants'
diets can be grouped into four general categories. De-
tailed data are listed in Appendix D. The categories are:

1. Infant behavior

2. External suggestion or prescription

3. Situational

4. Maternal health
Table IV.36 shows the distribution of these categories
by month of infant age at first introduction of non-breast
milk food. Each is described in more detail in the

following paragraphs.

135

TABLE IV.36--Categories of Maternal Decisions to Introduce
First Non-Breast Milk Food

 

 

Maternal Decision 1 2 3 4 5 6 Totals
Infant Behavior 2 2 1 4 4 0 13
External Suggestion l 0 0 l 5 l 8
Situational 0 2 l 0 2 l 6
Maternal Health 3 o o o o o ' 3

Mothers in the modal category-~infant behavior-—
emphasized the failure of their infants to behave in ways
they considered appropriate for infants of that age.

This included instances in which infants had not yet
begun sleeping through the night, as well as cases in
which infants who had previously slept through the night
began to awaken again for a night feeding. This category
also included infants who cried or exhibited other he-
haviors interpreted as indicative of a desire to nurse

at what the mother considered overly frequent intervals
during the day.

Although it is highly likely that many of these
mothers had received advice from other persons or from
the media which contributed to their introduction of

formula or solids, each claimed that she had arrived at

136
the decision herself. While a few mothers explicitly
stated that their infants' behavior indicated a lack of
satiety with breast milk, most treated the infants'
behavioral state as something unrelated to current diet,
but something which the foods introduced had the power
to ameliorate. For example, small amounts of cereal were
thought to cause infants to sleep through the night,
though their precise action was unknown.

Thirteen mothers were placed in this category.
Examination of the temporal distribution shows that these
mothers augmented the breast milk diet either early--
before the end of the second month--or 1ater--at four and
five months--after a long period of exclusive breast-
feeding.

Mothers in the second most frequent category of
reasons attributed their behavior to a suggestion from
someone external to the nursing couple that the infant
had a biological need not being met by the breast milk
diet. These included one instance in which a pediatrician
told a mother that her infant's weight gain was inadequate
due to insufficient calories in the breast milk. It also
included a number of cases in which physicians claimed
that breast milk could not provide a specific nutrient--
iron--which the infants needed at a particular age.

Also included in this category are mothers who responded
to pressure from relatives charging that infants of a

given age or size could not be adequately nourished on

137
breast milk alone. Seven infants are in this category.
All feeding changes but one took place in or after the
fourth month.

Most of these cases share common features. In all
but the instance of inadequate infant weight gain, the
existence of an unmet need in the infant was diagnosed
without a specific means of assessment. Physicians did
not test for iron levels and mothers-in-law did not weigh
or measure infants. The criteria for recommending a new
food was a combination of age and "bigness." Although
the mother herself may have been inclined to make the
dietary change at the same time, she attributed her be-
havior to an external source.

The third most frequent category of reasons for
introducing new foods was that based on situations ex-
ternal to the nursing couple. They included mothers
going back to work or to school and choosing to either
replace breast-feedings or wean completely. There were
five infants in this category, and the feeding changes
occurred randomly from the second through the fifth month.

Three mother-infant pairs fall into the fourth
category--maternal health. In these cases some health
problem specific to the mother prevented nursing. In
two instances mothers were hospitalized again within the
first month postpartum. In the other, the mother reported
she was unable to nurse due to sore nipples. All these

feeding changes occurred by the end of the first month.

138

The four reasons discussed above account for 28
mother-infant pairs. Of the remaining five mothers, two
could not articulate reasons for feeding new foods, one
cited a reason which did not match any others given by
other mothers (wanting to take advantage of her infant's
interest in solid foods so as not to have a "picky eater"),
and two exclusively breast-fed throughout their participa-
tion in the study.

From inspection of the data it appeared that regard-
less of reason for feeding the non-breast milk item, the
foods themselves followed the patterns of the entire
sample. That is, formula tended to be given in the
earliest months and solids, in the latest. Furthermore,
there appeared to be no effect of decision making category
on whether the new food was a dietary addition or replace-

ment (Table IV.37).

Maternal Nutritional Status

 

Analyses reported here were designed to examine the
relationship of breast-feeding practices to changes in
maternal nutritional status in the postpartum. These
changes include loss of body weight and changes in fat
stores. Characteristics of the sample as a whole will be
described first. Maternal changes will then be analyzed
in terms of the common feeding patterns identified earlier
in this chapter, months of breast-feeding, and other

details of the breast—feeding process.

139

mmH\mNH mH\- <o\mm mm\u mN\- u\ma :H\mm mHouoa

 

u\mm : u n u .- mm: suHoo: Homumuo:
<N\mm 1 ad 1 ¢ mm I HoCoHuoCuHm
mh\mH d mdddfi H I I 4 CoHummmmCm HMCHmuxm
mm\mm u mammm «mam H mm mm HoH>osom umonH
mHmuoa G m w m N H ConHooo HMCHoumz

sumo:

.moHHomwumo ConHooo
Ns ooom sHH: ummommnmoz ummHm Ho umosoomHoom Ho moHuHooH-usm.>H HHmme

140

Descriptive Statistics for the Total Sample

 

The amount of weight women lost between the last
pregnancy weighing and the first postpartum weighing
averaged 8.7 i 1.6 kg (Table IV.38). After a continued
steady loss to three months postpartum averaging 2.7 kg,
the weight loss plateaued, with only a slight additional
loss (0.6 kg) to six months (Figure IV.15). This loss
represents 1.9 i 3.6 kg above reported prepregnant weight
at six months. The degree of variability in weight loss--
indicated by the standard deviation--is lowest for that
period spanning the birth. Afterwards, it increases
rapidly to three months, and then, like the average
weight change itself, it stabilizes.

Midarm muscle area was calculated from arm circumfer-
ence (c) and triceps fatfold (8) using the formula (C ;"552
(Gurney and Jelliffe 1973). After a slight initial
increase during the first month postpartum, average
muscle area declines steadily to five months, followed by
a slight increase in the sixth month (Table IV.38). The
mean overall loss is 1.0 cm2 (Figure IV.16). The standard
deviation at each point ranges from i 2.5 to i 6.6 cm2.

In contrast, the long range trend in midarm fat area
%? - %?2 ) is an increase. By the end of the third month
the average mother has had an increase of 2.4 cm2 (Table
IV.39). There is some fluctuation in the mean fat area

increments after that time, but the cumulative effect at

six months is an increase of 1.9 cm2 (Figure IV.16).

141

 

sz

CoHHMH>oC oumoCmum
mev Coos

omoH ou uHoH usoHoz

sz

CoHumH>ma CHMCCmum
mev Coo:

mmOH quHmz Enuummumom

Hmmv HHNV Amps Hmhv Anny Anny Hmnv
N.N N.N H.N H.N N.N N.N G.N
N.H N.N N.N N.N G.N N.a N.m
Hmov HHNV Ammo Achy Anny Anny Ammo
G.N N.N N.N H.N N.N H.N G.H
o.NH H.HH N.HH G.HH N.oH G.N H.N
G N G N N H m
sumo:

.Eduummumom oCu mCHHCU mmOH pCmHmz HMCHoumz HO mmHCmmoz 0>HHMHCECUIImm.>H MHmHB

17

16

15

14

13

11

10

POSTPARTUM WEIGHT LOSS (kg)

FIGURE IV.lS--Maternal Postpartum Weight Status; Mean :

 

142

 

 

 

 

 

 

 

 

T- /-
N/
/ 1 1
L
1 1 J
-5.
/
r r v v v T v I
3 1 2 3 4 5 6

MONTHS POSTPARTUM

Standard Deviation.

' 10

’ 11

12

13

 

WEIGHT LEFT TO LOSE ABOVE PREPREGNANT WEIGHT (kg)

144

Hmmv AHNV Hmhv HGNV Hubs HNNV sz
hm.N 0H.m mv.N mm.N om.N mN.m COHumH>mQ UHCGCmym
vm.H Nm.N HN.N v¢.N mm.H mo. ANEOV Cow:

omCMCU omud Ham

Ammv HHNV Hmnv HGNV Anny Anny sz
h¢.N mm.N «v.0 Hm.m hm.N mm.N COHumH>mQ cumccmum
oo.HI MH.HI Nb.0l HN.OI HN.OI 0N. HNEOV Com:

mmCmCU moud OHUmCz

 

.ECuHmmumom
«Cu mCHHCC momma Ham UCm mHomCz EHCGHZ CH mmmCmCO m>HuwHCECUIImm.>H mHmde

145
The variability in fat area changes--represented by the
standard deviation-sis greater in the first and fifth
months, and lowest in the fourth and sixth.

The change observed in midarm muscle area constie
tutes an average decline of three percent from the first
postpartum measurement to the last. The increase in fat
area for the same period represents an average change of
nine percent.

Although the postpartum change in mean muscle area
was fairly smooth, there were enormous month to month
fluctuations in the measurements of individual women,
resulting in extremely large standard deviations such
as those in Months 4 and 5. Such fluctuations were too
great to have resulted from gain or loss of muscle mass
due to mobilization of pregnancy protein stores. They
could have been a reflection of variations in hydration
associated with the physiological changes of the post-
partum period and lactation. They also could have been
influenced by changes in muscles brought about by intensive
lifting and carrying of infants. Given these possible
confounding variables, the value of muscle area as a
measure of nutritional status in such women is question-
able. Since adipose tissue is affected by hydration to
a far lesser extent, fat area, but not muscle area, will
be used throughout these analyses as an assessment of

maternal body composition.

146

Maternal Measures by Common Infant Diet Patterns

 

Table IV.40 shows cumulative weight loss at each
postpartum measurement by infant diet patterns. Using
analysis of variance, no statistically significant
differences among groups were found at any time. Despite
the lack of statistical significance, trends in group
weight loss patterns are notable. Group lb--those mothers
who began replacing breast-feedings with formula in the
first month-~had the lowest rate of loss throughout the
study, with a total loss of 10.06 kg. This compared with
a mean loss of 12.29 kg in the remainder of the sample.'
Although Group 2a--those mothers who replaced breast-
feedings with formula during months two and three--
had an initial weight loss comparable with most other
groups, by six months its total loss was only 10.75 kg.
Thus, in terms of overall rate of loss, Groups 1b and 2a
appeared to lose weight most slowly and Groups la, 2b,
and 3, most rapidly.

When cumulative changes in midarm fat area were
analyzed by infant dietary pattern by month, there were
no statistically significant differences between groups
(Table IV.41). However, trends similar to those in
weight loss discussed above were present. Groups lb and
2a--those mothers replacing breast-feeds with formula--
had the greatest overall increase in fatness. All groups
increased in fat area during the first three months.

After that time the direction of change was variable.

147

ANN-G HN.H GN.NHNN.NH HN.HHNo.NH NN.NHNN.GH NG.NHGo.oH GG.NHNN.NH G

ANN-v No.H HN.NHNH.NH NN.HHNN.NH HN.HHGG.oH NG.NHNN.N GG.NHNG.NH N
ANN.V NH.H NN.NHNN.HH NG.NHNo.NH NN.HHHo.HH GG.NHNH.N HN.NHoa.HH a
AHH.G oo.H GH.GHNN.NH NH.NHGN.HH GN.GHHN.GH HN.NHNN.N NN.NHHN.HH N
ANN-V Na. HN.NHNG.oH HG.NHHN.GH oo.aHaH.oH GH.NHoN.N NN.NHNG.GH N
ANN-G HN. HN.HHGN.N NN.NHNN.N HN.NHNG.N NN.HHNN.N Go.NHHo.OH H
AHN.G GN. NG.HHGG.N NN.HHoH.N NN.HHHN.N HN. HNG.N mo.NHNN.N N\H
ioHGG m .HHG N ANHG mm Has mN HHHG sH 1N1 mH sumo:
<>oz¢ Esuummpmom

 

 

I .COHHMH>oQ
oumoCmum + mev Cum: .mCouw mCHpoom quwCH UCm CuCoz Esuuwmumom
wn mCHCmHoB ECuHmmon ummq EOHH mmOH uanoz HmCHouwz o>HumHCECOIIov.>H mqmfie

148

me.
mb.H
um.
VN.
mm.

wv.

 

 

 

 

HN.NHGN.H GN.NHNG.H NN.NHGG.N HN.HHNG.N HG.~HNG. G
HN.NHNN.N oa.HHHN.N HN.HHNG.N NN.NHGN.N Na.NHNN.H G
GN.NHNG.H NN.HHNH.N NN.NHNH.H HN.NHNN.N NG.NHNo.N G
GG.NHoo.N GN.NHHN.N oa.GHaG.N HN.HHNG.H NN.NHNN.N N
HN.NHNo.H NN.NHGH.H NN.NHNN.N HN.NHNN.N No.NHGN.H N
HN.NHNG. GN.NHoG. HN.NHGG. NN.HHNN.H HN.HHNH. H
AHHG N ANHG sN sac mN AHHG sH ANG mH sumo:

Enuummumom

.sz mCouw mCHomom qumCH
.COHHMH>OQ

ouprmum H HNEov Com:

.msouu mCHUmwm uCMMCH UCC CuCoz Esuuomumom >9
quEmHCmmmz ECuHmmumom umHHm Scum mmum pom EHMCHE CH omCMCU o>HumHCECUIIHv.>H MHmmB

149

Maternal Measures by Months of Breast Feeding

Table IV.42 shows cumulative monthly weight loss
analyzed in terms of breastffeeding. Although no statisti-
cally significant differences exist, mothers who weaned
before the third month had a substantially lower mean loss
(10.6 kg) than those who weaned later (12.3 kg). The
possible connection between overall weight loss and the
timing of weaning appears to be the result of differences
in weight loss velocity before and after weaning. In all
groups breast-feeding less than five months, weight loss
velocity before weaning is greater than after (Figure
IV.17). In all four of these groups there was either no
further weight loss or an actual gain in weight in the
month following weaning. Furthermore, the preweaning
velocity tended to be greater in those breast-feeding
longer. Thus, the net result is the relationship between
timing of weaning and weight loss. The sample sizes in
those groups being compared in Figure IV.17 are small and
the variation in weight change at any point is great.
Thus, while the trends in mean weight change are suggestive
of an association with weaning, they are not predictive
of the postpartum weight loss for any individual woman and
analysis of them does not produce statistically signifi-
cant results.

