Jaw,
Writ?
. a

ﬂu.
9R9:

:17
x

a
.V

an».

34»...
and

has“.

 

E .. . w
1 .3 \.

45:11. . f
7311.?

a?
e {anniha
if .. .3 .
1. 1- A .

30:24...”

.5 . :r
r. ‘31.}...3.‘ A

$¢3~33£
.3 .ﬁ.‘

f , 31...: .

~ ..

 

.. J .3? a Fﬂztmwmwugar ., $.54 , . ‘ 52...}: ‘ .35. V:
. ‘ ‘ . .T. as. ”aw.

lo 06 LIBRARY
Michigan State
University

This is to certify that the
dissertation entitled

VALIDATION OF A HISTORICAL PHYSICAL ACTIVITY RECALL
TOOL; THE EFFECTS OF PAST PREGNANCY PHYSICAL
ACTIVITY ON CURRENT PHYSICAL ACTIVITY, BARRIERS TO
PHYSICAL ACTIVITY, AND BODY SIZE

presented by

Patricia W. Bauer

has been accepted towards fulﬁllment
of the requirements for the

degree' In Kinesiology

 

[3%er [/2 (ﬁg/ml

Major Professor's Signature
08/25/05

Date

MSU is an Aﬁinnative Action/Equal Opportunity Institution

 

PLACE IN RETURN BOX to remove this checkout from your record.
To AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATE DUE

DATE DUE

DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

2/05 czrcfncrommo" Km '15‘

 

VALIDATION OF A HISTORICAL PHYSICAL ACTIVITY RECALL TOOL;
THE EFFECTS OF PAST PREGNANCY PHYSICAL ACTIVITY ON CURRENT
PHYSICAL ACTIVITY, BARRIERS TO PHYSICAL ACTIVITY, AND BODY
SIZE
By

Patricia W. Bauer

A DISSERTATION
Submitted to
Michigan State University

in partial fulﬁllment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY
Department of Kinesiology

2005

ABSTRACT
VALIDATION OF A HISTORICAL PHYSICAL ACTIVITY RECALL TOOL; THE
EFFECTS OF PAST PREGNANCY PHYSICAL ACTIVITY ON CURRENT
PHYSICAL ACTIVITY, BARRIERS TO PHYSICAL ACTIVITY, AND BODY SIZE
By
Patricia W. Bauer

Physical activity (PA) is a key element of health promotion and disease
prevention. Historically, PA was not universally recommended for pregnant women due
to limited research and fear of unknown risks to the mother and fetus. However, recent
studies have shown PA to have positive effects on a healthy pregnancy, as well as the
postpartum period. Methodology for assessing PA has not been consistent among
studies; therefore comparing results has been difﬁcult. Current research has focused on
an individual’s ability to recall historical PA, and relate this behavior to current disease
states. Using historical recall to assess past pregnancy PA could assist with determining
the relationship between pregnancy and long-term weight changes.

This investigation involved a six-year follow-up of a cohort Of women who
participated in a previous research study (Maternal Activity Measurement and
Assessment study). F ifty-six women had their PA energy expenditures quantiﬁed
carefully throughout pregnancies from between 1997-1999. Thirty of these women
agreed to participate in follow-up research. The purposes of this study were to evaluate
the ability of women to recall their PA during three time points previously assessed
during a pregnancy six years ago (20 weeks gestation, 32 weeks gestation, 12 weeks
postpartum); and to examine the effects of past pregnancy PA on current body size, PA

levels, and barriers to exercise. The Modiﬁable Activity Questionnaire (MAQ) was used

to assess current and past pregnancy PA. The Perceived Barriers Efﬁcacy Questionnaire
(PBEQ) was used to assess current barriers to PA and percent conﬁdence in overcoming
those barriers.

Results showed the MAQ values to be signiﬁcantly related to the original physical
activity recall (PAR) values at all time periods of interest. Correlations from this study
are similar (r=0.57-0.86) to those found in previous PAR and MAQ validation studies.
The women’s ability to recall past pregnancy PA did not differ based on current PA.

The median value for postpartum weight retention (PPWR) was 2.8 kg with a
wide individual range. When separated into current PA level groups (2 or S 20
MET/wk), the more active women showed signiﬁcantly lower weight, BMI and %fat,
and PPWR values compared to the less active women. A signiﬁcant inverse relationship
was found between current PA and BMI (r=-O.55) and %fat (r=-0.70). Past pregnancy
PA was found to be correlated with current PA at all time points of interest [20 weeks
gestation (r=0.49), 32 weeks gestation (r=0.7l), and 12 weeks postpartum (FO.73)].
Further analysis showed current PA to be the proximate cause of current body size;
though it is likely past pregnancy PA behaviors may inﬂuence current PA levels. The top
three barriers (and their frequencies) were time (22), motivation (7), and childcare (6).
Average percent conﬁdence score for overcoming these barriers was 65.6 % (i245).

Overall, the MAQ was found to be an accurate PAR for past pregnancy PA.
Sixty percent of women studied met current CDC/ACSM PA guidelines. Both recalled
and current PA were shown to be related to percent conﬁdence in overcoming barriers to

PA. Future studies should use larger heterogeneous samples and include longer follow-

up.

TABLE OF CONTENTS

LIST OF TABLES ................................................................................................ v

LIST OF FIGURES ............................................................................................... vii

KEY TO ABBREVIATIONS ................................................................................ ix

CHAPTER 1 INTRODUCTION ........................................................................... 1

References ............................................................................................................ 5

CHAPTER 2

LITERATURE REVIEW FOR PART I
Introduction ............................................................................................... 7
Physical activity and chronic disease ......................................................... 7
Physical activity measurement ................................................................... 10
PAR validity and reliability studies ............................................................ 11
Historical PARS ......................................................................................... 15
The Modiﬁable Activity Questionnaire ...................................................... 19
Summary ................................................................................................... 20
References ................................................................................................. 22

CHAPTER 3

VALIDATION OF A HISTORICAL PHYSICAL ACTIVITY

RECALL TOOL
Introduction ............................................................................................... 25
Methods ..................................................................................................... 27
Results ....................................................................................................... 31
Discussion ................................................................................................. 32
Future directions ........................................................................................ 35
References ................................................................................................. 37

CHAPTER 4

LITERATURE REVIEW FOR PART 11
Historical background ................................................................................ 39
Early animal studies ................................................................................... 40
Early human studies ................................................................................... 42
More recent human studies ........................................................................ 45
Postpartum weight retention reviews .......................................................... 52
PPWR studies ............................................................................................ 54
PPWR, gestational weight gain and race .................................................... 55
PPWR and lactation ................................................................................... 57
PPWR and physical activity ....................................................................... 61
Pregnancy and postpartum effects on lifestyle behaviors ............................ 67

iv

Self-Efﬁcacy and physical activity behaviors ............................................. 68

Self-Efﬁcacy and physical activity during and after pregnancy .................. 69

Summary ................................................................................................... 71

References ................................................................................................. 74
CHAPTER 5

THE EFFECTS OF PAST PREGNANCY PHYSICAL ACTIVITY
ON CURRENT PHYSICAL ACTIIVTY, BARRIERS TO PHYSICAL

ACTIVITY, AND BODY SIZE
Introduction ............................................................................................... 80
Methods ..................................................................................................... 88
Results ....................................................................................................... 94
Discussion ................................................................................................. 97
Future directions ........................................................................................ 102
References ................................................................................................. 104

APPENDIX A

ANTHROPOMETRIC EQUATIONS ................................................................... 133

APPENDIX B

THE MODIFIABLE ACTIVITY QUESTIONNAIRE .......................................... 134

APPENDIX C

THE PERCEIVED BARRIERS EFFICACY QUESTIONNAIRE ......................... 138

APPENDIX D

INFORMED CONSENT ....................................................................................... 139

LIST OF TABLES

Table 1. Descriptive statistics (MiSD) of total sample, and grouped
by current physical activity (220 MET/wk) ........................................................... 108

Table 2. Average physical activity values (MiSD) at all time
periods of interest (20 and 32 weeks gestation, 12 weeks
postpartum, and current) ........................................................................................ 109

Table 3. One-way AN OVA of recall ability at the past pregnancy
time points of interest (20 and 32 weeks gestations and 12 weeks
postpartum) by current PA group (220 MET/wk) ................................................. 110

Table 4. Correlations between measured (PAR) and recalled (MAQ)

past pregnancy PA and current PA at all three time periods of interest

(20 and 32 weeks gestation and 12 weeks postpartum) .......................................... 111
Table 5. Listed current barriers to exercise in rank order ....................................... 1 12
Table 6. Correlations between percent conﬁdence scores and MAQ

(recalled) energy expenditures at time periods of interest (20 and 32
weeks gestation and 12 weeks postpartum) ........................................................... 113

vi

LIST OF FIGURES

Figure 1. Correlations (r) between past measured (PAR) and
recalled (MAQ) physical activity at 20 weeks gestation ......................................... 1 14

Figure 2. Correlations (r) between past measured (PAR) and
recalled (MAQ) physical activity at 32 weeks gestation ......................................... 115

Figure 3. Correlations (r) between past measured (PAR) and
recalled (MAQ) physical activity at 12 weeks postpartum ..................................... 116

Figure 4. Correlations (r) between past measured (PAR) physical
activity and BMI at 20 weeks gestation ................................................................. 117

Figure 5. Correlations (r) between past measured (PAR) physical
activity and BMI at 32 weeks gestation ................................................................. 118

Figure 6. Correlations (r) between past measured (PAR) physical
activity and BMI at 12 weeks postpartum .............................................................. 119

Figure 7. Correlations (r) between past measured (PAR) physical
activity and percent body fat at 20 weeks gestation ................................................ 120

Figure 8. Correlations (r) between past measured (PAR) physical
activity and percent body fat at 32 weeks gestation ................................................ 121

Figure 9. Correlations (r) between past measured (PAR) physical
activity and percent body fat at 12 weeks postpartum ............................................ 122

Figure 10. Correlations (r) between past recalled (MAQ) physical
activity and BMI at 20 weeks gestation ................................................................. 123

Figure 11. Correlations (r) between past recalled (MAQ) physical
activity and BMI at 32 weeks gestation ................................................................. 124

Figure 12. Correlations (r) between past recalled (MAQ) physical
activity and BMI at 12 weeks postpartum .............................................................. 125

Figure 13. Correlations (r) between past recalled (MAQ) physical
activity and percent fat at 20 weeks gestation ........................................................ 126

Figure 14. Correlations (r) between past recalled (MAQ) physical
activity and percent body fat at 32 weeks gestation ................................................ 127

vii

Figure 15. Correlations (r) between past recalled (MAQ) physical
activity and percent body fat at 12 weeks postpartum ............................................ 128

Figure 16. Correlations (r) between percent conﬁdence and past physical
activity, both measured (PAR) and recalled (MAQ) at 20 weeks gestation ............ 129

Figure 17. Correlations (r) between percent conﬁdence and past physical
activity, both measured (PAR) and recalled (MAQ) at 32 weeks gestation ............ 130

Figure 18. Correlations (r) between percent conﬁdence and past physical
activity, both measured (PAR) and recalled (MAQ) at 12 weeks postpartum ......... 131

Figure 19. Visual representation of the past physical activity and current
body size relationship ............................................................................................ 132

viii

ACOG
ACSM
BIA
BMI
CDC
EE
EKG
FFM
IAQ
IOM
LBW

LTPA

MAQ
MET
NIDDM
PA
PAR
PBEQ
PPWR
SAQ

SE

KEY TO ABBREVIATIONS

American College of Obstetricians and Gynecologists
American College of Sports Medicine
Bioelectrical Impedance Analysis

Body Mass Index

Center for Disease Control and Prevention
Energy Expenditure

Electrocardiogram

Fat Free Mass

Interviewer Administered Questionnaire
Institute of Medicine

Low Birth Weight

Leisure Time Physical Activity

Maternal Activity Measurement and Assessment
Modiﬁable Activity Questionnaire
Metabolic Equivalent

Non-Insulin Dependent Diabetes Mellitus
Physical Activity

Physical Activity Recall

Perceived Barrier Efﬁcacy Questionnaire
Postpartum Weight Retention
Self-Administered Questionnaire

Self-Efﬁcacy

ix

SES Socioeconomic Status
SPAWN Stockholm Pregnancy and Women’s Nutrition

WIC Women Infant and Children program

CHAPTER I
Introduction

Background

Physical activity (PA) signiﬁcantly impacts health promotion and disease
prevention. Beneﬁts of a physically active lifestyle include reduced risk of chronic
disease, increased overall physical and mental well being, and enhanced weight
management (Surgeon General’s Report, 1996). The Centers for Disease Control and
Prevention (CDC) and the American College of Sports Medicine (ACSM) suggest 30
minutes of moderate activity most days of the week for health beneﬁts (Pate et al., 1995).
The American College of Obstetricians and Gynecologists (ACOG) recently updated
their guidelines for exercise during pregnancy to incorporate these CDC/ACSM PA
recommendations (ACOG, 2002).

Historically, PA was not universally recommended for pregnant women due to
limited research and fear of unknown risks to the mother and fetus. In recent decades,
research has Shown exercise during pregnancy should rarely be contraindicated, but
rather, may be beneﬁcial for both mother and child (Johnston, 1991; Clapp, 2000;
Larsen-Meyer, 2002; Wang, 1998). Maternal beneﬁts include improved cardiac function,
limited weight gain and fat retention, improved mental state, improved ﬁtness, and less
complicated labor. Fetal beneﬁts may include improved stress tolerance and enhanced
neurobehavioral maturation (Clapp, 2000).