There were no significant differences in midarm fat
area between groups based on months of breastffeeding.

Moreover, there was no discernable pattern of change when

 

150

 

AHN.G NG- No.mHNH.NH HG.HHoo.HH NN.NHNN.NH NG.¢HNN.NH No.NHNN.oH HN.HHNN.GH G
ANN-G HN. NG.NHGN.HH .NG.NHNH.OH NH.NHNN.NH HN.HHHN.NH NH.NHNN.N NN.HHNN.GH N
AHa.G No.H NG.NHHG.HH HH.NHNN.N .NG.NHoo.NH NG.NHGN.NH oo.NHNG.N HN.NHNN.N a
ANN.V NN. NN.NHNN.HH GN.HHNN.N HG.NHNN.NH «NH-«HHN.NH HN.HHON.N NH.NHGH.N N
AHN.V HN.H NG.NHNG.GH NN.HHNN.N NH.NHNH.NH HN.NHHG.HH .HN.HHNN.N NN.HHHN.N N
ioa.o NG.H NN.NHNG.N GN.HHNH.N GN.NHGN.HH NN.NHGH.oH NG- HGN.N «Na. HNN.N H
HNN.G GN. NN.HHWN.N NG..HGN.N NN.NHNN.N HH.HHHH.N HN..HHH.N oak Hon-N N\H
:onv m AHNG G 1N1 N AN1 4 ANG N 1G1 N ANG H .sumo:

Esuummumom

<>OZ<

sz mCHomom unmonm mo mCuCoz

l'lllll'l'l'l'I'Il'lll'llll"IIIII'IllllIIIlllllll|'|"l||'lll'|'llll'll'l'lll||"lll'll
lllllllll'III'IlllIII|Illllll'lllll'IIllllll"|llll|lll.lllllllll'l||I|I'|"I|'llllllllll

.CoHumH>oo UHCUCwum H mev Coo: .mCHomom ummmum mo mCuCoz UCM
CuCoz Eduuomumom an mCHCmHmS Esuummmnm ummq Eoum mmOH uanoz o>HumHCECUIINv.>H mHmde

151

12 1

10 ‘ ‘
.56 .57

 

 

 

 

12 T

1.25 .24

 

 

12 1

10 ‘ .77 .54

 

 

12 5

10 .50 .38

 

 

POSTPARTUM MONTH Pre- Post-
Weaning Weaning
Weight Change
Velocities

FIGURE IV.17-~Cumulative Postpartum Weight Loss by Month of
Weaning. Pre- and Post-Weaning Weight Change
Velocities (kg/month) Are Indicated. (0 =5-
Month of Weaning).

152
the overall rates of growth in fat were compared (Table
IV.43). It appears that early weaning was followed by
an immediate increase in fatness whereas later weaning
(after two months) was followed by a much lower increase
or a decrease. When numbers of women increasing and
decreasing in fat area in the month following weaning are
compared (Table IV.44), the same result appears. A
similar, but stronger, relationship exists between fat
change and the timing of the introduction of the first
non—breast milk food. Table IV.45 shows that introduction
of such food in the early months was followed by an
increase in maternal fat, whereas later introduction was

followed by a decrease (p < .01).

Maternal Measures and Details of Breast Feeding Practices
In order to study the association of maternal post-
partum weight and fat changes with the number and length
of nursing bouts, a series of Spearman correlation co-
efficients was computed for the entire sample. The
values for the breast-feeding variables were calculated
by averaging the number of all feedings and length of all
feedings reported within each month. Any woman who
reported any breastffeeding during a particular month is
included in the computation of the correlation for that
month. Thus, the sample size decreases from the first

to the sixth month as mothers wean their infants.

153

 

mCdSHEHmoshF82«

 

 

ANG.V NG. NN.NHNN.N NN.NHGH.N HN. HHH.H HN.NHHH. NN.NHNN.H- NN.NHGN.N G
ANG.G GN. NH.NHNN.N .GN.oHNN.N NN.NHGG.H HN.NHoo.N HN.NHNN. NN.NHNG.N N
AON.V NN.H GN.NHGN.N GN.HHNN.N .NN.HHNG.N GN.NHNN.H .NN.¢HNN. NN.HHNN.N a
ANH.G NN.H NN.NHNN.N HN.NHHN.N NN.NHNH. n._.NN.HH.._.o.H HN.NHoN.H HN.NHNN.N N
sic 3. NHHQH fioflq NfimflN SNWWT fiifiSH Qéflfia N
ANN.G GN.H No.NHoN. NH.NHHN.N HN. HHo. NG.NHNN.HI HN.HHoa.- «GN.HHNH.N H
AoHGG m IHNV G :NG N ANG a 1N1 N 1G1 N INC H suso:

Esunmmumom

¢>OZ<

Izv mmHooom-ummoum Ho msumo:

I .CoHu0H>mo UHMUCmum
+ HNEOV Cmoz .mCHUommlummoum mo mCuCoz oCm CuCoz Enuummumom ha
uCoEoqummz Ezuummumom umHHm Eoum moufl pom EHMCHE CH memnu o>HHMHCECUIImv.>H mqmfia

154

TABLE IV.44--Number of Mothers with Increasing and
'Decreasing Fat Measures in the Month
Following Weaning, by Month of Weaning.

 

 

 

 

 

'MOnth of Weaning“ f
Change in -
Fat Area 1 2 3 4 5 6 Totals
Increase 7 3 5 l 1 - 17
Decrease - l 3 2 1 - 7
Totals 7 4 8 3 2 - 24

TABLE IV.45--Numbers of Mothers with Increasing and
Decreasing Fat Measures in the Month
Following First Introduction of Non-
Breast Milk Food, by Month of Introduction.

 

MOnth of Introduction

 

Change in

Fat Area 1 2 3 4 5 6 Totals
Increase 16 7 5 7 7 - 42
Decrease l 2 2 3 10_ -_ 3 _ _-_21 -------
Totals 17 9 7 10 17 3 63
.__...___..___._.________.._.....g _________________________

155

The correlations between monthly maternal weight
loss and both breast feeding variables were not statisti-
cally significant (Table IV.46). A weak positive associa-
tion existed for Month 1 between weight loss and length
of feedings (p < .09), but it did not continue into the
later postpartum months.

Table IV.46 also shows the correlations of changes
in maternal fat area and the breast-feeding variables.
The length of feedings was not statistically correlated
with change in fat area until the‘sixth month. At that
time long feedings were associated with the greater loss
of fat; and short feedings, with the lower loss or gain
of fat (p < .02). The pattern of interaction of fat
area with number of feedings per day is more distinct.
Higher numbers of feedings per day are associated with
greater loss of fat; lower numbers of feedings, with
less loss of fat. The correlation is fairly strong in
the second month (p < .066) and highly significant in the
fourth month (p < .005).

To further explore the relationship between number
of feedings per day and fat area change, a series of
analyses were performed in which several possible con-
founding variables were controlled for. Mothers were
included in these analyses only if they exclusively
breastefed for at least two months and did not wean
before the fifth month, and if clear patterns of change

in their postpartum fatness could be identified (N=46).

156

TABLE IV.46--Spearman Rank-Order Correlation Coefficients:
Change in Maternal Weight and Fat with Length

and Number of Breast-Feedings.

 

Monthly weight.Loss with:

bkmtmm'of Feedings/Day

Correlation Coefficient .019
(N) (68)
Significance .434
Nfinutes/Feeding
Correlation Coefficient .170
(N) (66)
Significance .086
anthly Fat Loss With:
NmflmmcfifFaafingsflhw
Correlation Coefficient -.004
(N) (66)
Significance .487
Nfinutes/Feeding
Correlation Coefficient -.099
(N) (64)
Significance .219

.111 -.051
(68) (55)
.135 .355
-.029 .045
(68) (55)
.409 .374
.196 -.067
(61) (49)
.066 .325
-.125 -.113
(61) (49)
.169 .220

.142
(50)

.164

.068
(49)
.321

.394
(44)

.005

-0 037

(43)

.408

—. 033

(45)

.416

-.135

(44)

.193

.194
(37)
.126

-o 130

(36)
.225

-.O65
(41)

.343

.148
(40)
.182

-.194
(35)

.133

.363
(34)

.018

 

157
The means by which such patterns were identified will be
described below. The feeding criteria were aimed at
limiting unmeasurable differences in the intensity of
breast-feeding due to consumption of other foods by
'the infants. Data on breast-feeding details for these
analyses were taken from diet diaries at specific dates
of four, eight, twelve, sixteen, and twenty weeks. This
was to control for differences between infants in the
dates of diet diaries within the same month. Finally,
in these analyses changes in maternal fat area were
classified according to the overall pattern of such
changes. This was designed to control for minor monthly
fluctuations in fat area which might have been obscuring
overall trends in maternal nutritional status. To define
patterns of change in fatness, a least squares linear
regression line was first computed for each subject,
using days since parturition as the independent, and fat
area as the dependent variable. An "r" value was
calculated for each of the regression lines and a test
of significance was performed. Figure IV.18 shows the
distribution of significance levels for those 50 mothers
who met the infant feeding criteria described above.
There is a clear clustering of subjects at and below
the .10 significance level with several outliers at and
above the .30 level. Because of this distribution, .10
was used as the cut off for inclusion in these analyses

rather than the .05 level which is otherwise used

 

 

Flllllllllllllllllj

O
H

FIGURE

158

 

 

 

 

.1
__
:5:
.:3___BE]_._EJ
.05 .10 .15 .20 .25 .30 .35 .40 .45 .50

LEVELS OF SIGNIFICANCE

IV.18--Distribution of Significance Levels
for Regression of Postpartum Maternal
Fat Area on Days since Parturition.

159
throughout this dissertation as the upper limit of
statistical significance.

After narrowing the sample down to those for whom a
consistent pattern of change in fat area could be docu-
mented, the group was further subdivided according to
the direction of this change. All those women for whom
the slope of the regression line was negative formed one
group (Group A; N=l7). This was the expected pattern,
but numerically it was less common. Those with a positive
slope formed the second group (Group B; N=29). The
following analyses compare these two groups.

A comparison of maternal variables shows no signifi-
cant differences between these groups in prepregnant
weight, height, birthweight of the child, or length of
pregnancy (Table IV.47). Group A did gain significantly
more weight during pregnancy than did Group B, but after
the first month postpartum there was no significant
difference between the groups in the amount of weight
left to lose. Also, there were no significant differences
between the groups in the timing of introduction of non-
breast milk foods after two months.

Two major between group differences in breastefeeding
behavior were observed. The first confirms the relations
ship noted in Table IV.46: number of feeds per day is
linked to postpartum changes in maternal fatness. Group A
reported a significantly higher median number of feeds per

day at 4, 8, and 12 weeks than did Group B (Tables IV.48-50).

160

NN. NH.
GG. Na.
6N. 6N.
NN. w NH.
GN. NG.
NN. NH.H
No. HN.H
Go. NN.H
Ho. NN.N
HN. GH.
No. NH.N
unuunuun-uuleme ..... H----

o.NH N.HNN N.GH N.HNN smNmmc Nommmouma uo summoH

N. .H N.N a. H H.N loss usmHozsuuHm

N.N.H N.GGH N.N.H N.NGH Amos usoHom Hmmuoum:
G.N H G.N N.H H N.N smuuooumoa G sumo: uo omm
N.N H H.N H.H H N.N smummoumoa N sumo: Ho omm
N.N H H.N G.N H N.N smuumoumoa a sumo: Ho omm
N.G H N.H N.G H H.H smuuooumoa N sumo: uo omm
N.N H G.N H.N H N.N smuuooumom N sumo: mo omm
N.N H N.N N.N H N.G smuuomumoa H sumo: Ho omm

"uanmz HCMCmonmHm o>onm mm

N.N H N.NN N.G H N.NN loss usmHoz ummmmouomum

N.N H H.NH H.N H N.GH lost mHoo usmHoz somomooum

55m-mmmmm ------- m-mmmmm ------------------------ mmmmH-mmm

.CoHumH>mo CHMGCmum H Cmmz .m UCm H mmsouw

HON mOHQMH-Hmkr QEOOHDO SUM-mm @CM OflHUmEOQOHSUCAN HMQHGUN—lebwokfin HQQB

161

TABLE IV.48--Number of Breast-Feedings per Day at Four
Weeks for Infants of Mothers in Groups
A and B.

 

Feedings per Day'

 

 

 

 

Group A l 0 2 4 4 2 2 l 0 16
Group B 2 2 12 5 2 3 l 0 l 28
Totals 3 2 l4 9 6 5 3 l l 44

Median Test: Median = 7 Feedings/Day; x2 = 4.35; p < .05

TABLE IV.49--Number of Breast-Feedings per Day at Eight
Weeks for Infants of Mothers in Groups
A and B.

 

 

Feedings per Day

 

4 5 6 7 8 9 10 11 Totals

162

TABLE IV.SO--Number of Breast-Feedings per Day at Twelve
Weeks for Infants of Mothers in Groups
A and B.

.....

Feedings per Day

 

Group A l 3 3 3 1 3 0 2 0 16
Group B 2 10 9 2 2 O 0 0 1 26
Totals 3 13 12 5 3 3 0 2 l 42

The second difference between Groups A and B is in mean
length of individual feedings. Group A has shorter
feedings than does Group B at each point in time. These
differences approach statistical significance at four
weeks and are highly significant from 8 to 20 weeks (Table
IV.51). Thus, mothers with decreasing fat area during the
postpartum have, in general, more and shorter feedings
than do mothers whose fat area increases during the same
time. In the combined group of 46, minutes per feeding

is significantly and inversely correlated with number of

feedings per day at 4, 8, 16, and 20 weeks (Table IV.52).

163

TABLE IV.51--Minutes per Feeding for Groups A and B.
Mean : Standard Deviation.

 

 

 

Week Group A Group B t Sig.
4 19.1 i 7.4 24.4 i 11.0 1.89 .07
8 16.4 i 5.5 24.0 i 10 8 2.65 .01

12 14.8 i 4.3 21.2 i 8.3 2.89 .01

16 13.6 i 5.4 19.7 I 9.1 2.42 .02

20 12.2 i 3.6 17.1 i 7.0 2.60 .01

TABLE IV.52--Spearman Rank-Order Correlation Coefficient:
Minutes per Breast-Feeding with Numbers of
Feedings per Day.

 

Age of Infant (Weeks)

 

Correlation Coefficient -.514 -.468 -.221 -.424 -.269
(N) (44) (44) (44) (44) (42)

Level of Significance .001 .001 .075 .003 .043

CHAPTER V

DISCUSSION

Emergence of Infant Feeding Strategies

 

In order to identify major infant dietary patterns,
a series of analyses were done which assessed the patterns
of interaction of the various food behaviors which com-
prise infant diets. The following variables were con-
sidered: length of exclusive breast-feeding, age at
weaning, introduction of first non-breast milk food, and
mean duration of nursing bouts and number of feedings per
day in early infancy. From these it was apparent that one
variable, in a sense, formed the core around which all
the other dietary components were arrayed. This variable
was the length of exclusive breast-feeding. On the basis
of this three major dietary patterns are distinguishable.
Although they are defined primarily on the basis of this
one variable, the concordance of other variables supports
this division.

The first feeding group is composed of 18 infants
exclusively breastefed less than four weeks. Among these,

breast-feeding was never well established. In some cases,

164

165

maternal health problems precluded the establishment of
lactation. In the remainder, a variety of factors

working together led to the early introduction of non-
breast milk foods. The detailed decision making data

from Study II indicate that failure of infants to behave
as expected by the mother as well as external suggestion
of nutritional insufficiency led to early changes in the
infant diet. Data from Study I support this: of twelve
infants in Study I with this feeding pattern, four were
known to have been diagnosed by physicians as losing
weight beyond the expected initial perinatal weight loss.
Most of the remaining eight infants were supplemented for
other reasons. A large prOportion of mothers, for example,
reported that their infants cried and fussed too frequently,
as if unsatisfied and as a result, the mothers changed

the infants' diets.

In both studies, the new mothers were in many cases
without "doulas."1 Their own mothers were frequently there
to help in the early postpartum, but, never having breast-
fed an infant themselves, either could not give advice

appropriate to breastefeeding or suggested using non-breast

 

1Raphael (1966) devised the term doula as a label for
the person who "mothers the mother." She has noted that
such a role is culturally defined in most traditional
societies. Furthermore, she attributes the decline of
breast feeding in the 19505 and 19605 and increasing
incidence of lactation failure in the industrialized
countries to the isolation of new mothers from other sup-
portive, experienced older breast—feeders.

166

milk foods. In Study II where maternal and paternal
grandmothers' infant feeding modalities were known, all
women in Group 1 were, in a sense, first generation breast-
feeders. Five of six had been bottleéfed, and the other
had been breastefed for two weeks until her mother "quit
because it was so frustrating." Thus any advice these
grandmothers could give from personal experience would have
been from their 19503 experience with early solids, rigid
schedules, and few feedings per day.

It is possible that these mothers had a set of ex-
pectations for infant behavior which differed from those
of mothers in the other feeding groups. This seems un-
likely, however, since all were raising their first child
and most had read the same books and been exposed to the
same information in prenatal classes. On the other hand,
it is quite possible that due to the inhibition of the
let-down reflex by factors such as stress, anxiety, and
infrequent nursing as described by Jelliffe and Jelliffe
(l979b), these infants really were unsatisfied and there-
fore behaved in ways indicating it.

Mothers of infants in Group 1 followed two different
strategies for coping with the alleged insufficiencies
of breast milk. The first was to add food to the existing
breast milk diet (N=7). In most cases this was formula,
and it was consumed by the infant in small amounts immediately
after each nursing session. This feeding pattern was one

which physicians usually prescribed to mothers if there was

167

any question about milk sufficiency. The second strategy
was one of replacing entire breast-feedings with formula
(N=ll). In no case where information was available was

this pattern known to have been suggested by a physician.

This distinction between the use of formula and, in
two of eighteen cases, solids as additions or replacements
roughly predicts the time of weaning. "Replacements"
correlate with early weaning-~in all but one case, weaning
occurred by the end of the second month. In contrast, with
"addition" there was only one case in which the infant was
weaned by the end of the second month. One possible
reason for this difference in length of breast-feeding is
physiological. Since milk production is maintained by the
hormone prolactin, its production being stimulated by the
beginning of each nursing session, the decrease in number
of feedings per day in the "replacement" group should be
expected to decrease milk supply, thus hastening the wean-
ing process. Since in the "addition" category the number
of breast-feedings per day remains unchanged after intro-
duction of formula or other foods, no decrease in breast
milk supply would be expected.

One striking feature of the women in Study II who
follow the Group 1 feeding pattern is their almost total
lack of prenatal planning of their overall infant feeding
strategy. All had made the choice of breastffeeding over
bottle-feeding, but none had thought any further about the

use of other foods. Some seemed rather startled to be

168

asked about this and reported that family and acquaint-
ances often wanted to know whether or not they intended

to breast-feed, but none ever showed any interest in other
aspects of infant feeding. Thus, while they had every
intention of nursing their infants, they were quite
susceptible to suggestions from others about feeding their
infants in ways which, at least in the early weeks of life
when lactation was being established, could undermine their
chances for being successful breast-feeders.

In the remainder of the sample, all infants were
exclusively breast-fed over five weeks, and thus lactation
might be considered well established. They are divided
into moderate length exclusive breast-feeders (N=21) and
extended exclusive breast-feeders (N=4l), with the division
at the end of the third month. This separation is neither
arbitrary nor artificial. Instead, it reflects the contri-
bution of the new non-breast milk food to the total diet.
In those with the moderate length exclusive breast-feeding
referred to as Group 2, the formula or solids introduced
were almost always replacements for breast milk feedings,
while in the extended exclusive breast-feeding group
(Group 3), these foods were additions.

The reason for this differential impact of intro-
ducing either solids or formula into the diet may lie in
the fact that, althOugh both groups were consuming solely
breast milk at the time of the introduction, there were

important internal differences in the configuration of

169

the diets. Group 2 infants had, from birth, had fewer
breastffeeding bouts per day than Group 3. In Month 1,
for example, the median number of feedings was six for
Group 2 infants and eight for those in Group 3. Moreover,
the average number of feedings per day tended to decline
in Group 2 between the first and second months of life,
while at the same time many Group 3 infants actually in-
creased their daily feeding frequency. In some cases this
increase may have been unnoticed. However, in at least
one case it was deliberate. A mother in Study I, whose
pediatrician prescribed formula supplementation because
the infant failed to gain weight, instead chose to increase
the number of feedings per day to increase her milk supply.
(She also chose to change pediatricians to one more sup—
portive of and knowledgeable about breast-feeding!)

Thus when new foods were added to the diets of Group 2
infants, they may have reduced the children's appetites
to some extent, thereby reducing suckling. Then through
lowered levels of prolactin, the milk supply might well
have decreased, hastening a transition from breast milk
to a non-breast milk diet. The frequently earlier weaning
of Group 2 infants may be an end product of this process.
Of twenty-one infants in Group 2, eleven had been weaned
by the end of the study, compared with only one infant in
Group 3.

Although the new foods fed to Group 3 infants may have

had a satiating effect similar to that in Group 2, the

170

frequency of nursing bouts was already at such a high

level that any reduction associated with the addition of
new foods may not have appreciably affected prolactin
levels. Thus, there would have been no progressive de-
cline in milk production such as that hypothesized for
Group 2. This is supported by Delvoye et a1.'s observation
that a minimum of six feedings per day maintains high
prolactin levels (1977).

The actual reasons for the variability in the early
months in number of feedings per day are unknown. Although
it is tempting to describe Group 2 as implementing a
pattern of "scheduled" feedings and Group 3 as feeding "on
demand," this distinction cannot be made. A study of
individual infants' diet diaries shows that while some
nursed at the same time each day, these infants were a
definite minority. For most Group 2 and 3 infants, the
times of feedings and the intervals between them fluctuated
from week to week. Relatively little concern was expressed
by mothers over feeding schedules--beyond hoping for an
uninterrupted night's sleep. Many claimed to be feeding
their infants on demand and among these, the full range
of feeding frequencies was represented.

Thus, the term "on demand" may be more useful in
describing maternal intentions than in describing actual
maternal behavior. It is, in fact, the perception of
the infant's demand which initiates a feeding, rather than

the demand itself. Many factors can shape this perception.

171

Vis and Hennart (1978) have attributed the pattern of
extremely frequent feedings and apparent contentedness
of infants among the Kivu of Zaire to the continuous body
contact of mother and child. A mother soon learns to read
her infant's smallest behaviors and to react to them.
In contrast, in the developed countries such as the United
States, numerous factors now serve to distance a mother
from her child and thus reduce her ability to perceive
hunger or other needs. For example, cribs, strollers,
and infant seats now alleviate the need to hold infants.
In Lansing, Michigan it was observed that most pediatricians,
if asked, strongly discouraged an infant's sleeping in the
parents' bed or even in the parents' bedroom. One mother
in Study I took this advice so seriously that the infant
was given the home's only bedroom and the parent slept on
a rather broken down sleep-sofa in the living room.
Although it was not possible in this study to take
objective measures of the amount of time mothers spent
interacting with their infants, it was possible to separate
the families into two categories, based on the degree to
which their living situations forced close contact or
permitted distancing. Infants in Group 2 were far more
likely to live in two story dwellings (76%) than in single
story, compact ones (24%). Conversely, a majority of
Group 3 infants spent their early months of life in
compact dwellings (61%) and a minority, in dispersed ones

(39%). Thus, mothers living in situations which forced

172

them to be in close proximity to their infants fed them
more frequently in the first few months than did women
in larger dwellings. Among the latter, infants often
napped in a second story bedroom while the mother was
occupied on the first floor or even in the basement. An
infant's quiet fussing which might be audible to a mother
in a small apartment and perceived as hunger could go
unnoticed in a larger home. The time necessary for an
infant to gradually rouse itself to give a loud cry could
be sufficient to create a significantly longer interbout
interval, resulting in fewer nursing bouts per day.

Another factor which might have affected the fre-
quency of early feeding is breast-feeding advice received
during the early postpartum. Like Group 1, most Group 2
mothers from Study II were really first generation breast-
feeders. Some recognized this and, while accepting their
own mothers' help after the delivery, restricted it to
housekeeping activities, choosing to manage all baby care
themselves. In Group 3, many of the mothers and fathers
had been breastéfed themselves. Others who had not been
breast-fed did have sisters or friends who were experienced
breast-feeders, and the new mothers often chose these
women over their ineXperienced mothers for help after
coming home from the hospital. Therefore, Group 3 mothers
probably received infant feeding advice more appropriate
to breastffeeding than did Group 2. This may have

contributed to the observed differences in patterns of

173
breast-feeding.

Thus, the study of infant dietary data clearly shows
that breastffeeding is not a unidimensional behavioral
category. Within it are several distinctly different
infant dietary patterns. These patterns appear to be
relatively unplanned, instead evolving through the early
months of the infant's life as one action taken by the
mother selects certain later possible options for her and
eliminates others. Because of this, it is possible to
make predictions of later dietary patterns of infants
based on early feeding habits. Of the variables examined
in this study, number of nursing bouts per day appears

to be the best predictor.

Igteraction of Infant Diet and Infant Growth

The analysis of infant growth data shows that its
relationship to infant diet is a complex and dynamic one.
Two types of interaction appear to occur. As has been
assumed by much previous research, the foods and feeding
habits of infants certainly seem to influence their growth.
On the other hand, the way in which infants grow, particu-
larly in the early months of life, appears to affect the
manner in which they are fed. Together these form a
reinforcing feedback relationship. This section will
summarize the results of the data analysis which led to
the inference that these two aspects of the infant feeding
process exist. It will then show that supporting evidence

can be found in some of the most widely cited studies of

174
infant growth and feeding. Finally, it will attempt to
demonstrate that by changing one's assumption from that of
a unidirectional causal link between feeding and growth
to one of a more dynamic relationship such as that found
in these data, some of the inconsistencies which have been

the focus of so much attention and debate can be resolved.

Influence of Growth on the Dietary Pattern of Infants

 

This study demonstrates that, starting at birth, the
feeding regime to which any infant is subjected is not a
random event with respect to the size and growth of the
child. In the most general terms, it appears that the
feeding patterns of infants who are large at birth and
display rapid growth velocity in the early weeks of life
can be contrasted with those of infants which at birth
are relatively small and initially grow rather slowly.

Infants in Group 3--that is, those exclusively breast-
fed until at least the fourth month of life and not weaned
until the sixth--have at birth the greatest mean weight
for age (106.4% of NCHS median) and weight for length
(95.6% of median) of any of the three feeding groups.
Those in Group l--with the shortest exclusive breast-
feeding--have the lowest mean weight for age (102.6% of
median) and weight for length (93.7% of median). The size
superiority of the Group 3 infants is maintained through
five months of age in weight and four months of age in
weight for length, though the differences between groups

narrows over time. One detail of breast-feeding practice

175
is also linked to size at birth. Those infants whose
mothers settle into a pattern of shorter nursing bouts
during the first month tend to be heavier and fatter at
birth than those whose mothers report longer feedings.

Not only are Group 3 infants bigger at birth, but
their initial growth velocity in weight for age and total
fatfolds exceeds that of either of the other two groups.
Mean changes in percent of median weight for age in
Month 1 were -l.8%, 1.5%, and +3.3% for Groups 1 through
3, respectively. In total fatfolds, the average changes
were +5.9mm, +5.8mm, and +6.4mm for the groups. Although
Group 1 experienced a major dietary change during that
time, Group 2 along with Group 3 was exclusively breast-
fed.

There are several possible reasons why the feeding
patterns of these infants appear to be a function of
their early size and growth. First of all, there may be
behavioral differences between large and small infants.
If the former were more aggressive nursers or more vocal
in communicating their demands to be fed, one would expect
them to more rapidly establish the successful lactation
which is characteristic of their feeding group. Data to
document this are not available here, though when talking
about infants in general, many mothers tended to attribute
these behavioral traits to certain size infants.

A second possible explanation is that the relations

ship between infant size and feeding may exist as a part

176
of a larger complex of variables concerning maternal
health and health consciousness. Favorable life-long
living conditions and nutritional status (Weinstein 1981)
as well as conditions during pregnancy (Philipps and
Johnson 1977) have been linked to the birth of larger
infants. These factors, in turn, characterize the health-
awareness of women most likely to be successful, long term
breast-feeders (Sauls 1980). The direct associations in
this study of mothers' educational level and length of
lactation are consistent with such a link. Thus, the
correlation of birth size and feeding modality may in fact
simply be incidental to this more direct relationship.