Proposed beneﬁts of PA during the postpartum period include reduction of
excessive weight retention/gain, improvements in aerobic ﬁtness and insulin sensitivity,
increased bone mineral density, and enhanced psychological well-being (Larsen-Myer,

2002). Data from a longitudinal study of factors that affect weight changes during and

after pregnancy support the notion that return to pre-pregnancy weight is most likely to
occur in women who implement healthy exercise habits (Ohlin and Rossner, 1996). The
prevalence of physical inactivity is high among culturally diverse segments of the US.
population including low-income groups, the elderly, and women (Surgeon General’s
report, 1996). Life events unique to pregnancy and the postpartum period may place
women at greater risk for decreased PA compared to men (Downs and Hausenblaus,
2004)

During the postpartum period women may initiate lifestyle changes with a
renewed focus on weight loss. Conversely, changes in activity behavior toward a more
sedentary lifestyle due to family obligations may result in higher weight retention after
pregnancy. Thus, weight retained from a pregnancy can remain indeﬁnitely and may
increase a woman’s risk of becoming overweight or obese (Greene et al., 1988; Rookus,
1987; Rooney and Schauberger, 2002). The relationship between pregnancy and body
weight changes is complex and not well-understood. It is affected by lifestyle factors
before and after pregnancy such as eating behavior, PA, smoking, and breastfeeding
status (Linne, 2002). Methodology for assessing these variables, especially PA, has not
been consistent among studies; therefore comparing results can be difﬁcult.

PA can be measured in many different ways in ﬁeld studies including, heart rate
monitoring, accelerometry, and physical activity recall (PAR) (Montoye and Taylor,
1984). Each method has advantages and disadvantages with respect to cost, reliability,
and practicality with large study samples. Valid assessment of PA during pregnancy and
the postpartum period is essential for presenting and interpreting ﬁndings from studies of

the role of PA on weight gain and retention. To better obtain more complete data, PAR

tools are now being used to assess historical or lifetime PA. Studies using historical PAR
can give insight into the effect of lifetime or past PA on current disease status or risk.
However, there is strong reliance on a participant’s ability to recall past information with
historical PA studies and true validation is difﬁcult because in most cases, past PA was
not measured at the time it was performed.

PA participation is a lifestyle factor that can be a target for intervention studies
aimed at postpartum weight retention (PPWR). Self-efﬁcacy (SE) is one of many factors
that can affect PA participation. SE beliefs are deﬁned judgements concerning a person’s
ability to perform speciﬁc tasks in the face of obstacles to produce a given outcome
(McAuley, 1992). SE beliefs for exercise are important due to their signiﬁcant role in a
person’s decision, effort, and persistence to continue to exercise (Feltz, 1988). For
example, Miller et a1. (2002) found that SE affected the PA behaviors of mothers with
small children. The connections between SE factors, PA, and PPWR have not been
directly addressed. New information will help health care and exercise professionals to
determine the best intervention strategies to address the changes in lifestyle factors that
occur during the child-bearing years.

Study Summary

This investigation involves a six-year follow-up of a cohort of women who
participated in a previous research study (Maternal Activity Measurement and
Assessment study [MAMA]) (Pivamik et al., 2002; Stein et al., 2003). Fifty-six women
were studied meticulously throughout pregnancy, and their PA behaviors were carefully

assessed. Thirty of these women agreed to participate in follow-up research.

This dissertation is divided into two parts, each consisting of a literature review
and a manuscript. Some parts of the dissertation, especially the methods sections, will be
somewhat repetitive, but Should assist in reading comprehension.

Part I focuses on the ability of a woman to recall PA that occurred during a
pregnancy six years ago. Current PAR tools allow researchers to assess historical PA
levels. The strength of this portion of the dissertation is that we can assess the validity of
one such tool by comparing the study participants’ recall to actual data collected during
the pregnancy ﬁve years ago. Part I includes chapters 2 and 3. Chapter 2 is a literature
review regarding physical activity, and recall using current and historical questionnaires.
Chapter 3 consists of the introduction, methods, results, discussion points, and future
directions for the present study.

Part II focuses on assessing current PA behaviors and anthropometﬁcs of the
study participants. New ﬁndings are compared to data collected previously during the
original MAMA study, while also addressing barriers that may affect current PA
behaviors. Part 11 includes chapters 4 and 5. Chapter 4 is a literature review regarding
physical activity during pregnancy and postpartum weight retention. Chapter 5 consists
of the introduction, methods, results, discussion points, and future directions for the

present study.

References

American College of Obstetricians and Gynecologists (2002) Exercise during pregnancy
and the postpartum period. Technical Bulletin 267. Washington, DC: ACOG Press.

Clapp J (2000) Exercise during pregnancy: a clinical update. Clinics in Sports Medicine
19: 273-85.

Downs D, and Hausenblas H (2004) Women’s exercise beliefs and behaviors during their
pregnancy and postpartum. Journal of Midwifery and Women ’s Health 49: 138-44.

F eltz D (1988) Self-conﬁdence and sports performance. Exercise and Sport Science
Reviews 16: 423-57.

Greene G, Smiciklas-Wright H, Scholl T, and Karp R (1988) Postpartum weight change:
how much of the weight gained in pregnancy will be lost after delivery? Obstetrics and
Gynecology 71: 701-7.

Johnston E (1991) Weight changes during pregnancy and the postpartum period.
Progress in Food and Nutrition Science 15: 117-57.

Larsen-Myer D (2002) Effect of postpartum exercise on mothers and their offspring: a
review of the literature. Obesity Research 10: 841-53.

Linne Y, Barkeling B, Rossner S (2002) Long-term weight development after pregnancy.
Obesity Reviews 3: 75-83.

McAuley E (1992) The role of efficacy cognitions in the prediction of exercise behavior
in middle-aged adults. Journal of Behavioral Medicine I 6: 103-13.

Miller Y, Trost S, and Brown W (2002) Mediators of physical activity behavior change

among women with young children. American Journal of Preventative Medicine 23: 98-
1 03.

Montoye H, and Taylor C (1984) Measurement of physical activity in population studies:
a review. Human Biology 56: 195-216.

Ohlin A, and Rossner S (1996) Factors related to body weight changes during and after
pregnancy: the Stockholm Pregnancy and Weight Development Study. Obesity Research
4: 271-6.

Pate R, Pratt M, Blair S, Haskell W, Macera C, Bouchard C, Buchner D, Ettinger W,
Heath G, King A, et al. (1995) Physical activity and public health. A recommendation
from the Centers of Disease Control and Prevention and the American College of Sports
Medicine. JAMA 273: 402-7.

Physical activity and Health: A report of the Surgeon General (1996) Atlanta: US.
Department of Health and Human Services, Centers for Disease Control and Prevention,
National Center for Chronic Disease Prevention and Health Promotion.

Pivarnik J, Stein A, Rivera J (2002) Effect of pregnancy on heart rate/oxygen
consumption calibration curves. MSSE 34: 750-5.

Rookus M, Rokebrand P, Burema J, and Deurenberg P (1987) The effects of pregnancy
on the body mass index 9 months postpartum in 49 women. IntJ of Obesity 1]: 609-18.

Rooney B, and Schauberger C (2002) Excess pregnancy weight gain and long-term
obesity: one decade later. Obstet and Gynecol 100: 245-52.

Stein A, Rivera J, Pivarnik J (2003) Measuring energy expenditure in habitually active
and sedentary pregnant women. MSSE 35 : 1441-6.

Wang T, and Apgar B (1998) Exercise during pregnancy. American Family Physician
57: 1846-51.

CHAPTER 2
Validation of an historical physical activity recall tool

Introduction

The Centers for Disease Control and Prevention (CDC) and the American College
of Sports Medicine (ACSM) recommend that every American participate in moderate
physical activity (PA) for greater than or equal to 30 minutes on most if not all days of
the week (Pate et al., 1995). According to Healthy People 2010, PA is the number one
leading health indicator targeted for improvement, due to its relationship with all the
major chronic diseases ( www.healhtygeoﬁrie2010. gg). Physical activity is deﬁned as
any body movement produced by contraction of skeletal muscle that results in energy
expenditure (Casperson, Powell, and Christenson, 1985). According to Casperson et a1,
everyone performs PA in order to sustain life; however, the amount is mostly dependent
on personal choice and can vary greatly between different people as well as for the same
person over time (Casperson, Powell, and Christenson, 1985). In a 1984 review by
Montoye and Taylor of current “state of the art” assessments of habitual PA, the authors
stated that physical activities can be divided into those performed on the job
(occupational) and those performed during leisure time (Montoye and Taylor, 1984). The
authors believed that during the latter part of the twentieth century, leisure time activity
Should be considered the most important source of energy expenditure beneﬁcial to long
term health. Leisure time physical activity (LTPA) was labeled as the only volitional, off
the job pursuit that differentiates people on the basis of PA (Montoye and Taylor, 1984).
Physical Activity and Chronic Disease

In the 1980’s, results from two large longitudinal studies evaluating PA and

chronic disease (The Harvard Alumni and The F ramingham Study), were published.

These large longitudinal cohort studies continue to follow groups of people for 20 years
or more and assess the effects of PA on cardiovascular health as well as mortality and
morbidity (Paffenbarger et al., 1984; Paffenbarger et al., 1986; Kannel et al., 1986).
These studies have played a strong role in our current understanding of long-term PA
beneﬁts.

In 1986, Paffenbarger et a1. published a paper evaluating the role of PA in all-
cause mortality and longevity in college alumni. Men who entered Harvard University
through the years 1916-1950 were recruited to participate. Personal and lifestyle
characteristics data were recorded for over 16,000 men, with follow-up data on cause of
death and estimated length of life also collected (Paffenbarger et al., 1986). The authors
found that death rates declined steadily as energy expended on activity increased from
500-3500 kcal per week, beyond which rates increased slightly. Death rates were one
quarter to one third lower among alumni expending 2000 or more kcal per week than
among those who were sedentary. In general it was concluded that alumni mortality rates
were signiﬁcantly lower among the physically active (Paffenbarger et al., 1986).

The F ramingham Study began in 1948, following a cohort of 520 people from a
community in Massachusetts biennially to track the development of cardiovascular
morbidity and mortality. In 1986, Kannel et a1. published a paper from the F ramingham
study, examining physical activity and physical demand on the job and risk of
cardiovascular disease and death (Kannel et al., 1986). The overall results showed active
persons, as assessed by a 24-hour physical activity index, lived longer and suffered less
cardiovascular mortality. The cardiovascular and coronary heart disease mortality

improved with increasing level of PA at all ages, including the elderly (Kannel et al.,

1986)

In a summary of a workshop on public health aspects of PA and exercise, Powell
and Paffenbarger (1987) noted the beneﬁcial effects of PA were becoming more apparent
as the amount of research increased during the 1980’s. Ten scientiﬁc papers regarding
PA and health were presented and conclusions about the beneﬁts of exercise were made.
Topics ranged from epidemiologic research issues regarding physical activity to disease-
speciﬁc beneﬁts and risks, mental health, and promotion. Based on the presentations, it
was concluded that while physically active individuals were at lower risk for coronary
heart disease than less active persons, PA is a complex behavior for which there is no
standard measurement (Powell and Paffenbarger, 1987; Casperson, Powell, and
Christenson, 1985; LaPorte, Montoye, and Casperson, 1985: Siscovick, LaPorte, and
Newman, 1985; Iverson et al., 1985).

The Nurses Health Study was a large cohort study initiated in 1976, designed to
evaluate health behaviors of female registered nurses. Participants initially completed a
mailed questionnaire and every two years follow-up questionnaires were sent to obtain
updated information on potential risk factors and identify newly diagnosed cases of
coronary heart disease or other illnesses (Manson et al., 1999). In a prospective study by
Manson et a1. (1999), the association between total PA, walking, and vigorous exercise
and the incidence of coronary events was examined using approximately 72,000 female
nurses who were 40-65 years of age in 1986. Results showed a strong graded inverse
relationship between PA and the risk of coronary events. Sedentary women who became
active in middle adulthood or later had a lower risk of coronary events than their

counterparts who remained sedentary. The authors concluded that brisk walking and

vigorous exercise were associated with substantial and similar reductions in the incidence
of coronary events among women (Manson et al., 1999).
Physical Activity Measurement

To truly understand the role of physical activity in the prevention of chronic
disease it is important to ensure that PA measurement is accurate. In a review article
about the spectrum of PA, cardiovascular disease and health, LaPorte et al. (1984) noted
a lack of framework from which to measure physical activity, especially in epidemiologic
studies. The authors stated that in order to minimize PA measurement errors,
assessments must be valid, speciﬁc, reproducible, quantitative, practical, and inexpensive
(LaPorte et al., 1984). In a Similar article regarding methodologic issues in measuring
PA and physical ﬁtness, Haskell and Kieman stated that the two are complex entities
comprised of numerous diverse components that present a challenge in terms of accurate,
reliable measurement (Haskell and Kieman, 2000). PA can be assessed by a variety of
questionnaires, diaries and logs, and by monitoring body movement or physiological
responses. The selection of the measurement method depends on the purpose of
evaluation, nature of the study population, and resources available (Haskell and Kieman,
2000). Surveys are often used because of practicality and low cost of assessing PA in
large populations. Recall surveys have been used for time frames ranging from one day
to lifetime PA. Either precise details about PA and/or general estimates can be obtained
for the time frame of interest (Haskell and Kieman, 2000).

In 1986, Washbum and Montoye published a review on the assessment of PA by
questionnaire (Washbum and Montoye, 1986). Typical PA questionnaires require recall

of speciﬁc physical activity behaviors and can be immediate, by subjects maintaining a

10

detailed activity diary, or delayed, where individuals are asked to recall PA over a
speciﬁc time period. Delayed recall has been deemed the most practical and least
burdensome method, therefore is most commonly used (Washbum and Montoye, 1986).
The review assessed the validity, reliability, practicality, and relationship to disease of
several commonly used PA recall tools during that time period. Major survey differences
exist in the complexity of the questionnaires, time to completion, detail of recall
expected, time frame assessed, and type of measurement scale from which to determine
the outcome (Washbum and Montoye, 1986). Lack of a criterion assessment tool makes
adequate validation of PA questionnaires difﬁcult. To help with the validity and
reliability of a PA recall questionnaire, the researcher must consider population size,
amount of detail required for data analysis, as well as age, sex, socioeconomic status, and
racial make up of the target population (Washbum and Montoye, 1986). Since
questionnaires may be used to gather a variety of information from many diverse groups,
understanding the target population and pre-determining the outcomes of interest are
important steps in the instrument creation process. The validity and reliability of each
questionnaire then must be tested and documented before wide-spread use is acceptable.
Physical Activity Recall (PAR) validation and reliability studies

PAR tools have been used to examine the relationship between activity and
chronic diseases for more than 50 years. In 1962, Wessel, Montoye, and Mitchell
developed a PAR record and format to score activity questionnaires. During the fall and
winter of 1962-63 100 men and women completed the Self-Administered Questionnaire
(SAQ). Activity categories included occupational, leisure time, and combined. At this

time, an appointment was made for a meeting with a trained interviewer. An Interviewer

ll

Administered Questionnaire (IAQ) was then completed. Once the two administration
techniques were completed, three judges independently reviewed the responses and
classiﬁed the participant as very active, active, moderate, or light and the responses from
the SAQ and IAQ were compared (Wessel, Montoye, and Mitchell, 1965). The results
indicated that there was greater agreement when classifying women into activity groups
than men, and that the SAQ form was more satisfactory in classifying women than men.
Even with those promising results, disagreements existed in classifying both genders in
all categories. The authors noted that vague questions and point scale were the main
reason for disagreements, and concluded further work was needed in developing the tool
and scoring format (Wessel, Montoye, and Mitchell, 1965).

In 1967 the SAQ was revised as used as a part of the Tecumseh Community
Health Study (Reiff et al., 1967). The preliminary attempt to assess PA by self-
administration encountered some of the same misinterpretation problems as the original
SAQ. Therefore, the authors concluded that a personal interview may be necessary to
obtain accurate estimates. Questions were designed to gather information regarding
occupation, hours worked, transportation to and from work and participation in major
home repairs and maintenance. Leisure time sports, gardening and other physical
activities were checked on a list. Intensities for different physical activities as well as
energy expenditure values were collected and calculated. The overall results Show the
revised interview administered questionnaire was easier to use and score (Reiff et al.,
1967)

Many more efforts were undertaken to create PA questionnaires after the earlier

attempts in the 1960’s and 70’s. However, questions about their reliability and validity

12

remained. In 1984, Taylor et al. compared a seven-day activity recall tool, a self-report
PA log and direct PA measurement (Vitalog, minicomputer) to identify bias that might
exist among the three instruments (Taylor et al., 1984). Thirty White males from 34-69
years of age were recruited to participate. Some were members or cardiac rehabilitation
programs while the majority participated in an unsupervised YMCA exercise program.
All subjects completed daily self-report PA logs and underwent the seven-day interview
administered recall. Fifteen subjects the Vitalog, for the ﬁrst three of the seven day
report and recall periods (Taylor et al., 1984).

The self-report log was designed to be compatible with the recall, and required
subjects to rate their activities on a scale from light, to very hard and note the type and
duration of activity. The recall was administered by a trained interviewer and could be
completed in approximately 15 minutes. Subjects were asked to recall the types of
activities in which they participated, as well as estimate total hours spent in moderate,
hard, and very hard activities based on examples provided. The Vitalog monitor uses
three chest electrodes to monitor heart rate via R-R wave detection, as well as a mercury
motion sensor to detect body movement (Taylor et al., 1984).

The recall data were compared with the YMCA attendance logs for 10 subjects.
Correlations between the self-report and recall PA data were strongest for very hard
weekend activity (r=0.90) and weakest for hard and very hard weekday activity (r=0.39),
but all values were signiﬁcant at P=S0.05. Twenty-six classes were noted by the
YMCA log, all but two were reported by recall. Overall agreement between the self-
report and Vitalog frequency of episodes of moderate, hard, or very hard activity was

94%. The authors concluded that a seven-day recall accurately reﬂects mean energy

13

expenditure. They noted each method had advantages and disadvantages, but ease of use
and lower cost of the recall may give it the greatest potential for applicability to
population studies (Taylor et al., 1984).

In the 1990’s, other recall tools were analyzed for reliability and validity.
Examples included the Tecumseh, the Five-City, and the Minnesota Leisure Time
Physical Activity questionnaires. Ainslie et al. (2003) examined the correlations between
these questionnaires and the gold standard of doubly labeled water. The overall results
showed that the Tecumseh (r=0.64) and Minnesota Leisure Time Physical Activity
(r=0.74) validation studies, but not the Five City Questionnaire (r=0.42), showed
signiﬁcant agreement with doubly labeled water. Noted advantages of the recall
instruments included short time needed to complete the questionnaire, simple scoring
systems, low cost, ability to study large samples, and unlimited duration of use. The
authors concluded that the preferred method to determine energy expenditure principally
depends on factors such as the number of subjects, time period of measurement, and
ﬁnances available (Ainslie et al., 2003).

In 2001, Norman et al. conducted a study analyzing the validity and
reproducibility of self-reported total PA indexed by BMI. One-hundred and eleven men,
aged 47-78, completed a questionnaire and activity records. The questionnaire was sent
twice a year, and included questions on level of PA at work, hours per day of
walking/biking, home/household work, leisure time activity/inactivity, and sleeping.
Two seven-day activity records, performed six months apart, were used as the reference
method. Results showed signiﬁcantly higher validity correlations for self-report in men

below, versus above a BMI of 26 (Norman et al., 2001).

14

Historical PARS

In 1991, Blair et a1. assessed the reliability of long-term PA participation recall
by middle-aged men and women enrolled in a worksite health study. LTPA was assessed
at baseline, and energy expenditure spent in total, light, moderate, and vigorous activities
was calculated. The long-term recall of baseline LTPA was determined one to ten years
after the examination (Blair et al., 1991). Correlations were modest (ranging from
r=0.20-0.50) and percent agreement was 60-75% between baseline and recalled activity.
Multiple regression analysis suggested that recalled activity was a signiﬁcant predictor of
baseline activity, but recall interval and age were not important. The authors concluded
that questionnaire assessment of long-term PAR appeared reliable, length of recall
interval up to ten years was not an important factor, and that individuals were better able
to recall vigorous, compared to less intense activities (Blair et al., 1991).

In 1997, Torgen et al. (1997) published the results from a study that evaluated the
reproducibility of a questionnaire for the assessment of past and present PA. In 1970, a
survey regarding requirements of medical and social services was conducted with 2500
men and women between the ages of 18-65 living in Stockholm County, Sweden. In
1993, the previous survey participants, below the age of 59 years, were asked to
participate in a follow-up piece focusing on musculoskeletal function and health. Torgen
et al. (1997) then investigated a sub-sample of the 1993 study population to analyze the
reproducibility of the PA questionnaire data twelve months later.

Results Showed the reproducibility of questions on PA at work was higher than
that of during leisure time. The authors stated that reliable retrospective information

about LTPA could perhaps not be collected by self-administered questionnaires and other

15

methods such as interview-based questionnaires may be more suitable (Torgen et al.,
1997)

Previous studies support the use of PA recall surveys for different time frames in
similar populations, but what has not been conﬁrmed is the reliability of PAR from the
distant past. In 1999, Falkner et al. examined the quality of historical recall of PA in a
cohort of western New York residents followed Since 1960 in the Buffalo Health Study
(Falkner et al., 1999). The Buffalo Health Study was developed in the 1950’s to explore
the relations between blood pressure and social, environmental, and genetic factors in a
general population of randomly selected residents in Buffalo, New York. Surviving
residents from the original study who were 47 years old or less represented the follow-up
sample used. Two-hour recall interviews took place between August 1992 and August
1996 (Falkner etal., 1999).

Results showed that the recalled reports underestimated past weekday activities
when overall activity was examined. Estimates closer to the original 1960 data were
found when activity levels were examined. Both men and women were better able to
recall weekday light and moderate activities. Most participants underestimated overall
weekend-day activity, but males overestimated weekend-day hard activity. Correlations
for weekend-day activity were highest for summer sports in females, and winter sports
for both sexes. Correlations ranged from r=0.29-0.45. The authors believed that based
on the length of time between the original assessment and follow-up recall interviews, the
correlations were close to those from other studies where recalls were ten years or less
(Falkner et al., 1999). Differences between PAR and gender may need more study based

on these results. Also, with the low correlations for lower intensity activities such as

walking, distant past recall might be effected by activity intensity.

In 2002, the reproducibility of self-administered lifetime PA questionnaire was
assessed using a recall tool designed to assess duration, frequency, and intensity of
lifetime household and recreational activity (Chasen-Taber et al., 2002). A modiﬁed
version of the Historical Leisure Time Activity Questionnaire was used to assess PA of
female college alumni over four time periods (menarche to 21, 22-34, 35-50, and 51-65).
For assessment of questionnaire reproducibility, the PA questionnaire was repeated at the
end of the subsequent year. Interclass correlation coefﬁcients for reproducibility were
r=0.82 for total PA, 0.80 for moderate lifetime activity, 0.86 for vigorous activity, 0.87
for lifetime recreational activity, and 0.78 for household activity. The authors stated that
the PA questionnaire was reproducible and provided a useful measure of average lifetime
activity, but also made a point that the omission of household activities, especially for
women, may underestimate total activity (Chasen-Taber et al., 2002).

In 2002, Evenson et al. conducted a study involving female adult lifetime activity
(Evenson et al., 2002). The Women’s Health Initiative began in 1992 to investigate the
effects of selected interventions and risk factors on morbidity and mortality of cancer,
cardiovascular disease, and osteoporotic fractures among postmenopausal women.
Evenson et al used a sub-sample of women (approximately 70,000) from the health
initiative cohort to assess differences in vigorous PA at various times over the life span,
assess whether reports of participation in vigorous PA were associated with self-reported
current PA participation, and determine factors related to participation in vigorous PA by
race/ethnicity among women (Evenson et al., 2002). A questionnaire was used to assess

the women’s vigorous activity behavior at ages 18, 35, 50 years. The results showed that

17

prevalence of vigorous activity declined with age, with the largest decrease occurring
after age 50 for all ethnicities. Current vigorous activity was higher among women with
lower body mass indices, non-smoking, in excellent general health, and of higher
socioeconomic status across ethnic groups. Based on the results, the authors suggest
lower prevalence of vigorous PA in the postmenopausal period may indicate a complex
of health-related attitudes and behaviors beyond race/ethnicity. The authors also stated
that the pre-menopausal period may be a critical time to target interventions to help
women achieve and maintain their PA into the postmenopausal period (Evenson et al.,
2002)

In 2004 Bowles et al. determined the construct validity of self-reported walking,
running, and jogging activity on the basis of data from the Aerobics Center Longitudinal
Study (Bowles et al., 2004). Approximately 5,000 men and women underwent one or
more medical examination between 1976 and 1985 and completed a follow-up
questionnaire in 1986. The follow-up questionnaire assessed walking, jogging, and
running along with other strenuous activities for each year from 1976 through 1985. The
results indicated signiﬁcant correlations between recalled walking, jogging, and running
and treadmill times for each year through out the ten-year period. Engaging sufﬁcient
historical PA for health beneﬁts was associated with higher treadmill times and lower
BMIS for men and women and lower triglyceride levels for men (Bowles et al., 2004).

The establishment of certain historical recall tools has expanded the time frames
from which physical activity data can be gathered. Many tools used have been created
for speciﬁc populations and time periods, as suggested by Washbum and Montoye, and

others. What is not known is whether there is a speciﬁc format or framework that can

18

allow a researcher to target speciﬁc populations and time periods of interest, but work
from an established base.
The Modifiable Activity Questionnaire

The Modiﬁable Activity Questionnaire (MAQ) is an historical physical activity
recall tool designed for easy modiﬁcation to maximize the investigator’s ability to assess
PA in a variety of populations (Kriska and Bennett, 1997). Historical PA is activity
engaged in more than one-year prior to the time of assessment. The questionnaire can
assess past-year and past-week occupational and leisure activities, as well as extreme
levels of inactivity due to disability. The original version of the questionnaire also
assessed historical activity, which was used for retrospective studies of diabetes risk.

In 1990, the MAQ was used to examine the relationship of PA and diabetes in
Pima Indians (Kriska et al., 1990). The questionnaire was used to assess both historical
(lifetime PA) and more current (past-week, past-year) leisure and occupational PA. The
questionnaire was interviewer-administered, and included components that were deemed
necessary for comprehensive PA assessment of diabetes research in Native Americans.
Reproducibility of the past-year LTPA estimate was determined in 69 participants aged
10-59 years. The data were found to be reliable in all age groups with exception to the
10-14 year olds (Kriska et al., 1990). Validity of the current-activity question section
was determined via comparison with activity monitors. The authors concluded that the
questionnaire was found to be both reliable and feasible to use in the Pima Indian

population (Kriska et al., 1990).

In 1995, the MAQ was used further to examine the associate between glucose

tolerance and physical activity, but this time in middle-aged non-diabetic, multiethnic

l9

residents of Mauritius (Pereira et al., 1995). Mauritius is an island located in the
southwest Indian Ocean, and it’s inhabitants have a high prevalence of non-insulin
dependent diabetes mellitus (NIDDM). The survey was modiﬁed for this study by
expanding the occupational and leisure time activities to include popular options Speciﬁc
to the Mauritius population. Results showed total PA to be an independent predictor of
2-hour post-load glucose concentration after controlling for body mass, waist-hip ratio,
age, and family history of NIDDM. The authors stated that the data were supportive of a
potentially important role of PA in prevention of NIDDM in middle-aged inhabitants of
Mauritius (Pereira et al. 1995).