The third possible--and perhaps most readily docu-
mented--explanation for the association of growth and diet
in this study is that mothers have concerns about and
perceptions of infant needs which are specific to particular
sizes and shapes of infants. Concern was expressed by a
number of mothers in this study, for example, whose infants
were relatively low in birth weight. This was particularly
true of mothers who had been falsely led to believe they
would give birth to larger than average children. The
birth of a small infant in such cases seemed to cause
considerable anxiety over the child's health. Mothers
appeared to handle the infants more tentatively, often
having a great deal of trouble dressing and undressing
them for the first month or so. One of these mothers

(with a husband 6'2" tall and herself 5'7" with a 49 pound

177
pregnancy weight gain) gave birth to a six pound son.
The oversized clothes she had purchased in anticipation
of a nine pound infant further accentuated what to her
was his abnormally small size. Like other mothers in
similar situations, she initially worried a great deal
about his health and treated him--and made sure others
treated him--as though he were very delicate. On the
other hand, mothers of large infants tended to be more at
ease with them, treating them as more rugged and less
fragile individuals. The mother of the one set of twins
studied gave only conditional consent before the birth
to allow them to be measured. When each weighed over
seven and a half pounds, she readily agreed to participate.
She later attributed her earlier hesitation to the fear
that they might be small babies.

This variability in concern over the well-being of
infants based on size appeared to manifest itself in
attitudes toward infant feeding. Despite the growing
public awareness that "bigger" infants are not always
"better," such attitudes are not universal among new
mothers. And even among those who were seriously con-
cerned about obesity, there was a strong feeling that,
while fat babies are not necessarily the healthiest
babies, smaller ones are even more at risk. In discussing
the infant silhouettes used to stimulate discussion of
body shape, most mothers chose the fattest and thinnest

infants as less healthy and of those, finally chose the

m
thinnest as the least healthy, saying that it looked
malnourished. Some reasoned that a fat infant would
probably lose its "baby fat," but an overly thin one might
well be seriously ill.

Such concern over thinness was not limited to hypo-
thetical infants. Several mothers referred to their
infants' thin limbs and visible ribs as "bird-like" when
seen in the hospital after the birth or at home in the
early postpartum. Others commented that they would "have
to fatten up" their infants. Mothers such as this could
easily doubt the efficacy of a feeding method such as
breast-feeding with no immediate, tangible evidence (such
as an emptied formula bottle) of success or benefit. An
infant's cries which might, to a self-assured mother,
have a different meaning, might be more likely to be
interpreted as indicative of insufficient milk by a
mother already anxious about her infant's health.
Furthermore, the tension surrounding the feeding event
in those cases might well result in inhibition of the
let-down reflex. If a non-supportive atmosphere were
added to this, the chances of successfully establishing
lactation would probably diminish even further.

The later growth of infants whose mothers were
unable to successfully establish lactation indicates
possible overcompensation. By Month 3, Group 1 infants
were fatter than either Groups 2 or 3. In several cases,

infants who had had initial weight losses became the

 

179

fattest infants in the study. One such mother was told
by her pediatrician to overdilute her son's formula at
four months. When interviewed after that, she expressed
her frustration: "You know, at first I was so worried
because he wasn't gaining weight. Now we have to worry
because he's gaining too fast. You can't win--you never
know what is right."

While the small size of an infant sometimes worked
against mothers' efforts to breast-feed, large size was
used to advantage by others. Mothers with large infants
who wanted to exclusively breast-feed used the issue of
their infants' size to silence grandparents or pedia-
tricians who considered breast milk an inadequate source
of nutrients or thought supplements of solid food should
be added at an early age. When asked if anyone had
suggested a dietary change in the preceding month, these
mothers (such as one whose child weighed over twelve
pounds by one month of age) would laugh and poke at their
infants' robust bodies, replying, "Yes, but does he 123k
underfed?"

Thus, infants who were initially small and grew
slowly appear more likely to be exclusively breast-fed
for short periods of time. This may be due to a combina-
tion of factors, including maternal anxiety, inhibition
of the let-down reflex, and succumbing to external
pressures to supplement the breast milk diet. Larger

infants may be less likely to be the objects of anxiety

180
and their mothers, better able to maintain breast-feeding
in a nonsupportive social environment. Since the early
feeding of an infant has been shown in this study to be
a prime determinant of the future course of feeding,
the impact of infant size on early feeding behavior may
be an important factor in attempts to understand the

dynamics of the infant feeding process.

Influence of Dietarnyatterns on the Growth of Infants
While the previous paragraphs have discussed the
possible impact of size at birth and early growth patterns

on early and important feeding behaviors, the following
will examine what is, to some extent, the reciprocal
relationship. It will concentrate on the way in which diet
seems to affect growth. In a study such as this of a
free-living sample of 80 infants consuming almost 80
different diets, it is impossible to try to answer some

of the questions which have traditionally been asked in
infant nutrition research. The relative impact of formula
or breast milk, for example, becomes meaningless when
infants are consuming varying amounts, perhaps eating

solid foods as well, and, as has been argued above, are

of different sizes. However, there is a type of issue
which can be explored here better than in more controlled
studies. That is the impact on growth of minor variations
in dietary patterns in infants where mothers have decided
these are the optimal feeding strategies for their children.

Two of these variations were identified in the analysis of

181
the dietary intake data: 1) the use of the same food in
diets for the same age children as replacements or
additions to the existing diet, and 2) the use of different
foods to fulfill the same dietary role in children of
similar ages.

Those infants in Group 1, all exclusively breast-fed
for less than four weeks, provide the basis for comparing
the effects of the use of formula as a dietary addition
(Group la) and as a replacement (Group 1b). As stated
earlier, those infants for whom it served as an addition
consumed approximately two ounces after one or more
nursing bouts per day. Those for whom it was a replace-
ment consumed larger quantities of formula in place of
entire nursing bouts. Generally among the former, weaning
occurred after the second month and among the latter,
before the end of the second month.

Growth differences between these two groups were
marked during the first month. Although the "addition"
of formula suggests Group 1a infants might display
accelerated growth while "replacement" suggests Group 1b
infants would maintain steady growth, practically the
opposite was observed. Group la fell from an average of
104.4% of median birth weight to 96.6% of the median
weight for age at one month, while during the same period
Group 1b increased from 101.5% to 103.5%. The decreasing
percent of the median for la represents an actual loss

of weight in at least two cases. Group la remained at a

182
constant 93.6% of weight for length during the first
month, while Group lb rose from 93.8% to 100.1%. While
both increased in total fatfolds, that for la was slight
(+2.9mm), while the average increase for 1b was relatively
large (+9.4mm). By one month Group la was the thinnest
group of infants, compared to Group lb which was the
fattest. Those Group lb infants which were weaned during
Month 1 had the most marked increase in fatfolds during
that month.

Undoubtedly these early growth differences were due
to differences in caloric intake. The larger amounts of
formula in Group lb provided sufficient calories to
maintain growth, but decreasing breast-feeding frequency
in almost all cases eliminated any chances of breast-
feeding success. The smaller amounts of formula in
Group la were probably unable to compensate for the
apparent insufficiency of breast milk, but, by not inter-
fering with feeding frequency, allowed lactation to become
established. In fact, some mothers in Group la were later
able to reduce the formula supplementation and maintain
infant growth and satiety on a diet consisting almost
solely of breast milk.

Thus, the introduction of the same food into the
diets of similar aged infants produces different patterns
of early growth. A closer examination of the role the
new food has in the diet and the circumstances surrounding

its introduction provide plausible explanations.

183

The second diet variation noted in this study which
appears to have a significant impact on growth is the use
of different foods in similar dietary capacities for the
same age infants. The distinguishing characteristic of
Group 2 was the overwhelming use of newly introduced
foods as replacements for breast milk in the diet, rather
than as additions. This is consistent with the research
of Marlin et al. (1980) who found that before three months
of age, most dietary changes are caloric replacements.

In the present study, Group 2 has been further subdivided
into those mothers who replace breast milk with formula
(2a) and those who do so with solids (2b).

The patterns of growth for Groups 2a and 2b began to
show differences between one and two months, when the
first introductions of non-breast milk foods occurred.
From that time on, Group 2b displayed a slower overall
growth rate than Group 2a. In weight for age, Group 2a
went from 104.0% of the median at one month to 103.6%
at five months. At the same time, Group 2b dropped from
106.5% to 101.1%. The difference between the groups was
even more striking in weight for length. Between one and
five months, Group 2a increased from an average 99.7%
of the median to 102.0%, while Group 2b fell from 100.3%
to 90.0% in the same time period. The difference
between the groups was highly significant at five months.

After one month of age, Group 2b infants were

frequently smaller than infants in any other feeding group.

184
When the extreme growth quartiles were compared, both
Groups 2a and 2b tended to be overrepresented in the
lowest quartile of weight for age by three months. But
of the two, only the infants in Group 2b were overrepre-
sented in the lowest quartile of weight for length. Thus
while all the infants in combined Group 2 tended to be
small, those in 2b tended also to be thinner, and the
difference between the groups increased over time. This
plateauing of growth rates among infants consuming breast
milk and solids after two months is consistent with that
observed by Ferris et a1. (1979) for fatfolds.

These differences between the groups in growth seem
to reflect the caloric differences in the replacement
food. The mean caloric value of the initial replacement
formula was estimated to be 218 kcal, while that of the
solids was 48 kcal. Thus, the total daily energy intake
of the infants consuming solids probably declined immedi-
ately. Since breast milk and formula are more nearly
equivalent in calories, the energy intake of Group 2a
infants probably began to decline only after the decrease
in numbers of nursing bouts per day had an impact on milk
production.

Two reasons for introducing non-breast milk foods
predominate in Study II, where decision making factors
were specifically elicited, and in cases in Study I where
such information appears in fieldnotes. It was during

this period between one and three months that most mothers

185

who had planned to return to work did so. Also during
this time, the remaining mothers began to feel that their
infants were demanding to nurse too frequently, waking
too frequently at night, or exhibiting other behaviors
they deemed inappropriate. Mothers returning to work
almost always introduced formula into the diet. Those
seeking to change their infants' behavior used both
formula and cereal, but the latter was more common.

When the two reasons for the feeding change are
compared, the introduction of a new food because of the
physical separation of the nursing mothers and their

infants clearly implies replacement. When these mothers

 

discussed their plans, they did so in terms of which
feedings (usually morning and late evening) they would
retain as nursing bouts and which ones would be changed
to formula feedings by the caretakers.

The feeding of solids was not discussed in the same
way. Although the mothers of infants in Group 2b were
never asked, their attitudes and actions seem in retro-
spect to indicate that they were not consciously trying
to decrease their infants' breast milk intakes. They
often reported, for example, that they had begun to feed
their infants a certain amount of cereal at a specific
time of day. In doing so, the event was often described
in relation to either parental activities or non-feeding
activities of the infant (e.g., before bedtime, after

the parents' meal, or before the father came home from work).

 

186

Although the feeding categories defined here give
the impression that the introduction of solids or formula
are comparable options, little was observed to indicate
that mothers viewed them as interchangeable. Many mothers
in both studies attributed a pacifying effect to solids,
particularly cereal, and it appears that it was this
quality which those mothers in Group 2b sought. Formula
was instead classified as a milk. Thus, those returning
to work substituted a similar substance for the unavailable
breast milk.

Thus, the use of two different kinds of infant foods
with similar impacts on the existing diet produces
different patterns of growth. This can be explained by
comparing their caloric values. An examination of the
situations in which they tend to be used reveals the

rationales for their introduction.

SupportingiEvidence from the Infant Growth and Diet
Literature

 

Although other authors (Sauls 1980; Himes 1979) have
stated that samples of breast-fed infants are highly
selected and that those infants not thriving on breast
milk automatically drop into other feeding categories,
data to support this have never been explicitly presented.
However, in reviewing several of the growth studies cited
in Chapter II, similarities of data gathered in those

studies to those data presented here are apparent.

187

In three investigations, infants who were either
weaned or changed from breast milk to a mixed diet within
the first six weeks of life had slower rates of growth
during that period than did those infants who continued
to be exclusively breast-fed. Fomon et al. (1970) compare
the 45 infants who dropped out of their study with those
149 who adhered to the prescribed breast milk diet.
"Insufficient milk" was reported to be the reason behind
much of the attrition. Female infants who did not
complete the study had a weight gain velocity only slightly
lower than that of those in the study. However, the
difference for males was more pronounced, significantly
so in the second month. Holly and Cullen (1977) compare
breast-fed infants and those receiving both breast milk
and formula by six weeks. While the former had a per-
centage weight gain of 33% in the first month, the latter
increased by only 24%. Saarinen and Siimes (1979) show
that infants weaned to either of two types of formula
before one month of age had lower weights and lower weight
gain velocities from two weeks to two months of age than
did infants who continued to be exclusively breast-fed.
Thus infants making more significant early weight gains
do appear to be more likely to continue to consume a diet
composed solely of breast milk than those with slower
rates of growth.

The possible overcompensation in feeding by mothers

whose infants were either weaned or changed to a mixed

188

diet before one month of age is supported by evidence
from two studies. By six months of age the mixed-feeding
infants in Saarinen and Siimes study (1979) had gone from
their earlier weight deficit (in comparison to those
breast-fed) to a weight advantage. And in Huenemann's
study of six month old infants, she finds that those
breast-fed less than one month are overrepresented in the
upper 10% and underrepresented in the lower 10% of her
sample when ranked on an obesity index derived from
weight, fatfolds, and girth measures (1974).

Therefore, examples of the interactions of growth
and feeding observed in the present study can be found
in the infant feeding literature. However, they have
normally been presented without comment or statistical
analysis, perhaps because of the assumptions about the
interactions of infant growth and diet held by the

researchers.

Conflicting Evidence for the Impact of Diet on Growth

As shown in Chapter II, the current literature on in-
fant diet and growth is plagued by conflicting results.
Evidence can be cited both for and against the argument
that formula leads to heavier and fatter infants than does
breast milk. The same is true for early or late introduc-
tion of solid foods. This section suggests that an approach
to the study of_growth and diet which does not assume a
unidirectional relationship between growth and diet may be

useful in trying to resolve these inconsistencies. The

189
disputes over the effects of formula are perhaps most
easily resolved.