Recently, the MAQ has been used as a self-administered questionnaire to assess
past-year PA and compared to the interview-administered version and doubly-labeled
water (Vuillemin et al., 1999). Eighty-four subjects enrolled in an ongoing prospective
study completed both the self and interviewer-administered versions of the MAQ.
Excellent agreement was found between the self- and interview-administered mode of the
questionnaires, Spearman rank order correlations were r=0.88 and 0.92 for leisure
activity. Even though agreement between the two modes of questionnaire administration
was strong, the authors noted that presence of an interviewer may provide the individual
with a more structured framework for his/her response and may help the subject give
more information (Vuillemin et al., 1999).

Summary

Overall, the literature has shown that PA can affect cardiovascular disease risk

positively, and that it is an important health related factor to measure. PA measurement

has improved over the years, and the recall of current PA by different PAR

20

questionnaires has been shown to be accurate. The ability to assess historical PA,
ranging from past week to 10 years or more, may allow researchers to assess the affects
of distant past and lifetime PA on disease risk and mortality. Some historical recall tools
have been validated using different populations and time periods of interest.

Based on the established reliability and validity of the MAQ and the adaptability
of the questionnaire format for different populations and designs, we chose to use it as a
PAR tool for women ﬁve to six years postpartum. The tool has been modiﬁed to ﬁt
speciﬁc time frames of interest, second and third trimester of a pregnancy approximately
six years ago, as well as twelve weeks postpartum. The instrument was modiﬁed to be
used as an interview tool, not self-administered. It is believed the tool can be easily
modiﬁed for the pregnant population and Show Similar validity results as found in

previous research using different populations.

21

References

Ainslie P, Reilly T, and Westerterp K (2003) Estimating human energy expenditure: A
review of techniques with particular reference to doubly labeled water. Sports Medicine
33: 683-98.

Blair S, Dowda M, Pate R, Kronenfeld J, Howe H, Parker G, Blair A, and Fridinger F
(1991) Reliability of long-term recall of participation in physical activity by middle-aged
men and women. Am J of Epidemiology 133: 266-75.

Bowles H, Fitzgerald S, Morrow J, Jackson A, and Blair S (2004) Construct validity of
self-reported historical physical activity. Am J of Epidemiology 160: 279-86.

Casperson C, Powell K, and Christenson G (1985) Physical activity, exercise, and
physical ﬁtness: deﬁnitions and distinctions for health-related research. Public Health
Reports 100: 126-31.

Chasen-Taber L, Erickson J, McBride J, Nasca P, Chasen-Taber S, and Freedson P
(2002) Reproducibility of self-administered lifetime physical activity questionnaire
among female college alumnae. Am J of Epidemiology 155: 282-9.

Evenson K, Wilcox S, Pettinger M, Brunner R, King A, and McTiernan A (2002)
Vigorous leisure activity through women’s adult life. Am J of Epidemiology I 5 6: 945-53.

Falkner K, Trevisan M, and McCann S (1999) Reliability of recall of physical activity in
the distant past. Am J of Epidemiologz 150: 195-205.

Haskell W, and Kieman M (2000) Methodologic issues in measuring physical activity
and physical ﬁtness when evaluating the role of dietary supplements for physically active
people. Am J of Clin Nutr 72: 8541-50.

Iverson D, Fielding J, Crow R, and Christenson G (1985) The promotion of physical
activity in the US. population: the status of programs in medical, worksite, community,
and school settings. Public Health Reports 100: 212-24.

Kannel W, Belanger A, D’Agostino R, and Israel I (1986) Physical activity and physical
demand on the job and risk of cardiovascular disease and death: The Framingham Study.
American Heart Journal 112: 820-5.

Kriska A, Knowler W, LaPorte R, Drash A, Wing R, Blair S, et a1 (1990) Development
of questionnaire to examine relationship of physical activity and diabetes in Pima

Indians. Diabetes Care 13: 401-11.

Kriska A (1997) Modiﬁable Activity Questionnaire. MSSE 29: 873-78.

22

LaPorte R, Adams L, Savage D, Brenes G, Dearwater S, and Cook T (1984) The
spectrum of physical activity, cardiovascular disease and health: an epidemiologic
perspective. Am J of Epidemiology 120: 507-17.

LaPorte R, Montoye W, and Casperson C (1985) Assessment of physical activity in
epidemiologic research: problems and prospects. Public Health Reports 110: 131-46.

Manson J, Hu F, Rich-Edwards J, Colditz G, Stampfer M, Willett W, Speizer F, and
Hennekens C (1999) A prospective study of walking as compared with vigorous exercise
in the prevention of coronary heart disease in women. The New England Journal of
Medicine 341: 650-8.

Montoye H, and Taylor H (1984) Measurement of physical activity in population studies:
a review. Human Biology 56: 195-216.

Norman A, Bellocco R, Bergstrom A, and Wolk A (2001) Validity and reproducibility of
self-reported total physical activity-differences by relative weight. Int J of Obesity 25:
682-88.

Paffenbarger R, Hyde R, Wing A, and Steinmetz C (1984) A natural history of
athleticism and cardiovascular health. JAMA 252: 491-5.

Paffenbarger R, Hyde P, Wing A, Hsieh C (1986) Physical activity, all-cause mortality,
and longevity of college alumni. New England Journal of Medicine 31: 605-13.

Pate R, Pratt M, Blair S, Haskell W, Macera C, Bouchard C, Buchner D, Ettinger W,
Heath G, King A, et al (1995) Physical activity and public health. A recommendation
from the Centers of Disease Control and Prevention and the American College of Sports
Medicine. JAMA 273: 402-7.

Pereira M, Kriska A, Joswiak M, Dowse G, Collins V, Zimmet P, Gareeboo H, Chitson
P, Hemraj F, Purran A, and F areed D (1995) Physical inactivity and glucose tolerance in
the multiethnic island of Mauritius. MSSE 27: 1626-34.

Powell K, Thompson P, Casperson C, Kendrick J (1987) Physical activity and the
incidence of coronary heart disease. Annual Review of Public Health 8: 253-87.

Reiff G, Montoye H, Remington R, Napier J, Metzner H, and Epstein F (1967)
Assessment of physical activity by questionnaire and interview. Journal of Sports
Medicine and Physical Fitness 7: 135-42.

Siscovick D, LaPorte R, and Newman J (1985) The disease-speciﬁc beneﬁts and risks of
physical activity and exercise. Public Health Reports 100: 180-8.

23

Taylor C, Coffey T, Berra K, Iaffaldano R, Casey K, and Haskell L (1984) Seven-day
activity and self-report compared to a direct measure of physical activity. Am J of
Epidemiology 120: 818-24.

Torgen M, Alfredsson L, Koster M, Wiktorin C, Smith K, and Kilbom A (1997)
Reproducibility of a questionnaire for assessment of present and past physical activities.
Int Arch Occup Environ Health 70: 107-18.

US. Department of Health and Human Services. Healthy People 2010. Retrieved June
10, 2004 from, http://wwwhealthypeoplegov.

Vuillemin A, Oppert J, Guillemin F, Essermeant L, Fontvieille A, Galan P, et al (2000)
Self-administered questionnaire compared with interview to assess past-year physical
activity. MSSE 32: 1119-24.

Washbum R, and Montoye H (1986) The assessment of physical activity by
questionnaire. Am J of Epidemiology 123: 563-76.

Wessel J, Montoye H, and Mitchell H (1967) Physical activity assessment: by recall
record. American Journal of Public Health 55: 143 0-6.

24

CHAPTER 3
Validation of a historical physical activity recall tool

Introduction
Background

Physical activity (PA) is important in the prevention of chronic disease and an
integral part of healthy lifestyles. Epidemiological studies have shown that PA plays a
role in chronic disease prevention (Paffenbarger et al., 1986; Powell and Paffenbarger,
1987). Healthy people 2010, a comprehensive set of disease and health promotion
objectives, considers PA as an integral part of chronic disease prevention and has placed
it as one of the major health issues for the nation (http://www.healthypeople.gov).

PA measurement is a process that should be performed with as much precision
and accuracy as possible. Many tools are available to estimate PA on a daily, weekly, or
yearly basis (Kriska et al., 1988). Choice of instrument is a ﬁrnction of research
questions, study sample, funding, and other factors (Haskell and Kieman, 2000).
Instruments such as accelerometers and heart rate monitors can be worn and data are
collected as the PA is performed. From these measures, caloric expenditure can be
estimated. While fairly precise, these techniques are not generally used on large samples
where recall surveys are more practical and cost effective (Haskell and Kieman, 2000).
Physical activity recall (PAR) relies on subject self-report from which caloric expenditure
can be estimated. Recall surveys have been used for a variety of time frames and,
depending on the questions asked, can gather either precise details about physical activity
or general estimates of usual or typical PA participation (Haskell and Kieman, 2000).

Most PAR tools have been validated using other direct measurement techniques.

Stein et al. (2003) utilized accelerometry, HR monitoring and PAR and evaluated their

25

reliability and agreement during pregnancy and the postpartum period. The authors
found that the instruments were very reliable and showed good agreement on the group
level.

Many PAR tools have been validated for use with different populations for
assessing PA. The Modiﬁable Assessment Questionnaire (MAQ) is a tool that was
developed to address physical activity and diabetes risk in Pima Indians, but it has since
been validated with other populations (Kriska et al., 1990; Vuillemin et al., 2000). The
strength of the MAQ is that it can be modiﬁed for population to be studies and time
period. Physical activity can be recalled from the last week, to last year, to a speciﬁc
time period in a person’s life (Kriska, 1997). Modiﬁability will make this tool most
useful in evaluating historical PA during a speciﬁc past event, such as pregnancy.
Signiﬁcance of the problem

Historical PAR research has grown in recent years. Studies have focused on
adults’ abilities to recall PA and sport experiences from their youth, as well as
occupational activity and relate those to present health status and disease risk (Chasen-
Taber et al., 2002; Torgen et al., 1997). Validity of such a tool is difﬁcult and limited by
access to accurate historical data. A longitudinal prospective study with previously
recorded activity levels is needed to validate an historical recall tool. Also, the validity
of the MAQ for a past pregnancy is unknown. If it is determined that this tool is valid
with the postpartum population, information regarding past pregnancy PA can be
collected retrospectively and assist in determining the relationship of previous physical
activity and chronic disease risk of both mother and offspring.

Purpose of the study

26

In this study, the historical data measured during a pregnancy that occurred six
years ago will be compared to the subjects’ physical activity recall, allowing for the
validation of the MAQ for historical PA assessment. The purpose of this study is to
evaluate the ability of women to recall their PA during three time points previously
assessed (20 weeks gestation, 32 weeks gestation, 12 weeks postpartum) during a
pregnancy that occurred six years ago. The research questions are as follows.

1. How accurately can women recall physical activity performed during a

pregnancy that occurred six years ago?

2. Is a woman’s ability to recall physical activity performed during a

pregnancy that occurred six years ago related to current physical activity?
Methods
Participants

56 women participated in a study designed to determine the reliability and validity
of three commonly used physical activity measurement modalities from 1997-2000
(Pivarnik et al., 2002; Stein et al., 2003). Study participants were recruited from
obstetrical care clinics in the Mid-Michigan area. All participants delivered healthy,
singleton infants. The original MAMA study involved the collection of heart rate
telemetry, accelerometry, and a PAR over two consecutive days at 20 and 32 weeks
gestation, and 12 weeks postpartum. The researcher reviewed the two-day PAR data
with participants during each visit to the lab. When the women were recruited they were
informed that the investigators may wish to obtain some follow-up measures (on them
and/or their offspring) at a later time. Thus, participants were then contacted to

participate in the current study. Email, telephone calls, and letters were used to contact

27

the 41 of the original 56 participants believed to still live in the area. After many phone
calls, thirty women were located and agreed to participate, with only one refusal.
Instrumentation

The Modiﬁable Activity Questionnaire (MAQ) is a PAR instrument designed for
easy modiﬁcation to maximize the investigator’s ability to assess PA levels in different
populations (Kriska et al., 1990; Kriska, 1997). This questionnaire was used to assess
historical PA levels at time points from a pregnancy that occurred six years ago. Those
time points included 20 weeks and 32 weeks gestation, and 12 weeks postpartum.
Sample size, power analysis

Due to the follow-up design of this study, we could not increase sample size to
enhance statistical power. Therefore, correlation values needed to achieve meaningful
relationships and signiﬁcance in line with previous reports were calculated using the
available sample of 30 women. A review of other PAR tools and MAQ studies assisted
in the determination of correlation Signiﬁcance.