When several studies purporting to compare infants
"breast-fed" and "bottle-fed" are examined, their results
appear to depend on the way in which infants are assigned
to feeding groups. Saarinen and Siimes (1979) classify
as bottle-fed all infants who were initially breast-fed,
but were weaned by one month of age. In contrast, the
University of Iowa studies (Fomon et al. 1970; Fomon et
al. 1971) and the early study by Thomson (1955) consider
as bottle-fed only those infants EEZEE breast-fed.

The patterns of growth for breast-fed and bottle-fed
infants do not differ in the Fomon et al. (1970; 1971)
and Thomson (1955) studies. However, there are differ-
ences between the two groups in the Saarinen and Siimes
study (1979). Breast-fed infants have a high initial
growth rate. Bottle-fed infants recover and surpass the
breast-fed group in later infancy. These results look
very much like those obtained in the present study: early
growth superiority of those infants successfully breast-
fed and later overcompensation in cases where lactation
was unsuccessful. Thus, differences in growth patterns
among the three studies cited above may reflect the
feedback relationship of growth on diet and not simple
differences between the nutritional impacts of breast

milk and formula.

190

Interaction of Infant Diet and Maternal
Nutritional Status

 

Two assumptions about the relationship of infant
feeding and maternal nutritional status were examined:
1) that breastffeeding is associated with postpartum weight
loss, and 2) that breast-feeding leads to decreases in
maternal fat stores. This section will summarize the
results of those analyses. It will then propose possible
explanations for the patterns of anthropometric changes

observed.

Breast-Feeding and Weight Loss

No clear relationship was found between weight loss
and breast-feeding. When the data were analyzed by months
of breast-feeding, those weaning within the first two
months appeared to have the lowest overall loss during
the study. This may be due to the fact that the greatest
weight losses for the sample as a whole occurred within
the first three months. Those weaning before that showed
very little loss during that time. Weaning appears to be
accompanied by a temporary stabilization in maternal
weight.

When data were analyzed by infant feeding group,
the greatest initial loss was among mothers of Group la
infants and the lowest, among those of Group lb. Overall,
the lowest mean weight losses were in Groups lb and 2a.
However, none of these differences were statistically

significant. These results indicate that there may be

191
some effect of the proportion of the infant diet cone
sisting of breast—feeding on weight loss. Both Groups
lb and 2a replaced nursing bouts with formula feedings,
while other groups followed feeding practices which
maintained the child's demand on the mother for nutrients
at a higher level for a longer time.

The failure of these analyses to clearly confirm
the assumed relationship between weight loss and breast-
feeding may be due to three factors. First, mothers
may have compensated for the nutritional costs of
lactation by increasing their own dietary intake. When
questioned at six months, virtually none of the women
reported dieting or greatly increased activity level.

On the other hand, most reported little change in appetite
or intake from "normal" levels. However, no dietary
intake data are available to substantiate these state-
ments.

The second factor which may complicate attempts to
isolate a significant relationship between weight loss
and breast—feeding is the absence of a control group of
women never breast-feeding. Most of the calculations of
nutritional needs during lactation treat it as a physioé
logical state identical to that of the "reference female,"
with the demands of milk production simply added to an
otherwise normal body. Lactation may, in fact, be
accompanied by basic alterations in metabolic status.

The wide variety of physiological responses of women in

192
this study to breastffeeding suggests this. Thus women
breast-feeding even for a limited period of time may be
affected to the extent that they are an inappropriate
comparison group in analyses trying to isolate the effects
of lactation on weight. In fact, if a change in metabolism
accompanies breast-feeding--and the association of altered
weight loss velocity with weaning indicates it might--
the assumed relationship may not exist. Viewed from an
evolutionary perspective, there certainly is no apparent
selective advantage for a biological phenomenon which
constitutes a progressive and unchecked drain on the
mother. There would, however, be a very definite selec-
tive advantage in being able to support lactation without
physical depletion of the mother, especially in environ-
ments with an uncertain food supply, high risk of
infectious disease, or high work demands on females.
Given this, the lack of a strong relationship between
weight loss and breast-feeding makes sense.

Although lactation may have no significant effect on
weight loss, the prevalence of the assumption that it does
can nonetheless be explained. One explanation is related
to the simple input/output model of nutrition which persons
in this culture (and many others) have. Mothers in this
study frequently mentioned that they should be drinking
more milk than usual. While this is a suggestion fre-
quently made by dietitians and physicians, supposedly

because certain amounts of whole milk are convenient

193
measures of the daily nutritional supplement which has
been calculated as needed to support lactation (Food and
Nutrition Board 1980), this was often interpreted by

women to mean that milk input was necessary for milk

 

output. In fact, women with milk allergies or aversions
questioned their ability to successfully breast-feed.
Women also reported that they felt they should drink a
glass of water while nursing their infants, apparently to
provide the necessary liquid. An extension of this
input/output model is the belief that if a mother is
producing breast milk and not significantly increasing
food intake, the milk putput must come from her existing
body stores, and this will be reflected in weight loss.
Other ways by which it may be possible to accomodate the
energy drain of lactation are not always considered. Most
women, for example, have no notion of the possibility of
metabolic alterations accompanying lactation. Nor do they
consider the fact that motherhood may bring about re-
ductions in exercise and energy expenditure sufficient to
offset the cost of breast-feeding. For a woman used to
being active, just the act of breastffeeding may impose
restrictions on physical activity. One woman in the
present study, for example, reported daily totals of 365
to 423 minutes nurSing during the first two months postf
partum. There may also be a poor understanding of the
relative weights of various body components. While fat

may be mobilized to support lactation, it will result

194
in weight loss only if there is no corresponding increase
in the relatively heavier lean body mass. Once a mother
begins to habitually carry a growing infant, such an
increase in lean body tissue is very likely to occur.

One of the most impressive features of the postpartum
weight change data in this study is their high degree of
variability. Some women lost weight rapidly. In one case,
a mother with a pregnancy gain of 30 pounds was several
pounds below her prepregnant weight by two months post-
partum. One of the factors influencing her decision to
wean her son was her seemingly uncontrollable loss of
weight. In other cases, women actually gained weight
after the initial perinatal loss. Thus, at a population
level, some weight loss probably does accompany breast-
feeding. However, such a promise to women is meaningless
until the regulating factors of weight change and dietary

intake within individuals are better understood.

Breast-Feeding and Changes in Maternal Body Fat

The simple assumption that all women lose body fat
while breast-feeding was not confirmed by analysis of
data on midarm fat area. On the average, women increased,
rather than decreased, in upper arm fatness during the
postpartum period. There were variations in the amount of
increase, and these coincided with the degree to which
breast milk was the sole source of nutrients for the
infant. Mothers of infants in Groups lb and 2a-—all of

whom replaced breast-feeding bouts with formula--had the

 

195
greatest initial as well as cumulative increases in midarm
fat area. Variations in postpartum fatness also appeared
to be dependent on time since parturition. Weaning or
introduction of nonebreast milk food in the early months
was followed by an increase in fatness. When these same
feeding changes occurred late in the period of time
studied, they were followed by a decrease in fat.

When the relationship of breast-feeding to maternal
body fat was studied among long term breast-feeders who
were losing weight, the mean number of nursing bouts per
day in the first three months of lactation was found to
be closely linked to fat area change. Two patterns emerged.
Those with an overall decrease in midarm fat area (Group A)
nursed their infants significantly more times per day
than did those whose fat stores increased during the post-
partum (Group B). There existed a strong inverse corre-
lation of length of breast-feeding bouts with number of
bouts per day. Thus, two different nursing styles-~short,
frequent feedings and long, infrequent feedings--were
strongly associated with two patterns of change in maternal
fat stores.

Since one of the most direct effects of nursing
frequency is on serum prolactin levels, the results of
this study suggest that the concordant variation of
maternal fat may be'a function of the same hormonal system.
There are two physiological pathways through which this

might occur (Figure\L-l). The first is the mediation of

196

.umm hpom
accumumz can mocmsvmum mcflmnsz mo :oHuomumucH man MOM m>m3£uwm omNflmmnuommmIIH.> mmDon

mum>mq mzozmom umomzow»
a: Son» 5850mm scammeAull oszmaz azmnommm

mammHe mmomHo< zH ~9H>Heo<fi
mmmmHu szeommoqu

197
gonadal hormone production by prolactin. Serum prolactin
levels in nursing women are inversely proportional to
that of estrogen and progesterone (Donner and Wortman
1980). The level of progesterone, in turn, has been
found to be related to fat deposition (Hytten and Leitch
1971). Thus, women who nurse frequently (such as Group A
mothers in the present study) should be subject to greater
suppression of ovarian hormone production and therefore
less disposed to deposit fat in body stores than women
who nurse infrequently (Group B).

The second pathway linking nursing frequency and
fatness is mediated by the activity of the enzyme lipo-
protein lipase, which is responsible for the uptake of
lipids into adipose tissue. Since prolactin has been
found to induce a decrease in lipoprotein lipase activity
in peripheral adipose tissue and a corresponding increase
in that in the breast (Zinder et a1. 1974), women who
nurse frequently and have higher serum prolactin should
be less likely to increase in peripheral fatness than
women who nurse less frequently.

Thus, possible explanations can be formulated for
the observed patterns of fat change. These explanations
relate nursing behavior to variations in maternal physiof
logical parameters. Causes of these behavioral variations
and the interaction of the hypothesized hormonal patterns
with fecundity will be discussed in Chapter VI. Whether

the fat changes indicate overall changes in body composition

198
or fat patterning still needs to be investigated.

Although the interactions hypothesized in Figure V.l
seem to account for the overall trends seen in maternal
postpartum fat changes, these trends obscure a high
degree of variability. The correlation between fat change
and breast-feeding behavior was by no means perfect. Cases
were noted in which women nursed their infants only four
times per day. Yet their levels of midarm fat plummeted
without apparently extreme weight loss. As with the in-
vestigation of weight loss and breast-feeding, it is this
intra-population variability which is in many ways surpri-
sing. The participants in this study were purposely
selected by criteria designed to produce as homogeneous
a sample as possible. The fact that such variability
exists again points to the need to understand the inter-
actions of breast-feeding and maternal body changes in
individual women before expecting to understand the impact

of these interactions at the pOpulation level.

CHAPTER VI
SYNTHESIS AND CONCLUSIONS

Previous chapters have documented and discussed the
dynamics within each of three major components of infant
feeding systems involving breast feeding: infant diet,
infant growth, and maternal nutritional status. The
interactions among the three areas have been discussed
in only a limited way. This chapter will formulate a
more general biobehavioral model for infant feeding and
discuss the implications this model has for formulating
testable hypotheses for two areas of research. The first
is lactation amenorrhea and the critical body composition
hypothesis. The second is the adequacy of breast milk as

a sole, long term nutrient source for infants.

A General Model of Infant Feeding Systems
The central component in the model presented here is
the infant feeding behaviors which make up infant diet
(Figure VI.1). These include what is fed to an infant,
at what age it is fed, and perhaps most importantly, how
it is fed (e.g., the scheduling, frequency, and mode of

feeding). Some aspects of the physical environment can

199

200

have a direct effect on feeding behaviors. These include
factors which impose a physical separation on the mother-
infant pair. However, according to the results of the
present study, the most profound influence comes from the
social environment in which the nursing pair exists.
Figure V1.2 amplifies the components of this environment.
They include knowledge and beliefs about infant feeding,
growth, and behavior, as well as the entire range of
interpersonal relationships of which a mother is a part.
Together these form the basis for a mother's perception of
her infant's needs and of her options in meeting those
needs. It is this perception, then, which determines
what infant feeding behaviors will be implemented by the
mother, within the limits imposed by the physical environ-
ment. It is important to note that infant growth and
behavior affect feeding behaviors, but this relationship
is for the most part mediated by the social environment.

In an infant feeding situation where breast-feeding
is the sole source of infant nourishment, variations in
feeding behaviors affect infant growth indirectly, through
day-to-day changes in the parameters of maternal physiology
involved in milk production. The factors involved in
these changes are presented in Figure V1.3. Levels of
serum prolactin affect the quantity of milk produced and,
by regulating the level of lipoprotein lipase activity in
the breast, the composition of the milk. Prolactin

induced changes in peripheral lipoprotein lipase activity

201

.ucmfim>ow£om m>wuonp0ummu no: can .Hmflucmuom m>wuosooummu some
on mum: com: me uH .mwm3 ucmummmwp :H auwpcaomm Emma man mm: mmcflamwomwp ucmumwmfina

.mmoHmemsm Andaman: mo mumumEmumm
paw .cusouw ucmucH .uofl>mzwm maommm unnucH mo :OwuomumucH can no Hmoozlna.H> mmDon

NOOAOHmwmm Adzmmadz

\ mo 25.1555

ZOHBHmOMSOU
Mnom Adzmmadz

 

 

BzmzzomH>Zm AdUHmMmm

/ \

m0H>dmmm UZHQmmm BdeZH

 

 

 

 

 

6 \ mow>§Mm SEE / ”W

5305 9235 iv razmzzomgzm .EHoom

_

 

 

 

 

 

202

.AH.H> madman mmmv .=ucmficouw>cm Hmwoom=

mmOH><mmm UZHDmmh BZfih

 

 

mommz BdezH
mo monBmmummm Aflzmmadz
>
Aqdzmomzoz 024 A<zmomv

onsmopom qmzmmedz
v

 

a

moH><mmm oza
.mazomo .ozHommm az<mzH

BDOQ< mmmHAmm AdzmMde

 

mmmmeo .AmoHomz .quz<m

mBH3 monBU<mmBzH Aflzmmadz
l7

 

 

 

woouonwmm AT‘ _
Amzmma<z mo mmmamz<m<m

may no mucmcomEOUIIN.H> mmame

 

.3

.H> 95mg 008 .wmmoaowmmnm Hog—m: mo muwumfinmm: mo mudmconwfioulm .3 $505

\ EB;

 

2033850 A ..|

 

\ Ema 330:3

2038

SEQ mama /

ZHEmOmHH gammAlg Eam gmé

g mmmmfl \
559.6de g
_ . 5.550 Va:
283 Va: \

ZOHHUHHHHE MHHZ

III 983% GEE E

 

g Elam: NV—
MN

 

 

m\ SEEM? Va:

204

also affect milk composition, by varying the uptake of
dietary lipids into non-breast tissue and thus effectively
varying the lipids available to mammary lipoprotein
lipase. Maternal fatness and fecundity are hypothesized
to be byproducts of variations in the serum prolactin
levels, which vary themselves with different infant
feeding behaviors. According to the proposed model,
infant feeding behaviors also affect infant growth
directly in cases where non—breast milk foods are fed.
And, in addition to feeding behaviors and maternal physi-
ology of lactation, infant growth is influenced by infant
behaviors, especially through levels of energy expenditure.

Besides influencing feeding behaviors, the social
environment affects maternal physiological parameters
directly by either promoting or inhibiting the let-down
reflex. In breast-feeding situations which are supportive
and free of tension, this conditioned response to the
infant readily occurs, causing the release of oxytocin
into the blood.stream and the subsequent ejection of
milk from the alveoli. However, when the breast-feeding
context is one which is non-supportive or stressful, the -
let-down reflex is inhibited. Milk may be produced, but
without an effective letedown reflex it is unavailable to
the infant.

Three properties of this model make it a particularly
appropriate one for studying the infant feeding complex.

First, it can accommodate the full range of feeding

205
behaviors to which any infant may be subjected, not just
those associated with the lactation process. Second,
the model stresses the feedback cycle between infant
growth and feeding and the fact that this relationship
is always mediated by the social environment in which
infant feeding decisions are made. Third, it allows for
the impact that the whole range of infant feeding be-
haviors can have on maternal physiological parameters,
suggesting that observed differences in maternal post-
partum physiology may be explained, in part, by attention
to variability in infant feeding behaviors. Thus, as
social environments vary both inter- and intraculturally,
feeding behaviors, maternal physiology, and infant growth
patterns will all vary.

While the different types of data gathered in this
research support this model, other variables were not
examined which may invalidate the interactions proposed.
One of these is maternal diet. It is not known whether
variations in nutrient intake would affect the factors
entailed in milk production and in the regulation of
maternal body composition and fecundity. A second
variable not studied is the quality of the motherfinfant

interaction immediately following birth.

Lactation Amenorrhea and Maternal Bodnyat
The results of the present study have suggested
underlying patterns of change in the parameters of maternal

postpartum physiology (Figure VI.3). Further research

206
to confirm these patterns may help to resolve one current
controversy: that of the relationship of lactation
amenorrhea to maternal body fat. This controversy origi—
nated with the publication of the "critical body composition
hypothesis" by Frisch and colleagues (Frisch et a1. 1973;
Frisch and McArthur 1974). Frisch cited evidence which
she said demonstrated that a minimum percentage of body
fat is necessary for ovulation to occur. Although her
original argument dealt with menarche, it has been extended
to include amenorrhea experienced by victims of starvation,
female athletes and dancers, and lactating mothers.
Although the etiology of the failure to ovulate is vague
in Frisch's work, the hypothesis at first was widely
entertained. It appeared quite logical from the perspective
of adaptation that an organism--particularly a member of a
species which invests high levels of energy in raising only
a few offspring—-should be fecund only if able to support
a pregnancy and the offspring produced.

Nonetheless, Frisch's hypothesis has been severely
criticized on a number of grounds. First, it is charged
that the measures of body composition on which she bases
her conclusions are not adequately sensitive to pick up
the small variations in nutritional status she claims
are responsible for the presence or absence of ovulation
(Johnston et a1. 1975; Trussell 1978). Second, it is
charged that the work has used inappropriate statistics

(Trussell 1980). Finally, she has been criticized because

207
research by others fails to replicate her findings
(Huffman et al. 1978).

The results reported here suggest that female body
fat levels are a side effect of a more direct relationship
between infant feeding practices and fecundity, mediated
by the prolactin induced gonadal hormones. To substantiate
this, the following hypothesis and corollary should be
tested:

Hypothesis: Lengthening interbout intervals

will lead to increased production
of gonadal hormones and the re-
sumption of ovulation.

Corollary: If a mother's energy intake exceeds
that expended for non-lactation
activity and maintenance, ovulation
will be accompanied by increase in
her levels of body fat.

Therefore, as a side effect of hormonal change and
subject to other unrelated influences, body fat may or may
not vary with ovulation. If a mother changes breast-
feeding frequency, thereby allowing levels of serum pro-
lactin to drop and levels of the hormones necessary for
ovulation to rise, ovulation should occur. But body fat
levels will probably increase only if there is adequate
caloric intake to support this rise (see Table VI.1).
Therefore, changes in fatness with ovulation cannot be
expected to occur in all populations or even among all
women from the same population.

Thus, changes in infant feeding behavior rather than

in maternal body fat should show the better correlation

with the termination of lactation amenorrhea. While some

8
0
2

mmcmso o:

mmmmuocfl

umm atom
Hangman:

mm» mmmmuocfl mmmmuomp mucmamuwswmu Ummoxw uo: moon
mm» mmmmuOCa wmmmuomp mucwfimuwsvmu mpmmoxm

COfifiMHDer mQCOEHOE HMUMCOU CflUOMHOHm wv—MHCH UHHOHMU HMGHGUMZ

 

coflumuucwocoo Esumw

 

.mxmucH OHHOHMU mo mam>mq mcw>um> cufl3 hocmsqwum
mcflomomnummmum pommmuomo mo muommmm Andaman: owuflmmnuommmnlaquw mqmde

209
studies exist which document ovulation and feeding behavior
(Knauer 1981; Chen et a1. 1974) and others, ovulation and
maternal nutritional status (Delgado et a1. 1978; Chowdhury
1978), none has collected all three types of data. Doing
so may help to provide a satisfactory resolution of the

current debate over the critical body composition hypothesis.

The Adequacy of Breast Milk

 

The second controversy to which the biobehavioral
model proposed here is relevant concerns the nutritional
adequacy of human milk to provide for maintenance and
growth of infants. In 1979, Waterlow and Thomson published
a paper entitled, "Observations on the Adequacy of Breast
Feeding." In it they cited reports from developing
countries which indicated that infant growth frequently
faltered at about three months of age. They attempted
to show that energy requirements after three months are
such that breast milk may be insufficient to meet them.
Waterlow and Thomson's paper was not well received, since
it appeared at a time when the adequacy of breast milk was
a crucial issue in attempts to limit marketing of infant
formulas in developing countries as well as in the pro-
motion of late introduction of solids in Europe and the
United States. Jelliffe and Jelliffe (l979a), for
example, replied by citing other studies in which no
growth impairment had been reported and by claiming the

nutrient composition of human milk was indeed sufficient

210
for infant needs to six months of age.

The results of the present study suggest that breast
milk may well be nutritionally inadequate (that is, fail
to support steady weight gain) at three months in some
populations, but that this has less to do with the
requirements of infants at this specific age than with
changes in feeding habits likely to occur at about this
time. To confirm this, the following hypothesis and
corollaries can be tested:

Hypothesis: Lengthening the interbout interval
will lead to decreased quantity and
decreased fat content of breast
milk produced.

Corollary I: Infants for whom this breast milk
is a sole source of nutrients
will manifest indications of
reduced growth rates after the
interbout interval is lengthened.

Corollary II: In societies where changes in
maternal behaviors which affect
feeding frequency commonly occur
at a particular time in an infant's
life, population-wide alterations
in infant growth rates will occur
at that age.

Declining frequency of breast-feeding has been shown
in the second and third months of life. According to the
proposed model of infant feeding, two factors may be
responsible for this. One is a decrease in total milk
output and the other is a decrease in breast milk fat
content due to a reduction in mammary lipoprotein lipase
activity. Thus, in other populations where similar

patterns of decreasing feeding frequency occur, the same

infant growth trends might be expected. If these changes

211
in feeding behavior are imposed by institutionalized
patterns of maternal behavior, the trends may be even more
pronounced.

Existing studies support portions of the proposed
hypothesis and corollaries. Data from the Gambian studies
of maternal and child health and nutrition (Rowland 1978;
Rowland and Paul 1981; Whitehead et al. 1978), for example,
show that the growth of breast-fed infants begins to
falter in the third month, regardless of the time of year
at which the child reaches this age. The investigators
hypothesize that breast milk is inadequate to provide
sufficient calories at that age. But when the ethno-
graphic data which Rowland (1978) gives are examined, it
is evident that it may not be breast milk per se which
is inadequate, but the quantity and composition of that
produced by these particular mothers.

Immediately after birth, these infants are fed very
frequently-~in a true demand pattern. By three months
however, their mothers' social roles have changed. They
are expected to return to work in the fields, and they
leave the infants behind in the care of siblings (Rowland
1978). An infant is nursed thereafter on the mother's
schedule of availability. Thus, with a longer interbout
interval, milk production may decrease at this time.
Rowland and Paul's (1981) data on concurrent changes in
maternal body fat further support this line of reasoning.

They find that by 13 weeks, mothers' fatfold measures

212
begin to increase, indicating increases in body fat.
If this is indicative of increases in peripheral lipof
protein lipase activity accompanying changes in infant
feeding frequency, it also suggests the occurrence of the
hypothesized decrease in mammary lipoprotein lipase
activity and the consequent decrease in breast milk fat
content proposed in the model above. Thus, infant growth
retardation at three months of age may be the result of
concomitant changes in maternal availability for breast-
feeding which in turn causes a change in breast milk
quantity and composition.

There are indications from the literature that such
an association of maternal breast-feeding behavior and
particular ages of infants is not unusual. Other popu-
lations have similar demands placed on mothers when
infants are about three months old. Chen et a1. (1974)
note that among mothers in Bangladesh and other South
and Southeast Asian areas, conceptions are highly seasonal,
occurring shortly after the harvest. Thus, these infants
are approximately three months old at the subsequent
harvest. It is at this time that mothers devote in-
creasing amounts of time to agriculture, at the expense
of infant feeding. Although the effect on the infants is
not documented, Chen et a1. note that the change in feeding
is often sufficient to allow resumption of ovulation (1974).
Therefore, apparent inadequacies of breast milk may appear

at three months of age in these pOpulations as well.

213
Collection of the sort of data necessary to test

the proposed hypothesis may help to resolve the question
of breast milk adequacy by directing attention to sources
of variability in the milk produced by individual mothers
over the course of lactation. Consideration of pressures
exerted on such mothers by their physical and social
environments may then help to explain observed differences
in the apparent inadequacy of breast milk to provide for

satisfactory infant growth.

Conclusions

 

The original objectives of this research were to
identify major infant dietary strategies and to document
their interaction with infant growth and patterns of change
in maternal postpartum nutritional status. From the re-
sults of the analysis, several summary statements can be
made:

1. The patterns of infant feeding in the first

month of infancy are predictive of later
dietary components. The frequency of nursing
bouts in the first months of life is one of
the best predictors of dietary practices

in later infancy.

2. Infant size at birth and early growth are
associated with both early and late infant
feeding patterns. The less the deviance of
an infant from culturally approved patterns

of growth, the lower the likelihood that the

214
infant will experience a change in feeding
regime.
3. A non-breast milk food can serve as either
a replacement for breast milk or an addition
to it. Which of these two dietary roles it
takes determines the impact of such a food
on infant growth, on the course 0f breast-
feeding, and on maternal nutritional status.
4. The common assumptions of loss of weight and
fat during lactation belie the complexity of
the associations. Both the frequency and
duration of breast-feeding correlate with
observed patterns of maternal body change.
Hormonal and enzymatic action induced by
difference types of breast-feeding practices
may be the underlying modifiers of weight

and body composition in the postpartum

period.

This study was designed to limit known sources of
genetic and behavioral variability. Only low-risk,
Caucasian primiparae with healthy, normal, full term
infants were included. Nonetheless, throughout the study
the high degree of variability in both behavioral and
biological factors was obvious. That distinct relation-
ships between infant feeding strategies and infant growth

on one hand and infant feeding strategies and maternal

215
nutritional status on the other have emerged from this
indicates that in studies of other populations it will be
important to document ranges of variability as well as the
averages in order to produce a coherent picture of the

infant feeding process.

REFERENCES CITED

Adair, L.J.

1980 Biocultural determinants of infant feeding
practices and physical growth. Doctoral
dissertation, University of Pennsylvania,
Philadelphia.

Ahn, C.H. and W.C. MacLean, Jr.
1980 Growth of the exclusively breast-fed infant.
American Journal of Clinical Nutrition
33:183-192.

Aldrich, C. Anderson
1942 Ancient processes in a scientific age:
Feeding aspects. American Journal of
Diseases of Children 64:714-722.

American Academy of Pediatrics
1978 Breast-feeding. Pediatrics 62:591-601.

1980 On the feeding of supplemental foods to
infants. Pediatrics 65:1178-1181.

American Medical Association Council on Foods
1937 Strained fruits and vegetables in the
feeding of infants. Journal of the
American Medical Association 108:1259—1261.

Auerbach, K.G.
1978 A study of timing of first solid foods:
Comparison of breast- and bottlefeeding
mothers. Birth and the Family 5:27—31.

Bass, M.A., L. Wakefield, and K. Kolasa
1979 Community nutrition and individual food
behavior. Minneapolis: Burgess.

Blackburn, M.W. and D.H. Calloway
1976 Energy expenditure and consumption of
mature, pregnant and lactating women.
Journal of the American Dietetic Associ-
ation 69:29-37.

216

C.

217

Brans, Y.W., J.E. Sumners, H.S. Dweck, and G. Cassady
1974 A noninvasive approach to body composition
in the neonate: Dynamic skinfold measure-
ments. Pediatric Research 8:215-222. '

Brozek, J.
1963 Quantitative description of body composition:
Physical anthropology's "fourth" dimension.
Current Anthropology 4(1):3-39.

Bryant, C.A.

1978 The impact of kin, friend and neighbor
networks on infantile obesity. Unpublished
doctoral dissertation, University of
Kentucky, Lexington.

Charney, E., H. Goodman, M. McBride, B. Lyon, and R. Pratt
1976 Childhood antecedents of adult obesity. New
England Journal of Medicine 295:6-9.

Chen, L.C., S. Ahmed, M. Gesche, and W.H. Mosley
1974 A prospective study of birth interval
dynamics in rural Bangladesh. Population
Studies 28(2):277-297.

Chowdhury, A.K.M.A.
1978 Effect of maternal nutrition on fertility
in rural Bangladesh. In W.H. Mosley, ed.,
Nutrition and Human Reproduction. New
York: Plenum.

Clein, N.W.

1943 Streamlined infant feeding. A feeding routine
utilizing earlier addition of solid foods and
fewer feedings. Journal of Pediatrics
23:224-228.