Agreement between measured PA and self-reported PA via PAR tools for the
same time periods have been established in studies comparing direct measurement
techniques such as Vitalog and Caltrac accelerometers, to exercise logs and recall (Stein
et al., 2003; Taylor et al., 1984). Correlations have ranged ﬁ‘om 0.28 to 0.90 in studies
that have assessed the validity and reliability of PAR tools (Kriska, 1997; Norman et al.,
2001; Vuillemin et al., 2000; Wareham et al., 2002). In studies that have measured
lifetime or historical leisure time physical activity levels, correlations have ranged from

r=0.33 to 0.70 (Chasen-Taber et al. 2002; Torgen et al., 1997). Based on our sample size

28

of 30, conversion tables indicated that an r=0.36 would be signiﬁcant for an alpha level
of P<0.05, while an r=0.46 would signiﬁcant for an alpha level of P<0.01.
Data Callection Procedures

In the original MAMA study, actual PA data for all the three previously
mentioned time points were collected using a two-day exercise PAR. During those two
days, women wore watches with alarms that sounded on the hour. This alarm served as a
reminder that they should record any activity performed within the previous hour. In
addition, upon arrival to the lab, the study coordinator debriefed each participant to be
sure that she recalled her previous day’s activities as accurately as possible (Stein et al.,
2003). During the follow-up visit, participants came to the Human Energy Research Lab
at Michigan State University for one 45-60 minute visit. Upon arrival the participant was
reminded of the study aims and protocols, after which written informed consent was
Obtained. Next, physical activity measures for 20 and 32 weeks gestation, 12 weeks
postpartum, and current (last week) were collected. The historical recall portion of this
study involved the repeated use of The Modiﬁable Activity Questionnaire (MAQ). Each
participant was administered the MAQ for each time period in the order of most distant
past (20 weeks gestation) to most current (last week).

Participants were asked to recall their PA levels during a usual week for 20 weeks
gestation (the most historical time period of interest), then 32 weeks gestation, 12 weeks
postpartum of a pregnancy that occurred six years ago, and ﬁnally current (last week). In
order to assess past and current PA using the MAQ, participants were asked to circle all
activities (from a pre-determined list generated from pilot studies) they had participated

in during the time period of interest. Appendix B shows examples of the MAQ activity

29

list. The circled activities were then listed in the “activity” box of the questionnaire. The
frequencies, in days per week, and duration, in hours per bout, were established for each
chosen activity. Hours per week for all activities chosen were calculated by multiplying
frequency by duration. The MET value for all listed activities was determined using the
compendium for physical activities developed by Ainsworth et al. (2000). Estimated
energy expenditure for each activity was calculated and multiplied by the hours per week
in order to calculate a weekly value, for example Kcal/week,
(MET*WT)(dur/60)*freq=Kcal/wk (Kriska, 1997). The estimated costs of each activity
were derived from the most recent compendium of physical activities (Ainsworth et al.,
2000). After the average weekly energy expenditure values were calculated, those data
were divided by seven to derive daily energy expenditure, then divided by the maternal
body weights for each study period in order to obtain energy expenditure relative to
weight ([Kcal/day]/kg=Kcal/day/kg). Please refer to appendix A for the energy
expenditure equations used with the MAQ described in this section.
Data analysis

In the original MAMA study, actual PA data for 20 and 32 weeks gestation and
12 weeks postpartum were collected using a two-day exercise PAR. Leisure time energy
expenditure values (kcal/day/kg) calculated from the PAR used in the original study and
those from the MAQ used in the follow up study were analyzed based on the research
questions for this dissertation. Statistical signiﬁcance for all analytical procedures was
set at an alpha level of P=S0.05. The statistical program of JMP version 4 was used for
all statistical procedures (Sall, Lehman, and Creighton, 2001). During the data analysis

process it was discovered that there were two participants whose PA at all time periods

30

was 2-3 standard deviations above the sample norm. Further inspection showed the
women, though more active than the rest of the sample, to be with in the regression line
and therefore, remained part of the analyses. Some analyses also involve comparisons
based on current PA levels. Therefore the participants were separated into two groups
based on those who met the current CDC/ACSM PA recommendations of 20 MET/wk or
more, and those who did not.

Research Question One: How accurately can a woman recall physical activity performed
during a pregnancy from six years ago?

Data analysis included calculation Of Pearson correlations for the measured
leisure time energy expenditure values (kcal/day/kg) of each time period (20 and 32
weeks gestation and 12 weeks postpartum) and the recalled values for those same time
periods.

Research Question Two: Is a woman ’s ability to recall the physical activity performed
during a pregnancy ﬁ'om six years ago related to current physical activity?

Data analysis included a one-way ANOVA to compare difference between
measured and recalled past energy expenditure values at the different time periods (20
and 32 weeks gestation and 12 weeks postpartum), between subjects meeting
CDC/ACSM PA guidelines (220 MET/wk), and those who did not.

Results

Descriptive characteristics and other select physiological measures of the entire
sample can be found in Table 1. The averages for the sample included age 36.2 (:ES.0)
years, weight 65.5 kg ($12.9), height 165.9 cm (:l:6.9), and parity 2.5 children (:I:0.94).

Sixty percent of the sample met the CDC/ACSM guidelines for current PA (220

31

MET/wk). Refer to Table 2 for the average PA in MET/wk at all time periods of interest
(20 and 32 weeks gestation, 12 weeks postpartum and current). The results for Part I of
this dissertation are organized by research question.

Research question one stated: How accurately can a woman recall physical
activity performed during a pregnancy from six years ago? Results of the Pearson
correlations on the entire subject population for the MAQ recalled and PAR measured
leisure time energy expenditures (kcal/day/kg) were at 20 weeks gestation (r=0.57;
P<0.01), at 32 weeks gestation (r=0.85; P<0.01), and at 12 weeks postpartum (r=0.86; P<
0.01). Refer to Figures 1-3 for results.

Research question two stated: Is a woman ’s ability to recall the physical activity
performed during a pregnancy from six years ago related to current physical activity?
Results showed that differences between recalled (MAQ) and measured (PAR) leisure
time energy expenditure (kcal/day/kg) did not differ between women who exceeded or
did not exceed CDC/ACSM recommended twenty METS per week. Differences in recall
ability, or recalled (MAQ) minus measured (PAR) energy expenditure were small. Refer
to Table 3 for results.

Discussion

The purpose of this study was to evaluate the ability of women to recall their PA
during three time points previously assessed (20 weeks gestation, 32 weeks gestation, 12
weeks postpartum) during a pregnancy that occurred six years ago. Based on this
purpose, the research questions included: how accurately can women recall physical

activity performed during a pregnancy that occurred six years ago; and, was a woman’s

32

ability to recall physical activity performed during a pregnancy that occurred six years
ago related to current physical activity?

The results regarding the ability of a woman to recall her past pregnancy PA
showed the MAQ energy expenditure values to be signiﬁcantly related to the original
PAR energy expenditure values at all time periods of interest (20 and 32 weeks gestation
and 12 weeks postpartum; Figures 1-3). The correlations from this MAMA follow-up are
similar (r=0.57-0.86), if not stronger, to those found in previous PAR tool validation
studies, as well as MAQ validation studies.

When Blair et al. (1991) evaluated the reliability of long-term recall of
participation in PA by middle-aged men and women, correlations ranged from r=0.20-
0.50. The Blair et al. sample included 451 participants, but we have no knowledge
whether any of the women were pregnant. The length of recall for the Blair et al. study
was greater (ten years) than that for this study (Six years), which may have lessened their
reliability estimates.

In a study similar to that of Blair et al., Bowles et al. (2004) assessed construct
validity of self-reported historical physical activity. The results showed signiﬁcant
correlations between recalled walking, running, jogging, and treadmill times each year
throughout a 10 year period (Bowles et al., 2004). The correlations ranged from r=0.40-
0.61, very similar to those found in this study. The sample size was again much larger
(n=5,063) and included both genders. The time for follow-up varied in the Bowles et a1
study, ranging from 1 to 10 years. The strong correlations between original and recalled

energy expenditure found in the MAMA follow-up compare well with those from other

33

long-term recall studies and would suggest the MAQ to be a valid tool to assess past
pregnancy PA.

Chasen-Taber et al. (2002) examined the reproducibility of a self-administered
historical PAR with adult females. The age of the study sample ranged from 39-65 years,
and though it was not mentioned, some women may have been pregnant during at least
one of the time periods assessed (menarche to 21, 22-34, 35-50, and 51-65). Correlation
coefﬁcients for reproducibility ranged from r=0.70-0.87. These are similar to the
MAMA follow-up correlations found between actual and recalled PA for 32 weeks
gestation and 12 weeks postpartum six years ago. The correlations for the Chasen-Taber
et al. study are very high for the recall of PA from over 10-20 years ago.

Results from the present study can also be compared to those found in other
investigations that have used the MAQ to assess past PA. Kriska et al. (1990) examined
the relationship of PA and diabetes in Pima Indians, comparing the MAQ to values
gathered via the Caltrac activity monitor (Kriska et al., 1990). The correlation between
the two tools was relatively strong (r=0.62, PS 0.05) with a sample size of 17. Despite
the difference in culture between the two populations, sample size was similar to this
study as are the correlations at all time periods. In 1995, Pereira et al. (1995) used the
MAQ to assess the relationship between physical inactivity and glucose intolerance on
Mauritius adults. The MAQ was compared to a seven-point categorical activity scale
(CAS) to establish internal validity. Though correlations were not reported, the MAQ
and CAS PA measures were in agreement (PS 0.001) for both males and females, and PA
differed among the three activity categories (inactive, moderate, and high). Sample size,

culture, and gender of the Mauritius study (n=2790) differed from out investigation.

34

Even with these differences, it appears that the MAQ correlates well with different PAR
tools and populations.

While our study showed women are able to recall their PA during a pregnancy
some six years ago, present physical activity level does not appear to inﬂuence this
ability. Because we are the ﬁrst to perform this analysis during pregnancy, direct
comparisons with similar previous studies are not possible. Norman et al. (2000)
assessed the validity and reliability of self—reported total PA in 286 males, and examined
differences by relative weight. The results showed the PAR tool correlated well with
seven-day activity records (r=0.56). In addition, the authors found that the participants
with a BMI S 26.0 kg-mz, were better able to recall their physical activity levels
compared to heavier subjects. Also, the lighter subjects performed the greatest about of
physical activity. The Normal et al. study did not assess historical PA, but the resulting
differences in self-reported PA by relative weight may indicate the same trend.
Limitations

Limitations of the present study include small sample Size and a homogeneous
sample in terms Of physical activity. Therefore, the range of activities and recall would
be restricted, compared to a more diverse group, lowering statistical associations between
our participants’ current PA levels and ability to recall past pregnancy PA. Even with the
limitations noted, MAQ recall correlations compared favorably to those reported in
previous historical recall validation studies. Therefore, it appears the MAQ, used in this
MAMA follow-up is an appropriate tool to assess past PA in postpartum women, six
years after delivery, regardless of their current activity participation levels.

Future Directions

35

Based on the results of this validation study, we recommend that the MAQ be
used in future long-term past pregnancy PA recall studies. Even with studies that did not
originally measure PA during a past pregnancy, investigators can retrospectively assess
PA and have reasonable conﬁdence in the accuracy of their data. New ﬁndings may add
to our knowledge of the relationship between PA during pregnancy and long-term body
weight changes or other chronic disease risk issues. The MAQ should also be used with
larger, more heterogeneous samples to determine it’s validity with different ethnic and
cultural groups. The MAQ should also be used in studies that include longer follow-up

(ten years or greater) to determine it’s accuracy at recalling PA in the more distant past.

36

References

Blair S, Dowda M, Pate R, Kronenfeld J, Howe H, Parker G, Blair A, and Fridinger F
(1991) Reliability of long-term recall of participation in physical activity by middle-aged
men and women. Am J Epidemiology 133: 266-75.

Bowles H Fitzgerald S, Morrow J, Jackson A, and Blair S (2004) Construct validity of
self—reported historical physical activity. Am J of Epidemiology 160: 279-86.

Chasen-Taber L, Erickson J, McBride J, Nasca P, Chasen-Taber S, and Freedson P
(2002) Reproducibility of self-administered lifetime physical activity questionnaire
among female college alumnae. Am J of Epidemiology 155: 282-9.

Haskell W, and Kieman M (2000) Methodologic issues in measuring physical activity
and physical ﬁtness when evaluating the role of dietary supplements for physically active
people. Am J of Clinical Nutrition 72: S541-50.

Kriska A, Sandler R, Cauley J, LaPorte R, Hom D, and Pambianco G (1988) The
assessment of historical physical activity and its relation to adult bone parameters. Am J
of Epidemiology 127: 1053-61.

Kriska A, Knowler W, LaPorte R, Drash A, Wing R, Blair S, et al (1990) Development
of questionnaire to examine relationship of physical activity and diabetes in Pima
Indians. Diabetes Care 13: 401-11.

Kriska A (1997) Modiﬁable Activity Questionnaire. MSSE 29: S73-78.

Norman A, Bellocco R, Bergstrom A, and Wolk A (2001) Validity and reproducibility of
self-reported total physical activity-differences by relative weight. Int J of Obesity 25 :
682-88.

Paffenbarger R, Hyde P, Wing A, Hsieh C (1986) Physical activity, all-cause mortality,
and longevity of college alumni. New England Journal of Medicine 31: 605-13.

Pereira M, Kriska A, Joswiak M, Dowse G, Collins V, Zimmet P, Gareeboo H, Chitson
P, Hemraj F, Purran A, and Fareed D (1995) Physical inactivity and glucose tolerance in
the multiethnic island of Mauritius. MSSE 27: 1626-34.

Powell K, Thompson P, Casperson C, Kendrick J (1987) Physical activity and the
incidence of coronary heart disease. Annual Review of Public Health 8: 253-87.

Sall J, Lehman A, and Creighton L (2001) JMP Start Statistics: a guide to statistics and
data analysis using JMP and JMP IN software, 2nd edition, Canada: Duxury.

Stein A, Rivera J, Pivarnik J (2003) Measuring energy expenditure in habitually active
and sedentary pregnant women. MSSE 35 : 1441-6.

37

Torgen M, Alfredsson L, Koster M, Wiktorin C, Smith K, and Kilbom A (1997)
Reproducibility of a questionnaire for assessment of present and past physical activities.
Int Arch Occup Environ Health 70: 107-18.

US. Department of Health and Human Services. Healthy People 2010. Retrieved June
10, 2004 from, http://www.healthypeople.gov.

Vuillemin A, Oppert J, Guillemin F, Essermeant L, Fontvieille A, Galan P, et a1 (2000)

Self-administered questionnaire compared with interview to assess past-year physical
activity. MSSE 32: 1119-24.

38

CHAPTER 4

The effects of past pregnancy physical activity on current physical activity, barriers to
physical activity, and body size.
Historical Background

Information regarding what activities a pregnant women should and should not
participate in has been available since the late 1700’s (Artal, Drinkwater, and Wisewell,
1991). In the 1920’s and 30’s some restrictions for maternal exercise were published, but
were based on social and cultural biases of the time, rather than empirical evidence
(Artal, Drinkwater, and Wisewell, 1991). In the 1920’s and into the 40’s prenatal
exercise programs were developed. Books directed toward self-powered, natural
birthings are considered the basis for current prenatal exercise (Fairbairn, 1924;
Fairbairn, 1926; Randell, 1945). In 1949 the United States Children’s Bureau published
an article which included physical activity recommendations (Children’s Bureau
Publication, 1949). The recommendations stated that moderate physical activity was
good for a pregnant woman who was without illness or complication, and that she could
continue housework, gardening, daily walks, and swim occasionally (Children’s Bureau
Publication, 1949).

During the 1970’s women of child-bearing age became more active (largely due
to increased sports opportunities for high school girls and collegiate women) and
concerns about vigorous physical activity during pregnancy arose. It was not well-
understood how a woman’s physiological responses to physical activity would affect the

maternal-fetal unit, so evidence based recommendations were not available. Physicians

and other health care professionals had strong concerns about the effects of physical

39

activity on maternal weight gain, blood pressure, glucose tolerance and insulin function,
and temperature regulation.
Early Animal Studies

The ﬁrst studies designed to systematically examine the maternal and fetal
responses to physical activity during pregnancy primarily used animals as subjects
(Clapp, 1980; Bagnall et al., 1983; Lotering, Gilbert, and Longo, 1983). In 1980, Clapp
found that strenuous exercise with near-term ewes resulted in a rapid increase in maternal
temperature, and decreased rate of blood ﬂow to the placenta and fetus by more that 50%
(Clapp, 1980). Despite this severe exercise challenge, the fetal lambs showed no signs of
stress after an hour of continuous maternal exercise, indicating the placenta was able to
compensate for the demanding cardiovascular responses.

A few years later, Bagnall, Mottola, and McFadden examined the effects of
strenuous exercise on maternal rats and their developing fetuses. Twenty female rats
were mated and divided into exercise and control groups after 19 days of treadmill
running acclimatization (Bagnall et al., 1983). The exercising rats continued to run
throughout gestation, while the controls did not. The rats were weighed daily, including
the day following birth. Offspring numbers and weights were recorded. Results showed
the fetuses were spared from any major effects of maternal exercise, with average birth
weight and litter size being unaffected. There was a signiﬁcant maternal weight
difference between the exercise and control group mother rats at the end of gestation.
The control group was 36 grams heavier at the end of gestation and lost weight after

delivery, but remained to 14.6 grams heavier than the exercise group. The authors

40

suggested the data reﬂect a decreased amount of amniotic ﬂuid and maternal body fat in
the exercising group (Bagnall et al., 1983).

To study whether maternal exercise causes fetal hypoxia, Lotgering, Gilbert, and
Longo measured maternal and fetal temperatures and blood gases, then calculated uterine
oxygen consumption of pregnant sheep in response to different treadmill exercise
regimens (Lotgering, Gilbert, and Longol983). The results showed that total uterine
oxygen consumption was maintained during exercise, despite the reduction in uterine
blood ﬂow. The authors believed hemoconcentration and increase oxygen extraction
helped maintain uterine oxygen consumption and suggested maternal exercise did not
represent a major stressful or hypoxic event to the fetus (Lotgering, Gilbert, Longo,
1983).

Reviews published in 1985 and 1987 summarized the information generated by
the animal studies of the 1980’s (Clapp, 1987; Lotgering, Gilbert, and Longo, 1985). The
authors reached three main conclusions 1) acute exercise during pregnancy had
signiﬁcant effects of maternal physiology without detrimental effect on the fetus in late
pregnancy, 2) little was understood about the maternal-fetal effects of exercise during
pregnancy with humans, 3) even with the reassuring information about exercise during
late pregnancy in large animals, human studies would be needed to answer the many
questions about the safety of exercise during pregnancy (Clapp, 1987; Lotgering, Gilbert,
and Longo, 1985). Major ﬁndings from animal studies conducted to this point in time
showed that despite some strong physiological responses of pregnant animals to exercise,
the maternal-fetal unit remains unharmed. Whether these results could be generalized to

humans had not yet been conﬁrmed. However, the animal studies provided important

41

insight into maternal-fetal adaptations to maternal exercise that proved valuable for future
research on humans.
Early Human Studies

The ﬁrst few human studies of exercise during pregnancy were designed to
evaluate both maternal and fetal effects. In 1974, Knuttgen and Emerson assessed
physiological responses to pregnancy at rest and during exercise. Thirteen women with
normal pregnancies participated in this study. and were tested at four-week intervals
throughout gestation as well as six-weeks postpartum. Both steady state treadmill
walking (4.5km/hr at 4% grade) and cycling (60 W) were performed. Oxygen
consumption, heart rate, and ventilation were evaluated. Preme and postpartum
values were also compared based on exercise mode, treadmill (weight bearing) or cycle
ergometer (non-weight bearing) (Knuttgen and Emerson, 1974). The results showed
pulmonary hyperventilation developed early in pregnancy and persisted at rest as well as
during both exercise modes. Oxygen consumption and body weight showed consistent
increases during pregnancy. Resting heart rate was more variable during pregnancy, and
a consistent increase was observed in six of the thirteen subjects. Results also showed an
increase in oxygen cost of treadmill walking, but not in cycle ergometry during
pregnancy. It is likely that the decreased exercise economy that occurred was a result of
pregnancy related weight gain (Knuttgen and Emerson, 1974).

Also in the 1970’s, Pomerance, Gluck, and Lynch sought to determine if the fetus
beneﬁted from maternal physical ﬁtness (Pomerance et al., 1974). The researchers
investigated whether length of gestation, length of labor, birth weight, length, head

circumference, and one-minute APGAR scores would differ based on maternal physical

42

ﬁtness levels. Study participants included forty-one women who delivered at a naval
hospital in California (Pomerance et al., 1974). Cycle ergometry was used to assess
maternal physical ﬁtness. Aerobic capacity was estimated using a submaximal protocol
(Astrand and Rodahl, 1970). The women were followed throughout the pregnancy and
delivery. The results showed an inverse relationship between physical ﬁtness scores and
pre-pregnancy weight, with lighter women being more physically ﬁt. Although
physically ﬁt multiparas tended to have shorter labors than primiparas, ﬁtness was not
signiﬁcantly related to infant birth weight, length, head circumference, or one-minute
APGAR scores (Pomerance etal., 1974).

In 1981, Artal et a1. assessed maternal cardiovascular and metabolic responses to
normal pregnancy. Study volunteers included twenty-three healthy pregnant women in
their third trimesters. Cardiovascular and metabolic measurements were recorded during
and after lS-minutes of light (2.3 METS) treadmill exercise (Artal et al., 1981).
Although exercise induced a signiﬁcant increase in maternal heart rate and a shortening
of the R to R intervals on EKG, both returned to baseline by 30-minutes of recovery.
Exercise also induced signiﬁcant transitory increases in glucagon, norepinephrine, and
epinephrine concentrations, all of which were reversed during recovery. The light
exercise did not result in a change in glucose or cortisol concentrations. The authors
concluded that light exercise for a brief duration elicited appropriate and transitory
cardiovascular and metabolic responses during normal pregnancy (Artal et al., 1981).

Even as the amount of research regarding the physiological and maternal-fetal
effects of exercise during pregnancy increased, concerns still existed. A major concern

was the risk of physical activity causing a decrease in birth weight or Shortened gestation,

43

predisposing the infant to preterm delivery and subsequent health problems. In 1983,
Berkowitz et al. examined the effects of physical activity on spontaneous preterm
delivery. The authors found that women who exercised were half as likely as non-
exercisers to deliver prematurely and found no evidence that physical activity was related
to an increase the incidence of either smaller than average babies or premature labor
(Berkowitz et al., 1983). However, it still remained unclear whether physical activity
during pregnancy would limit calories available to the fetus, resulting in grth
retardation, and/or cause uterine contractions leading to preterm deliveries. There was a
great need for more human physical activity and pregnancy studies involving prospective
cohort designs.

As a result of these few early human studies, the American College of
Obstetricians and Gynecologists (ACOG) published the ﬁrst guidelines regarding
exercise during pregnancy in 1985 (ACOG, 1985). Based on the limited research
available at the time, the recommendations were cautious. Concern for maternal-fetal
temperature regulation and the effect of exercise increasing the risk of preterm birth and
low birth weight (LBW) guided the recommendations. The authors indicated that
maternal heart rates should not exceed 140 beats per minute and vigorous exercise should
be limited to no more than 15 minutes per bout (ACOG, 1985).

After the publication of these ACOG guidelines, research regarding pregnancy
and physical activity increased signiﬁcantly with more than 500 papers being published
between 1985 and 1994. Many of the outcomes reﬂected the results of previous studies
showing no evidence that recreational exercise increases the risk of LBW or premature

labor, and may decrease the incidence of both (Klebanoff, Shiono, and Carey 1990,

44

Rabkin et al 1990). Studies Showed that maternal physical activity actually increased
placental grth and functional capacity, possibly providing a mechanism to positively
affect birth weight (Clapp and Rizk, 1992; Jackson et al., 1995).

More recent human studies

Lokey et al. published a meta-analysis of the effects of physical exercise on
pregnancy outcomes in 1991. Eighteen studies were included in this analysis. All studies
were written in English, included women who exercised during pregnancy, and
quantiﬁed outcome variables. Results showed that women who exercised during their
pregnancies did not differ from sedentary women for any of the measured outcome
variables; including maternal weight gain, infant birth weight, length of gestation, length
of labor, and APGAR scores. Many exercise programs exceeded the suggested 1985
ACOG limitations, without apparent adverse effects. The authors concluded that overall,
aerobic exercise performed for an average of forty-three minutes per day, three days per
week, at a heart rate up to 144 beats per minute, did not appear to be related to adverse
maternal or fetal effects in a healthy normal pregnancy (Lokey et al., 1991).

Based on encouraging results of studies performed after the 1985 guidelines,
ACOG published a revised, updated version in 1994 (ACOG 1994). New evidence had
failed to support concerns about harmful maternal-fetal effects ﬁ‘om exercise performed
during a normal healthy pregnancy. To the contrary, reviews analyzing the more recent
studies indicated that maintaining a physically active lifestyle during pregnancy could
enhance maternal health and improve the delivery process (Clapp, 1994; Crowell, 1995).
The updated 1994 recommendations no longer included heart rate or vigorous exercise

duration limitations for healthy mothers. Continuing aerobic exercise throughout

45

pregnancy was encouraged for mothers who were active before pregnancy. In addition,
individual health and exercise screenings and prescriptions were suggested for all
pregnant women (ACOG, 1994).

Many studies performed after the 1994 ACOG guidelines were prospective and
the majority supported the less stringent revised guidelines (Sternfeld et al., 1995; Kardel
and Kase, 1998; Hatch et al., 1998). In 1995, Stemfeld et a1 conducted an investigation
of the effects of participation in aerobic exercise on pregnancy outcome (Sternfeld et al.,
1995). Study participants included 388 women followed from approximately sixteen
weeks gestation through delivery. Frequency, duration, and type of exercise prior to and
during the ﬁrst trimester of pregnancy were collected by an in-person interview.
Telephone interviews were used during the second and third trimesters (Sternfeld et al.,
1995). Based on information collected, the women were grouped into different activity
levels for each time period of interest. The levels included; level I= aerobic exercise
(excluding walking) for 20-minutes at least 3 times per week, level II= aerobic exercise
for 20 minutes at least 3 times per week, if and only if vigorous walking was included,
level III= aerobic exercise for 20-minutes less than 3 times per week, level IV= aerobic
exercise less that once per week. Results showed mean birth weight was unrelated to
physical activity performed, either prior to, or during pregnancy. Gestational age and
weight gain were also unrelated to maternal physical activity level. Pregnancy related
symptoms, however, were inversely related to physical activity with women who
exercised in greater amounts earlier in pregnancy reported fewer discomforts in later
gestation. The authors concluded that aerobic exercise performed during pregnancy at

levels great enough to produce or maintain exercise-training effects did not adversely

46

affect birth weight or other maternal or fetal outcomes, but may be associated with fewer
perceived pregnancy-related discomforts (Sternfeld et al., 1995).

Stemfeld et al. (1995) were able to examine different activity levels at speciﬁc
time periods throughout gestation and assess relationships between those activity levels
and pregnancy outcomes. A similar prospective study was conducted in Oslo, Norway in
1998.. Women were given the choice to follow either a high or medium intensity exercise
program through pregnancy up to six-weeks postpartum (Kardel and Kase, 1998).
Results Showed no differences between program intensity and duration of labor, birth
weight, or APGAR scores at delivery. The authors concluded that healthy and well-
conditioned women may exercise during pregnancy without compromising fetal grth
or development (Kardel and Kase, 1998). Hatch et al. (1998) examining maternal leisure
time activity and timely delivery, also found no adverse effect of maternal exercise
during pregnancy on gestational length. In fact, women performing relatively high levels
of activity (greater than 1000 kcal per week) appeared to be at reduced risk of
Spontaneous preterm birth (Hatch et al., 1998).