Cunningham, A.S.
1977 Morbidity in breast-fed and artificially fed
infants. Journal of Pediatrics 90:726-729.

Davies, D.P., O.P. Gray, P.C. Elwood, C. Hopkinson, and
S. Smith
1977 Effects of solid foods on growth of bottle-
fed infants in first three months of life.
British Medical Journal ii;7-8.

Delgado, H., A. Lechtig, E. Brineman, R. Martorell, C.
Yarbrough, and R.E. Klein
1978 Nutrition and birth interval components: The
Guatemalan experiences. In W.H. Mosley, ed.,
Nutrition and Human Reproduction. New York:
Plenum.

218

Delvoye, P., M. Demaegd, J. Delogne-Desnoeck, and C. Robyn
1977 The influence of the frequency of nursing
and of previous lactation experience on
serum prolactin in lactating mothers.
Journal of Biosocial Science 9:447-451.

Dennis, K.J. and W.R. Blytheway
1965 Changes in body weight after delivery.
Journal of Obstetrics and Gynaecology
72:94-102.

deSwiet, M., P. Fayers, and L. Cooper
1977 Effect of feeding habit on weight in infancy:
Lancet i:892-894.

Dixon, W.J. and F.J. Massey, Jr.
1969 Introduction to statistical analysis. New
York: McGraw-Hill.

Egli, G.E., N.S. Egli, and M. Newton
1961 The influence of the numbers of breast
feedings on milk production. Pediatrics
27:314-317.

Evans, T.J.
1978 Growth and milk intake of normal infants.
Archives of Disease in Childhood 53:749-760.

Ferris, A.G., V.A. Beal, M.J. Laus, and D.W. Hosmer
1979 The effect of feeding on fat deposition in
early infancy. Pediatrics 69(4)397-401.

Fomon, S.J.
1974 Infant nutrition. Philadelphia: Saunders.

Fomon, S.J., L.J. Filer, T.A. Anderson, and E.E. Ziegler
1979 Recommendations for feeding normal infants.
Pediatrics 63:52-59.

Fomon, S.J., L.J. Filer, Jr., L.N. Thomas, and R.R. Rogers
1970 Growth and serum chemical values of normal
breastfed infants. Acta Paediatrica
Scandinavia Supp. 200.

Fomon, S.J., L.N. Thomas, L.J. Filer, E.E. Ziegler, and
M.T. Leonard
1971 Food consumption and growth of normal
infants fed milk-based formulas. Acta
Paediatrica Scandinavia Supp. 223.

Food and Nutrition Board -
1980 The recommended dietary allowances, ed. 9.
Washington, D.C.: National Academy of Sciences.

219

Frisch, R.E. and J.W. McArthur
1974 Menstrual cycles. Fatness as a determinant
of minimum weight for height necessary for

their maintenance or onset. Science 185:
949-951.

Frisch, R.E., R. Revelle, and S. Cook'
1973 Components of weight at menarche and the
initiation of the adolescent growth spurt
in girls: Estimated total water, lean

body weight and fat. Human Biology 45:469-
483.

Galletti, F. and A. Klopper
1964 The effect of progesterone on the quality
and distribution of body fat in the female
rat. Acta Endocrinologia 46:379-383.

Gerber Products Company
1977 Nutrient values. Gerber Baby Foods,
Professional Communications Department,
Gerber Products Company. Fremont, Michigan.

Goldsby, J.M. '
n.d. PCTL9D Anthropometry Subroutine Set.
Atlanta, Georgia: Center for Disease Control.

Green, L.W. '
1970 Manual for scoring socioeconomic status for

research on health behavior. Public Health
Reports: 85:815-827.

Grosvenor, C.E., F. Mena, and D.A. Schaefgen
1967 Effect of nonsuckling interval and duration
of suckling on the suckling-induced fall in
pituitary prolactin concentration in the
rat. Endocrinology 81:449-453.

Gussler, J.D. and L.H. Briesemeister
1980 The insufficient milk syndrome: A biocultural
explanation. Medical Anthropology 4(2):
147-174.

Gurney, J.M. and D.B. Jelliffe
1973 Arm anthropometry in nutritional assessment:
Nomogram for rapid calculation of muscle
cirCumference and cross-sectional muscle
and fat areas. American Journal of Clinical
Nutrition 26:912-915.

220

Guthrie, H., and E. Kan

1977 Infant feeding decisions--timing and
rationale. Environmental Child Health
264-266.

Himes, J.H.
1979 Infant feeding practices and obesity.
Journal of the American Dietetic Association
75:122-125.

Hodgson, S.V.
1978 Early infant feeding and weight gain.
Journal of the Royal College of General
Practitioners 28:280-281.

Holly, D., and D. Cullen
1977 A comparison of weight gain in breastfed
and bottlefed babies. Public Health
(London) 91:113-116.

Hood, L.J., J.A. Faed, P.A. Silva, and P.M. Buckfield
1978 Breastfeeding and some reasons for electing
to wean the infant: A report from the
Dunedin Multidisciplinary Child Development
Study. New Zealand Medical Journal 88:
273-276.

Huenemann, R.L.

1974 Environmental factors associated with
preschool obesity. 1. Obesity in six-month-
old children. Journal of the American
Dietetic Association 69(5):480-487.

Huffman, S.L., A.K.M.A. Chowdhury, and W.H. Mosley
1978 Postpartum amenorrhea: How is it affected

by maternal nutritional status? Science
200:1155-1157.

Hytten, F.E., and I. Leitch
1971 The physiology of human pregnancy. Oxford:
Blackwell Scientific Publications.

Hytten, F.E., and A.M. Thomson
1961 Nutrition of the lactating woman. In S.K.
Kon and A.T. Cowie, eds., Milk: The Mammary
Gland and Its Secretion, Vol. II. New
York: Academic Press.

Jelliffe, D.B., and E.F.P. Jelliffe‘
1978 The volume and composition of human milk in
poorly nourished communities. A Review.
American Journal of Clinical Nutrition.
31:492-515.

221

1979a Adequacy of breast-feeding. Lancet ii:
691-692. '

1979b Human milk in the modern world. Oxford:
Oxford University Press.

Jerome, N.W., G.H. Pelto, and R.F. Kandel
1980 An ecological approach to nutritional
anthropology. In N.W. Jerome, R.F. Kandel,
and G.H. Pelto, eds., Nutritional Anthro-
pology. Pleasantville, NY: Redgrave
'Publishing Company.

Johnston, F.E., A.F. Roche, L.M. Schell, and H.N.B.
Wettenhall
1975 Critical weight at menarche: Critique of
a hypothesis. American Journal of
Diseases of Children 129:19-23.

Kannel, W.B., and T.R. Dawber
1972 Atherosclerosis as a pediatric problem.
Journal of Pediatrics 80:544-554.

Katz, S.H.
1977 Toward a new concept of nutrition. In L.S.
Greene, ed., Malnutrition, Behavior, and
Social Organization. New York: Academic
Press.

Kemberling, S.R.
1979 Supporting breast feeding. Pediatrics
63:60-63.

Klaus, M., and J.H. Kennell
1970 Mothers separated from their newborn infants.
Pediatric Clinics of North America 17:
1015-1038.

Knauer, M. -

1981 Breastfeeding and the return of postpartum
ovulation in urban women. Paper presented
at the 50th Annual Meeting, American
Association of Physical Anthropologists,
Detroit.

Konner, M., and C. Worthman
1980 Nursing frequency, gonadal function, and
birth spacing among £Kung hunter-gatherers.
Science 207:788-791.

LaLeche International
1963 The womanly art of breastfeeding. London:
Tavistock.

222

Marlin, D.W., M.F. Picciano, and E.C. Livant
1980 Infant feeding practices. Journal of the
American Dietetics Association 77:668-676.

Meyer, H.F.
1952 Essentials of infant feeding for physicians.
Springfield: Charles C. Thomas.

1958 Breast feeding in the United States: Extent -
and possible trend. Pediatrics 22:116-121.

1968 Breast feeding in the United States. Report
of a 1966 national survey with comparable
1946 and 1956 data. Clinical Pediatrics
7:708-715.

Michigan Department of Public Health
1980 Recommendations for feeding healthy infants:
Reference for health professionals. Public
Health Paper No. 2, Lansing.

Naismith, D.J. and C. Ritchie
1976 The effect of breast feeding and artificial
feeding on bodyweight, skin fold measure-
ments and food intakes of forty-two
primiparous women. Proceedings of the
Nutrition Society 34:116A.

Newman, B.C., H. Delgado, R. Martorell, and R.E. Klein
1980 Maternal anthropometric changes during
pregnancy and lactation in rural Guatemala.
Presented at the 49th Annual Meeting,
American Association of Physical Anthro-
pologists. Niagara Falls, New York.

Philipps, C., and N.E. Johnson
1977 The impact of quality of diet and other
factors on birth weight of infants.
American Journal of Clinical Nutrition
30:215-225.

Quandt, S., and C. Ritenbaugh
1979 Variations in infant growth associated
with feeding patterns. Paper presented
at the 48th Annual Meeting, American
Association of Physical Anthropologists.
San Francisco.

Raphael, D.
1966 The lactation-suckling process in the
matrix of supportive care. Ph.D. thesis,
Columbia University. Microfilm #69-15,
580.

22 3 — r" . x . .
I - l - ‘ .-v /
[A . «’1 V 1 _,, ‘1 , ' . .- I

-\ ‘ _ tn , ._A' , - /

Ritenbaugh, C., and G. Harrison
1980 Mexican-American and Anglo infant feeding
patterns: A comparative study. Presented
at the Annual Meeting of the American
Anthropological Association. Washington,
D.C.

Ross Laboratories
1976 Boys and girls: Birth to 36 months.
Physical Growth NCHS Percentiles.
Columbus, Ohio.

Rotch, T.M.
1907 An historical sketch of the develOpment of
percentage feeding. New York Medical
Journal 85:532-537.

Rowland, M.G.M.
1978 Factors affecting lactation capacity.
Presented at Infant and Child Feeding
Symposium, Michigan State University, East
Lansing, October 16.

Rowland, M.G.M., and A.A. Paul
1981 Factors affecting lactation capacity:
Implications for developing coutries.
In J. Bond, W. Weil and G. Leveille, eds.,
Infant and Child Feeding. Academic Press.

Saarinen, U.M., and M.A. Siimes
1979 Role of prolonged breast feeding in infant
growth. Acta Paediatrica Scandinavia
68:245-250.

Sackett, W.W., Jr.
1953 Results of three years experience with a new
concept of baby feeding. Southern Medical
Journal 46(4):358-363.

Sauls, H.S.

1979 Potential effects of demographic and other
variables in studies comparing morbidity of
breast-fed and bottle-fed infants.
Pediatrics 64(4):523-527.

Schneider, M.H., and B.S. Worthington-Roberts
1981 A new look at breast-feeding. In B.S.
Worthington-Roberts, eds., Contemporary
DevelopmentsiJINutrition. St. Louis: Mosby.

224

Scow, R.O., C.R. Mendelson, O. Zinder, M. Hamosh, and
E.J. Blanchette-Mackie
1973 Role of lipoprotein lipase in the delivery
of dietary fatty acids to lactating mammary
tissue. In C. Galli, G. Jacini, A. Pecile,
eds., Dietary Lipids and Postnatal Develop-
ment. New York: Raven Press.

Shukla, A., H.A. Forsyth, C.H. Anderson, and S.M. Marwak
1972 Infantile overnutrition in the first year
of life. British Medical Journal 4:507-515.

Siegel, S.
1956 Nonparametric statistics for the social
sciences. New York: McGraw-Hill.

Sjolin, S., Y. Hofvander, and C. Hillervik
1977 Factors related to early termination of
breast feeding. A retrospective study in
Sweden. Acta Pediatrica Scandinavia
66:505-511.

1979 A prospective study of individual courses of
breast feeding. Acta Pediatrica Scandinavia
68:521-529.

Stunkard, A., E. d'Aguili, S. Fox, and R.D.L. Filion
1972 Influence of social class on obesity and
thinness in children. Journal of the
American Medical Association 221(6):
579-584.

Taggart, W.R., R.M. Holliday, W.Z. Billewicz, F.E. Hutten,
and A.M. Thomson
1967 Changes in skinfolds during pregnancy.
British Journal of Nutrition 21:439-450.

Taitz, L.S.
1971 Infantile overnutrition among artificially
fed infants in the Sheffield region.
British Medical Journal i:315-3l6.

Thomson, J.
1955 Observations on weight gain in infants.
Archives of Disease in Childhood 322-327.

Trussell, J.
1978 Menarche and fatness: Reexamination of the
critical body composition hypothesis.
Science‘200:1506-1509.

225

1980 Statistical flaws in evidence for the Frisch
hypothesis that fatness triggers menarche.
Human Biology 52(4):711-720.

Tyson, J.E.
1977 Nursing and prolactin secretion: Principal
determinants in the mediation of puerperal
infertility. In P.G. Crosignani and C.
Robyn, eds., Prolactin and Human Reproduction.
London: Academic Press.

Tyson, J.E., H.G. Friesen, and M.S. Anderson
1972 Human lactational and ovarian response to
endogenous prolactin release. Science 177:
897-900.

Vis, H.S., and P.H. Hennart
1978 Decline in breast feeding: About some of
its causes. Acta Pediatrica Belgium 31:
195-206.

Waterlow, J.C., and A.M. Thomson
1979 Observations on the adequacy of breast
feeding. Lancet ii:238-242.

Weil, W.B., Jr.
1977 Current controversies in childhood obesity.
Pediatrics 91(2):175-187.

Weiner, J.S., and L.A. Lourie (eds.)
1969 Human biology: A guide to field methods.
IBP Handbook #9. Philadelphia: F. A.
Davies Co.

Weinstein, R.S.
1981 Nutritional status and biological varia-
bility in an urban Mexican population.
Presented at the 50th Annual Meeting,
American Association of Physical Anthro-
pologists, Detroit.

Whitehead, R.G., M.G.M. Rowland, M. Hatton, A.M. Prentice,
E. Mueller, and A. Paul.
1978 Factors influencing lactation performance
in rural Gambian mothers. Lancet ii:l78-
181. '

Wilkinson, P.W., and D.P. Davies
1978 When and why are babies weaned? British
Medical Journal 1:1682-1683.

226

Zinder, O., M. Hamosh, T.R.C. Fleck, and R.O. Scow
1974 Effect of prolactin on lipoprotein lipase
in mammary gland and adipose tissue of

rats. American Journal of Physiology 226(3):
744-748.