The previous studies involved women who were active before pregnancy, but
questions remained about the effects of starting an exercise program during pregnancy on
maternal-fetal health and birth outcome. A study evaluating the initiation of regular
exercise in early pregnancy was conducted by Clapp et al. in 2000. The women were
divided into two groups, exercise and controls. The exercising women participated in 20
minutes of weight—bearing activity (treadmill, step aerobics, or stair stepper) three to ﬁve
times a week at 55% to 60% of preconception aerobic capacity (Clapp etal., 2000). The

authors found that beginning a moderate physical activity regimen of weight bearing

47

exercise in early pregnancy results in enhanced fetoplacental growth. The offspring of
the exercising women were signiﬁcantly longer and heavier than those born to control
women (Clapp et al., 2000). The authors believed that this amount of exercise could
easily be performed by most women during a normal, low-risk pregnancy.

Not all studies conducted after the 1994 ACOG guidelines showed that maternal
exercise enhanced fetal growth and birth weight. In 1995, Bell, Palma, and Lumley
assessed the effect of continuing a Vigorous exercise program into late pregnancy on birth
weight (Bell, Palma, and Lumley, 1995). Study participants were women who had
participated in vigorous physical activity prior to pregnancy. A control group consisted
of healthy pregnant women who did not participate in regular vigorous exercise. Seven-
day food records and exercise diaries were kept at one time point during each trimester.
The results showed women who continued to participate in 30 minutes of vigorous
exercise more than four times a week at twenty-ﬁve weeks gestation gave birth to infants
315 grams lighter than those born to controls (Bell, Palma, and Lomley, 1995). The
authors did not discuss any possible confounders regarding birth weight, such as
pregravid weight and gestational weight gain. Also, even if babies born to vigorous
exercisers were lighter, there was no mention of an increased risk of low birth weight.
The decrease in mean birth weight may have indicated the exercising women gave birth
to babies who were leaner, but no less healthy than those born to controls.

In two separate studies, Clapp et al. found that exercising women delivered lighter
infants with less fat mass compared to babies born to controls (Clapp et al., 1999; Clapp
et al. 2002). The authors found although gestational age at delivery, axial length, and

head circumference were similar, the exercise group offspring were lighter and leaner

48

than those of the control group (Clapp et al., 1999). Clapp et al. (2002) also examined
the effects of exercise volume at different times during pregnancy. Currently exercising
women were randomly assigned to one of three groups after eight weeks gestation (Clapp
et al., 2002). One group performed moderate exercise throughout pregnancy, another
group decreased volume of activity, and the third group increased exercise volume in mid
to late pregnancy. Results showed that a high volume of moderate-intensity exercise
performed in mid and late pregnancy reduced fetoplacental grth symmetrically,
whereas reduction in exercise volume resulted in greater increase in fat mass and heavier
babies (Clapp et al., 2002). Although birth weight was lower as exercise volume
increased, there was no indication of increase risk of delivering a low birth weight infant.
In a review paper published in 1997, Stemfeld assessed the relationship between
physical activity and pregnancy outcome (Sternfeld, 1997). Proposed risks of
redistribution of uterine blood ﬂow, fetal hypoxia, hypertherrnia and teratogenic effects,
and increase risk of pro-term labor were addressed based on the dual stresses of
pregnancy and physical activity. Even after considering these potential risks, Stemfeld
stated that the literature generally shows neutral or somewhat positive effects of physical
activity on pregnancy outcome. Based on lack of evidence for harmful effects for healthy
well-nourished women, Stemfeld suggested exercise during pregnancy appears to be safe
and subject to few restrictions. This ﬁnding supports the overall tone of the revised 1994
ACOG recommendations (Sternfeld, 1997). A 1998 review by Pivarnik also stated that
the evidence at that time appeared to indicate participation in moderate to vigorous
activity throughout pregnancy may enhance birth weight, while more severe regimens

may result in light offspring (Pivarnik, 1998). In addition, Pivarnik indicated most

49

studies that had shown unfavorable birth outcomes had not addressed some important
potential confounders of birth weight (such as SES) and that physical activity had not
been well quantiﬁed throughout gestation. The author suggested that future studies
included randomized trials in which women with out a history of chronic physical
activity are assigned to an exercise or physical activity or control group (Pivarnik, 1998).
Two additional reviews regarding the effects of physical activity on birth weight stated
that the research has been inconclusive, primarily because many studies have examined
different types, timing, durations, and intensities of activity (Larsen-Meyer, 2002;
Johnston, 1991).

Based on knowledge gained from studies conducted after the 1994
recommendations, ACOG revised their physical activity guidelines again in 2002 (ACOG
2002). The main recommendation states that in the absence of either medical or obstetric I
complications, 30 minutes or more of moderate physical activity a day on most, if not all,
days of the week is recommended for pregnant women (ACOG, 2002). These guidelines
follow the physical activity recommendations for healthy adults put forth jointly by the
Centers for Disease Control and Prevention (CDC) and the American College of Sports
Medicine (ACSM) (Pate et al., 1995). The guidelines also address concerns for
competitive athletes and suggest regular exercise during pregnancy may provide
additional beneﬁts to women with gestational diabetes (ACOG, 2002).

The newest ACOG guidelines suggest pregnant women participate in “moderate
physical activity”, a term that has been difﬁcult to deﬁne and separate from “exercise”.
Physical activity is deﬁned as any bodily movement produced by skeletal muscle that

results in energy expenditure (Casperson et al., 1985). Physical activity is performed to

50

sustain life, but the amount is subject to personal choice and varies greatly from person to
person and over time. Exercise is a subcategory of physical activity because it is
planned, structured, repetitive, and purposeﬁrl in nature (Casperson et al., 1985). The
CDC/ACSM and ACOG guidelines may be easier for most pregnant women to meet due
to the fact that they recommend physical activity rather than structured exercise
programs, allowing more options.

In 2003, the Canadian Clinical Practice Obstetrics Committee and the Canadian
Society for Exercise Physiology Board of Directors created a Clinical Practice Guideline
for exercise in pregnancy and the postpartum period (Davies et al., 2003). The objective
was to design guidelines advising obstetric care providers of the maternal, fetal, and
neonatal implications of aerobic and strength conditioning exercises in pregnancy. The
authors completed a MEDLINE search from 1966-2002 for articles related to studies of
maternal aerobic and strength conditioning in previously sedentary populations as well as
in previously active populations, the impact of aerobic and strength conditioning on early
and late pregnancy outcomes, and neonatal outcomes (Davies et al., 2003). After
reviewing the literature six main recommendations were created 1) all women without
contraindications should be encouraged to participate in aerobic and strength
conditioning exercise as a part of a healthy lifestyle during pregnancy; 2) reasonable
goals of aerobic conditioning in pregnancy should be to maintain a good ﬁtness level
without trying to reach peak ﬁtness or train for an athletic competition; 3) women Should
chose activities that will minimize the risk of loss of balance and fetal trauma; 4) women
should be advised that adverse pregnancy or neonatal outcomes are not increased by

exercise; 5) initiation of pelvic ﬂoor exercises in the immediate postpartum period may

51

reduce the risk of future urinary incontinence; 6) women should be advised that moderate
exercise during lactation does not affect the quantity or composition of breast milk or
impact infant growth (Davies et al., 2003). These guidelines are similar to those written
by ACOG in 2002, recommending physical activity during pregnancy while providing
more detailed suggestions directed toward the health care community rather than the
pregnant woman.
Postpartum Weight Retention (PPWR) Reviews

Now that ACOG guidelines indicate a healthy pregnant woman can and should be
physically active during pregnancy, there has been a gradual shift in research to focus on
potential beneﬁts of exercise during pregnancy beyond the perinatal period. For
example, long-term effects of pregnancy on a woman’s body size and lifestyle have yet to
be determined. Some research has shown weight gained and retained after one pregnancy
can increase with successive pregnancies (Rookus et al., 1987; Rooney and Schauberger,
2002; Greene et al., 1988). As weight increases a woman may become at risk for
overweight or obesity after her childbearing years. Assessing the effects of the many
variables that may play a role in postpartum weight retention may be difﬁcult. There
have been at least 31 factors identiﬁed in the relationship between pregnancy and weight
retention. Examples include parity, pre-gravid weight, maternal pregnancy related
weight gain, age, race, lactation, and physical activity (Johnson, 1991; Linne, Barkeling,
and Rossner, 2004). Including all relevant factors in a study warrants a complex design.
In addition different measurement tools may be used for the same variables (e.g., physical

activity and lactation) leading to mixed and/or incomparable results.

52

In a 1991 review by Johnston, the author discussed trends in gestational weight
gain since the 1980’s, methodological problems in the assessment of weight changes in
pregnancy, the proposed factors involved with gestational weight gain and retention, and
weight changes in the postpartum period (Johnston, 1991). This detailed review included
information from 204 published studies. The author concluded that reports on weight
loss during the postpartum period indicate a large degree of variability in total weight
change. Overall, the evidence suggested each successive birth adds approximately one
kg of body weight above that normally gained with age (Johnston, 1991).

In 1995, Crowell reviewed approximately 37 studies on weight change in the
postpartum period. Study topics ranged from weight gain and fetal growth, to factors
affecting postpartum weight loss. The purpose of this review was to assemble the
literature so a nurse-midwife could provide sound, clinical advice to pregnant women at
risk for postpartum weight retention (Crowell, 1995). The main conclusions were that a
woman needs six months to remove her pregnancy related weight gain and that women
with average weight gain during pregnancy retain one kg after each child. The same
average weight retention was also noted in the Johnston review and both values are above
the 0.2 kg gained per year due to age. In addition, postpartum weight loss occurs earlier
and to a greater degree in women of lower parity, age, and pre-pregnancy weight
(Crowell, 1995). Both the Johnston and Crowell reviews state that an average amount of
pregnancy related weight retention is one kg, but the variation among women may range
from -12. to +265 kg (Ohlin and Rossner, 1990). Other studies have also shown a

similar average in weight retention, along with a similarly wide variation. A study by

53

Ohlin and Rossner (1990) showed average weight retention to be 0.5 kg, though 14% of
the sample retained 5 kg or more (Ohlin and Rossner, 1990).

Larson-Meyer reviewed studies that examined the effect of postpartum exercise
on mothers and their offspring (Larson-Meyer, 2002). The studies were grouped into
categories of those addressing the effect of postpartum exercise on changes in body
weight, body composition, and energy balance (n=6) and those addressing the inﬂuence
of maximal and submaximal exercise on human-milk lactic acid concentration (n=9)
(Larsen-Meyer, 2002). Most studies included 25 or fewer subjects who were followed
for six months or less. The main ﬁndings were that postpartum exercise can improve a
woman’s aerobic ﬁtness, insulin sensitivity, high-density lipoprotein-cholesterol level,
and psychological well-being. It was not clear whether postpartum exercise enhances
body weight or fat loss. As for lactating women, regular exercise had not been shown to
promote great weight loss, though a number of studies suggested exercise has no adverse
effects on milk production (Larsen-Meyer, 2002).

Postpartum Weight Retention Studies

McKeown and Record (195 7) published one of the ﬁrst papers that evaluated
pregnancy and postpartum weight development. The authors conducted a prospective
study following women’s weight changes from 3-24 months postpartum. Mean weights
were plotted for age and parity showing three consistent trends; an increase in weight
with age, an increase in weight increment with age, and an increase in weight increment
with parity. The authors hypothesized a biological origin where increases in weight of
women who have their ﬁrst child after 35 years of age may be due to an association

between obesity and low fertility (McKeown and Record, 1957). Since the McKeown

54

and Record study, many others have determined the strongest indicator of weight
retention after a pregnancy is the amount of weight gained during that pregnancy
(Rooney and Schauberger, 2002; Ohlin and Rossner, 1995; Schauberger etal., 1992;
Ohlin and Rossner, 1990; Greene et al., 1988; Muscati et al., 1995; Linne and Rossner,
2003).

Using a prospective design, Greene et al. analyzed interpregnancy weight change
and found that the mean weight gain between pregnancies was 0.9 kg and on average
women retained 0.9-2.3 kg between pregnancies after accounting for age related weight
gain (Greene et al., 1988). Study subjects were women who had participated in the
Collaborative Perinatal study in the United States from 1959-1965. From the original
58,760 women, the sample was restricted to 7,116 who had enrolled for more than one
pregnancy and had singleton births. The study conﬁrmed previous ﬁndings that women
retain some of their pregnancy related weight gain. Half the participants retained two
pounds in the two years between the beginning of one pregnancy and the beginning of
another (Greene et al., 1988). The authors also found that at least one in ten women
retained 15 pounds or more associated with pregnancy, and pregnancy weight gains of 20
pounds or more were associated with postpartum weight retentions (Greene et al., 1988).
PPWR, Gestational Weight Gain, and Race

Many studies have shown that postpartum weight retention is related to
gestational weight gain, those that gain more often retain more (Ohlin and Rossner, 1990,
1994, Greene et al., 1988, Rooney and Schauberger, 2002). Past and current guidelines
have reﬂected the tension between recommending the appropriate amount of weight gain

for fetal development, without resulting in signiﬁcant postpartum weight retention. Prior

55

to the 1960’s, average recommended weight gain during pregnancy was between 8-9 kg
(Crowell, 1995). In 1970, the Committee on Maternal Nutrition recommended a higher
gestational weight gain (10.9 kg) based on evidence that maternal weight restriction had
negative effects on infant morbidity and mortality rates (Ohlin and Rossner, 1999). In
1990, the Institute of Medicine (IOM) created new recommendations based on a
woman’s pre-pregnancy BMI (IOM, 1990). The BMI groups and weight ranges include;
low pregravid weight (BMI <19.8; 12.5-18.0 kg), normal pregravid weight (BMI 19.8-
26.0; 11.5-16.0 kg), high pregravid weight (BMI >26.0-29.0; 7.0-11.5 kg), and very high
pregravid weight (BMI >29.0; ~ 6.8 kg).