APPENDICES

APPENDIX A

LETTERS OF INTRODUCTION AND CONSENT FORMS

MICHIGAN STATE UNIVERSITY

227

”WMWY MW‘W‘W

“WWW
“WWW
mmmnmlcm

CONSENT FORM

I, authorize the participation of myself and my infant(s)
as subjects in the postnatal portion of the project
entitled Incremental Weight Gain in hem which is being conducted by

Dr. Cheryl Ritenbaugh of Michigan State University and such qualified female
assistants as she may need and employ. I understand that these assistants will
be under her supervision at all times during the project.

 

The nature and general purpose of this part of the project have been explained

to me. The procedures to be performed (Appendix) have been fully described to
me by - . ‘

I understand that no risk to my infant's or my health and wellbeing is associated
with: our participation in this project.

Dr. Ritenbaugh or her assistants have offered to answer any inquiries I might wish
to make at any time during the course of the project.

I also authorize Dr. Ritenbaugh to obtain from my prenatal and hospital charts
and from my infant's hospital chart the information listed in the Appendix. I
understand that no other parts .of my medical records will be available to her.

I understand that at no time will any information obtained regarding me or my
infant(s) be made available to anyone not connected with the study. Any published
reports on this study will 'present data in cumulative form, with no references to
individual subjects. .

I.~.understand that I ’am free to withdraw myself and/or my infantCs) from partici-

pation in any part of this project at any time I may wish to do so without penalty
or prejudice.

I understand that nothing in the above waives any of my legal rights.

Date

 

_§ignatme of individual fir consent

 

Si gnatureTaf representative from
the study

MICHIGAN STATE UNIVERSITY 223 .

”WWWWY mLm-W-m

mumm
mammalian
MENOSMPAMCW

Appendix
DATA TO BE COLLECTED
Mother.
Monthly measurements (also at 7 days post-partum)'
weight '

minimum ankle and wrist and maximum calf circumferences
mid-arm circumference

fatfold thicknesses at triceps and subscapular locations
maximum abdominal circumference

distance pubic tubercle to xyphoid process

Bimonthly interview
24-hour dietary recall
feed habit questionnaire

Infant
Brazelton Behavioral Assessment (at 1 and 7 days)

Monthly measurements (also at l and 7 days}
weight
recumbent length
head, arm, and chest circumferences
fatfolds at triceps, biceps, and midabdomen

Materials to be completed by the mother
weekly--one 24-hour diet journal for the infant
monthly--report on eating habits, health, and activity level of infant

,At completion of the study

mother's prenatal and hospital chart
urine protein and sugar
hematocrit and hemoglobin
blood pressure
blood sugar, if available
placental weight -

infant's hospital chart
birth and daily weight
length
APGAR scores at l and 5 minutes
feeding modality (breast, bottle, supplemental sugar water, etc.

229

MICHIGAN STATE UNIVERSITY

 

DEPARTMENT OF ANTHROPOLOGY ' EAST LANSING ° MICHIGAN ' 4cm

COLLEGE or SOCIAL SCIENCE
COLLEGE OF HUMAN MEDICINE
CO“! 0! OSTBOPATHIC MEDICINE

CONSENT FORM

l. I, , authorize the participation of myself and my
infant(s§n:2elubjects in the study of third trimester pregnancy growth and
infant growth and feeding practices which is being conducted by Sara Quandt
of Michigan State University.

2. The nature and general purposes of the project have been explained to me.
and the procedures to be performed (Appendix) have been fUIly described.

3. I understand that no risk to my infant's or my health and well being is
associated with our participation in this project.

5. I understand that at no time will any information obtained regarding me
or my infant(s) be made available to anyone not connected with the study.
Any published reports will present data in cumulative form, with no
reference to individual persons.

6. I understand that I am free to withdraw myself and my infant(s) from
participation in the study at any time I may wish to do so.

7. I understand that nothing in the above waives any of my legal rights.

Date

 

 

Signature of individual giving consent

 

Investigator -- Sara Quandt

230

APPENDIX
Data to be Collected

Mother

1. Prenatal interview: plans for infant feeding, previous experience with
infants, expectations for own infant's behavior

2. Measurements biweekly during pregnancy and post partum at 2 weeks, 3 and
6 months

weight

height (one time only)

midarm and calf circumferences

triceps and subscapular fatfold thicknesses

Infant

1. Measurements at l day, 2 weeks, and monthly from l to 6 months

weight

recumbent length

midarm, head, chest, and thich circumferences

triceps, biceps, subscapula, midabdominal, and thigh fatfold thicknesses

2. One 24-hour diet diary per week

3. Questions on infant growth, behavior, and feeding to be answered by mother
at measuring sessions.

Sara Quandt

Department of Anthropology
Michigan State University
353-2950

231

MICHIGAN STATE UNIVERSITY

 

”AWOFAWMY EASTLANSNG'MICHIGAN'm

mmmm
mumume
MNWAWW

M E M 0 R A N D U M

T0: Participants in the project, Incremental Weight Gain in Pregggggy_

FROM: Cheryl Ritenbaugh, Ph.D. Assistant Professor
Sara Quandt a Debbie Gregg. Research Assistants

DATE: November 15. 1977

When we wrote the first memorandum asking your help in the study of weight gain in
pregnancy, we indicated that we intended to continue monitoring your’weight and
that of your infant for a short time after delivery. Since that time we have re-
ceived funding for a longer and more detailed follow up of the infants. In this
letter we would like to tell you more about this portion of the project and ask
your continued participation.

Previous studies of weight gain in pregnancy have ended at delivery. _Those studies
which have been conducted on infant development have begun at birth with little
information on the preceding pregnancy. Since we have collected such detailed data
on the prenatal progress of you and other women, we are in the unique position of
being able to study the correlation between this progress and the early growth and
development of infants. To do this, we would like to extend the postnatal portion
of the project to up to 6 months after delivery. During this period we are inter-
ested in gathering data on the normal body adjustments of mothers after birth and
on the growth, eating habits, and activity levels of their infants.

One part of the postnatal study mentioned in the first memorandum'was an evaluation
of your infant's behavior at about a week after delivery. This would involve using
a safe, nationally-used rating scale which would be explained to you in detail
before asking your permission to do it. Also, we would like to weigh and measure
you and your baby once a month. The measurements are similar to those used during
your pregnancy, and the new ones for your baby will be explained to you. On severa
occassions we would like to do a diet recall on a 24-hour period followed by a shor'
questionnaire on your eating habits. Since it will be important to us to know what
your baby is eating, we will provide you with a form on which you can record at you‘
convenience what your baby is eating during one 24-hour period each week. If’you
are breast feeding, we would want you to record the time and length of each nursing
session for that day.

we are interested in having you and your infant continue with us in this project.

If for some reason you will be unable to do so for the full 6 month period and
would still like to help, we would be most grateful fbr any participation possible.

Jr

232

MICHIGAN STATE UNIVERSITY

 

DEPARTMENT OF ANTHROPOLOGY EAST LANSING ' MICHIGAN ' 48814

COLLEE Of SOCIAL SCIENCE
COLLEGE OF HUMAN MEDICINE
COLmE OF OSTEOPATHIC MEDICINE

MEMORANDUM

To: Potential participants in study of third trimester pregnancy growth and
infant feeding and growth

From: Sara Quandt, Graduate Research Trainee, Department of Anthropology

Date: July 25, l979

I am a biological anthropologist with a strong interest in human growth and
nutrition. Currently I am conducting a study which involves pregnant

women in the Lansing area. This letter is to explain the project and to ask your
help in its completion.

During a study ‘I helped to conduct last year on weight gain during pregnancy
and its effect on infant growth, I became quite interested in two different
topics. The first is the variability of third trimester weight and body changes

in pregnant women and their possible relation to factors such as onset of labor
and sex of the child. The second is the interaction of nutrition and growth and
the way in which mothers decide on a feeding plan for their infants. When I began
reading the scientific literature on this subject, I discovered that most previous
studies have looked at infants only in terms of whether bottle or breast-fed.

They have generally ignored other factors such as the frequency of feeding and

the timing of introduction of solids. For this reason, it is hard to apply the
results of these studies to other infants. I think there are probably many
different patterns of feeding which lead to healthy babies. What I would like

to do is to find out what the variability in infant feeding is and how that relates
to growth. Additionally, I would like to learn about the ways mothers decide
among various ways to feed their infants.

You and the other women who are being asked to participate in this study were
chosen from healthy women having their first babies under the care of several
obstetricians in the Lansing area who are cooperating with me. Your parti-
cipation would involve, first,an interview before your baby is born so I could
find out your feelings on infant feeding and you could get to know a little more
about me and this study. Also, I would want to weigh you and do a few measurements
periodically before delivery. After delivery, I would measure your daughter

or son at regular intervals--about every four weeks--until age 6 months. The
measurements would include weight, length, and several others which would each

be explained to you. When measuring, I would also want to ask a few questions
about any changes you had made in your infant‘ s feeding, as well as general
questions about how things were going. The measuring would take place at your
home--unless you preferred that we met elsewhere--and would take 30 minutes or
less. Since it would be important to know what your baby was eating, I would
provide you with a form on which to record at your convenience what your baby ate
during one 24- hour period every week.

I will be calling you within the next few days to answer any questions you

m-I-‘LL L.... L-..- .411 L- 4-4-ana-I-All «In lent-tune. an {nunIuae undid-In m In +511-

APPENDIX B

DISTRIBUTION OF MOTHERS RETURNING INFANT DIET DIARIES

APPENDIX B

DISTRIBUTION OF MOTHERS RETURNING INFANT DIET DIARIES

Figure B.l shows the distribution of number of
diaries returned by mothers during the study. For those
women who dropped out of the study before six months, a
projected number of returned diaries is entered. This is

based on rate of return before leaving the study.

 

0 2 4 6 8 10 12 14 16 18 20 22 24

Number of Diaries

FIGURE B.l--Distribution of Mothers Returning Diet
Diaries.

233

APPENDIX C

INFANT OUTLINES USED IN DATA COLLECTION

APPENDIX C

INFANT OUTLINES USED IN DATA COLLECTION

The six infant outlines on the following pages (Figure
C.1) were used to elicit maternal ideas of appropriate and
inappropriate body shape. Each was traced on a blank card.
These were then shuffled and presented as a group to the

mothers.

234

235

 

.mmcflauso ucmwcHIIH.U mmDUHm

 

236

.Ao.u:ooc H.o mmDon

 

APPENDIX D

REASONS FOR INTRODUCING NEW FOOD TO A

BREAST MILK DIET

APPENDIX D

REASONS FOR INTRODUCING NEW FOOD TO A

BREAST MILK DIET

Infant Behavior

 

ID#

48 B3 had slept through the night at an
earlier age; had now stopped sleep-
ing through the night.

51 M wants to tide the baby over to
four hour feedings. B currently
wants to nurse too frequently.

57 B couldn't go long enough without
nursing. Kept crying, making
sucking motions with mouth, want-
ing more food.

58 B not satisfied on breast milk.
Kept crying, making sucking
motions with mouth, wanting more
food.

61 B seemed to want to eat all the

time. Because of hot, humid
weather this annoyed M.

lC=cereal; Fo=formula; Fr=fruit

Foodl/=+2/Month

Fr + 4

2"+" represents addition; "=" represents replacement

3B=baby; M=mother

237

62

66

68

72

73

74

80

81

238

B was crying "constantly." M said, Fo
"Not satisfied on breast milk."

M thought it would get infant to F0
sleep through the night.

B wanted to nurse frequently--at C
one to two hour intervals day

and night.

B crying a lot, colicky, wanted F0

to nurse "all the time."

B fussy in the evening, breast milk C
did not pacify.

B waking one or more times during C
the night.
M trying to get baby to sleep C

through the night.

B not satisfied after nursing. Fr
M thought breast milk not enough.
B fussed after feedings.

External Suggestion

 

50

53

54

55

Pediatrician told M to start cereal C
because infants that age need iron
not obtainable from breast milk.

Pediatrician noted insufficient Fo
infant weight gain; said M's breast

milk supply "inadequate" to provide

needed calories.

Pediatrician told mother infants C
that age need additional iron; M

could use cereal or iron drops

(told her certain ill effects

possible from drops).

Pediatrician told mother infants C
that age need additional iron; gave

her choice of cereal or iron drops

(told her certain ill effects

possible from drops).

67

69

76

77

239

M had been receiving increasing

pressure from mother-in-law, who
thought the baby was "old enough
to need real food."

M reports feeling pressure from
husband's family, particularly

mother-in-law who thought baby

"so big he needs food? mother-
in-law also doubted whether the
daughter-in-law‘s small breasts
could supply enough breast milk.

Doctor told M he thought the baby
was "big enough" to need more
iron than breast milk could
provide.

Doctor told M that the baby needed
iron now and that breast milk did
not contain any.

Situational

 

52

64

65

70

71

M returning to work (teaching);
needed to have something for baby-
sitter to feed infant during day.

M returning to work (nurse in
hospital); wanted to get infant
used to food so the babysitter
could feed him during the day.
M planned to nurse at night.

M returning to work (teaching);
wanted to be the person to introduce
infant to solid foods which she
would eat while at the babysitter's.

M going back to graduate school classes;
needed something for father to feed
infant to tide him over until M re-
turned home to nurse him.

M returning to work (secretarial);
was satisfied she had given infant

a good nutritional start; now wanted
to wean so babysitter could keep
infant content with formula during
the day.

F0

F0

Fr

Fo

Mater

59

63

79

240
nal Health

M hospitalized at bout one week post-
partum for hemorrhage, D & C; tried
to keep nursing baby, but found milk
supply affected by anesthetic;
hospital staff impeded attempts to
have infant in room with mother.

M hospitalized for appendectomy at

2 weeks postpartum; was not allowed
to nurse because of possible effects
of drugs on infant; milk supply had
diminished by the time allowed to
nurse, so continued with formula
infant had received while she was in
hospital.

M reported breasts engorged, nipples
too sore to nurse.

Other

49

75
78
56

60

B was beginning to display an interest
in food; picking up parents' food and
putting it in mouth; mother wanted

to start solids while B was interested
to avoid having a picky eater later in
life.

M could not articulate reason.
M could not articulate reason.
B exclusively breast fed.

B exclusively breast fed.

F0

F0