Johnston believed that the newest IOM recommendations are based on optimal
infant outcomes rather than maternal health considerations (Johnston, 1991). Lederrnan
investigated the outcomes of studies analyzing the effects of the IOM weight gain
recommendations on postpartum weight retention and birth weight (Lederman, 2001).
She concluded that weight gains above recommended levels were associated with
increased maternal fat and pregnancy complications. The author suggested that health
care providers can help reduce obesity risk by monitoring maternal weight; promoting
healthy pre-pregnancy weight, pregnancy weight gain and postpartum weight loss; and
encouraging the maintenance of an active postpartum lifestyle (Lederman, 2001). The
question remains whether the recommendations work for all populations. It is unclear
whether all races would beneﬁt from following the IOM guidelines for gestational weight
gain, or whether cultural differences play a role. It should be noted that the current 1990
IOM guidelines suggest that adolescents and Black women should strive to gain weight at

the upper end of the recommended range.

56

In 1995, Boardley et al. investigated the relationship between maternal factors
(including diet and physical activity), and postpartum weight change by race. Study
participants were Black and White women seven to twelve months postpartum, who
attended special supplemental food programs for Women, Infant, and Children (WIC)
nutrition. The results showed Black women retained more postpartum weight than
Whites. The overall sample showed breast-feeding was not related to postpartum weight
change, but signiﬁcant differences in breastfeeding were found, between the races. Other
cultural or lifestyle differences were found, including time to return to work, marital
status, and smoking. Black women also had lowered physical activity scores than Whites
(Boardley et al., 1995). These race related differences are supported by other studies also
assessng the IOM guidelines and their effect of postpartum weight retention (Parker and
Abrams, 1993; Keppell and Taffel, 1993; Schieve et al., 1998). Taken together, results of
these studies suggest lifestyle factors related to race have a signiﬁcant impact on
postpartum weight retention beyond gestational weight gain.

PP WR and lactation

The effect of lactation on postpartum weight retention has been investigated
because of its role on maternal caloric expenditure and subsequent energy intake. The
average woman utilizes an additional 300 calories a day when breastfeeding, but whether
that extra caloric expenditure results in greater weight loss has not been clearly
established (Brewer et al., 1989; Dewey et al., 1993). Lactation data from the McKeown
and Record study Showed consistent, but not large differences in weight based on
duration of exclusive breast-feeding. Between months 3-12, weight loss increased

slightly as lactation duration increased in the women who breastfed for 3 months or

57

greater, but this effect was almost eliminated by 24 months (McKeown and Record
1957)

Rookus et al. (1987) investigated the effect of pregnancy and lactation on body
mass index (BMI) nine months postpartum using a case-control design. Changes in BMI
of 49 women (from pre-partum through nine months postpartum) were compared with
BMI changes over the same period of time in a control group of 400 non-pregnant
women. Lactation results showed that women who breastfed for more than two months
gained 0.6 kg-m2 more weight than the non pregnant women, whereas women who used
bromocriptine to stop lactation lost body mass (Rookus et al., 1987). The results
suggested that maternal obesity may be associated with breastfeeding for long periods of
time. The authors proposed that women who breastfeed may increase dietary intake
beyond what is needed to provide sufﬁcient energy to sustain breastfeeding, thus
nullifying the effect of lactation on weight loss (Rookus et al., 1987). Use of a
nonpregnant control group was different from most previous studies, and the lactating
comparison sample of 49 women was not large. Most previous studies have included at
least 100 to 200 subjects (Haiek et al., 2001; Walker and Graves, 1997; Brewer et al.,
1989)

Haiek et al. (2001) assessed the rate of a nine-month postpartum weight loss
differences in 236 women, according to infant feeding method. Infant feeding methods
included exclusive breast-feeding, mixed feeding, and predominantly bottle-feeding. The
objective of the study was to clarify whether women from various socioeconomic and
ethnic backgrounds lose weight at different rates as a function of lactation status. Most

participants lost weight, but at 182 days postpartum, 31 of 181 weighed more than

58

immediately after delivery (Haiek et al., 2001). Results also showed infant feeding
method did not affect rate of postpartum weight loss. These results supported those of
Rookus et al. (1987), and the authors suggested that breast-feeding women might eat
more, and therefore lose weight more slowly. The authors also suggested that breast-
feeding mothers may differ from bottle-feeding mothers in important attitudinal aspects
(Haiek et al., 2001).

The results of Haiek et al. (2001) are supported by another study examining bottle
vs. breast-feeding women. Walker and Graves explored the relationship between
lifestyle variables with postpartum weight gain and body image attitudes of bottle- and
breast-feeding women (Walker and Graves, 1997). A mail survey was sent to 513
women whose names were obtained from newspaper birth announcements, and
participants were eligible if they resided in Austin, Texas, had a singleton birth, and a
published mailing address. Of the 513 surveys mailed, 245 were returned and 207 met
the exclusion criteria. Items collected ﬁom the survey included anthropometric,
reproductive, and sociodemographic variables (Walker and Graves, 1997). No
signiﬁcant differences were found between breast and bottle-feeding groups for pre-
pregnancy body mass index, gestational weight gain, time since delivery, or social
variables. Breastfeeding women were more educated and more likely to be living with
the infant’s fathers. The results showed feeding method was not associated with
postpartum weight gain in the sample as a whole. Bottle-feeding women with higher
PPWR exercised less, had higher fat intakes, and were more dissatisﬁed with body image

than mothers with lower gains. Breast-feeding mothers with high or low PPWR did not

59

differ on any lifestyle factors (Walker and Graves, 1997). This study was a retrospective
cross sectional design (four months postpartum) that relied on self-reported data.

Brewer et al. (1989) conducted a study analyzing postpartum changes in maternal
weight and body fat depots in lactating versus non-lactating women. No signiﬁcant
differences were found between the groups at three months postpartum. The authors
suggested that greater insulin sensitivity in lactating women may enable them to better
conserve energy for use in milk production. The authors also stated that three months
may be the turning point in metabolic efﬁciency (insulin sensitivity and energy
conservation for milk production) for lactating as well as non-lactating women. This
supports the ﬁnding of Rooney and Schauberger (2002) suggesting that women who
breastfeed for more than 12 weeks may have less long-term weight retention or gain. The
authors suggest lactation plays a role in postpartum weight and body fat loss, but that
current RDA values for total calories may be too high to permit such losses (Brewer et
aL,1989)

Though most results Show no Signiﬁcant weight loss change with breast-feeding
women, some investigators have found effects of lactation on postpartum weight loss.
Dewey et al. (1993) examined maternal weight loss patterns during prolonged lactation.
Subjects included mothers participating in the DARLING study, a longitudinal
comparison of growth, nutrient intake, and morbidity of matched cohorts of breast-fed
and formula-fed infants (Dewey et al., 1993). The women were divided into groups
based on breast-feeding practices. The breast-feeding group consisted of 46 mothers
whose infants received less than 120 m1 of other milks until 12 months of age. The

formula-feeding group consisted of 39 mothers whose infants were weaned completely to

60

formula by three months. Weight and body fat were monitored through 24 months in
both groups. The results showed no difference in weight loss at 1-3 months postpartum,
but breast-feeding women showed a signiﬁcant difference in weight loss from 3-6
months, suggesting women should breastfeed for six months or greater if they expect
lactation to enhance weight loss. The authors suggested that their study results indicate
lactation promotes weight loss during the ﬁrst year postpartum and that the differences
found in other studies maybe a function of inadequate study design (Dewey et al., 1993).

In 1997, Janney et al. performed a longitudinal study examining lactation and
weight retention between infant feeding methods through 18 months postpartum. One
hundred ten 20-40 year-old women were recruited during their third trimesters from
birthng education and obstetric practices in Ann Arbor, Michigan (Janney et al., 1997).
Weight was evaluated at two-weeks, two-months, four-months, six-months, and 12 and
18 months postpartum. Information about feeding practices, reproductive history, diet,
and physical activity were collected at each visit. Regardless of infant feeding method,
all participants lost weight from two-weeks to 18-months postpartum, with slower rates
of weight loss occurring after lZ-months. Overall, less weight was retained by the
lactating women than their non-lactating counterparts. Even women who breast-fed for
less than four months retained less weight than women who bottled-fed their infants.
Once lactation was discontinued, the rate of weight loss slowed. The breast-feeding
women achieved their pre—pregnancy weights approximately six months before those who
bottle-fed their infants (Janney et al., 1997).

PPWR and physical activity

61

Physical activity is another lifestyle factor that may play a role in PPWR. The
effects of physical activity on PPWR appear mixed, but differences in study design,
measurement tools, and outcome measures may explain those results. Larson-Meyer
(2002) reviewed approximately 60 cross-sectional and randomized control trials from the
past 12 years, addressing the effect of postpartum exercise on weight loss and or energy
balance in women, most of whom were lactating (Larson-Meyer, 2002). Overall, the
studies suggest that moderate exercise without calorie restriction does not promote
greater weight or fat loss. The author speculated that exercise may promote greater
caloric intake and or reduced energy expenditure from non-physical activities, preventing
a negative energy balance. Larson-Meyer concluded that beneﬁts of exercise during the
postpartum period are important it is not clear whether the physical activity promotes
greater body-weight or body-fat loss after childbirth (Larson-Meyer, 2002). This review
only included studies evaluating postpartum exercise, not exercise during pregnancy, and
its effect on PPWR. Since the early 1990’s there have been more studies with
prospective and longitudinal designs assessing the relationship between physical activity
during and after pregnancy on PPWR.

In 1994, Ohlin and Rossner assessed trends in eating patterns, physical activity,
and sociodemographic factors in relation to postpartum body weight development. This
prospective investigation was part of the Stockholm Pregnancy and Weight Development
Study, where 1493 women were followed from the beginning of one pregnancy to 1 year
postpartum. The study was designed to identify risk factors for postpartum weight
retention using data collected from routine pregnancy records and questionnaires at six

and twelve months postpartum (Ohlin and Rossner, 1994). Mean weight retention was

62

0.5kg above pre-pregnancy weight, while 14% of the sample retained 5 kg or more. The
authors found that weight retention one year postpartum was greater in women who a)
increased their energy intake during and after pregnancy, b) increased their snack eating
after pregnancy, c) decreased their lunch frequency during and after pregnancy. Women
who retained greater or equal to 5 kg at one year postpartum were less physically active
in their leisure time when compared with women with lower postpartum weight gains
(Ohlin and Rossner, 1994). Of the women who gained the most weight (>10 kg), 23%
were inactive in their leisure time throughout all study time periods, compared with 4%
of those with lower weight gains. The results indicated that nutrition habits and physical
activity during the one-year postpartum follow-up were important for facilitating weight
loss. The authors also believed that the degree of leisure time activity might, in addition
to affecting energy expenditure, also reﬂect a greater health consciousness (Ohlin and
Rossner, 1994).

Harris et al. (1997) examined the impact of pregnancy on long-term weight gain
of primiparas in England and identiﬁed potential risk factors for maternal obesity. The
study design was retrospective involving a repeat-pregnancy and examination of change
in maternal body weight from the beginning of one pregnancy to the beginning of a
second. Two hundred forty-three women with singleton births participated in the study.
The results Showed no signiﬁcant long-term increase in mean maternal body weight
following the ﬁrst pregnancy, and while most mothers retained 1.0kg or less, 24.7 %
retained 1.54 kg or more. Mothers with higher BMIS at the beginning of the ﬁrst
pregnancy, who gained more weight during pregnancy, gained signiﬁcantly more weight

from one pregnancy to the next (Harris etal., 1997). The same mothers also gave birth to

63

heavier babies and had longer intervals between pregnancies. The authors concluded that
pregnancy may be associated with a permanent increase in maternal body weight simply
because it is a time of positive energy balance when some women gain excessive weight.
The authors did not speculate as to the cause, but stated that other factors (such as pre-
pregnant BMI) determine whether long-term weight gain actually occurs. The authors
also suggested that higher weight gains of some women may have been the result of
fundamental differences in physiology and or lifestyle that put them at higher risk for
excessive weight gain, irrespective of pregnancy (Harris et al., 1997).

Sampselle et al. (1999) studied the role of exercise and postpartum well-being by
evaluating reported patterns of postpartum physical activity and identifying beneﬁts or
risks associated with that activity at six weeks after delivery (Sampselle et al., 1999).
Approximately 1000 women completed a questionnaire at a Six week clinic visit. Body
weight measured at the clinic was then compared to reported pre-pregnancy weight. The
results showed women with higher physical activity levels retained less weight than their
less active counterparts. Participation in vigorous exercise at six weeks postpartum was
signiﬁcantly associated with psychosocial well-being. The authors stated that although
the outcomes of physical activity have been positive in the general population, exercise
has rarely been an element of the postpartum care plans (Sampselle et al., 1999). This
study was part of a large longitudinal study, which suggests results might be more
generalizable compared to those obtained from cross-sectional design.

In 2002, Rooney and Schauberger investigated the relationship between excess
pregnancy weight gain and long-term obesity, a decade after delivery. The original study

was conducted in Wisconsin from 1989-1990. Seven hundred, ninety-ﬁve women were

64

observed through pregnancy and six months postpartum to examine factors that affect
weight loss. The women completed surveys at 2, 4, 6, 8, 12 and 24 weeks postpartum on
breastfeeding, use of tobacco or alcohol, sexual activity, return to work, and frequency of
exercise. In 1999, a follow-up was conducted and weights were obtained through
medical record review (Rooney and Schauberger, 2002). The average weight change
from pre-pregnancy to follow-up was a gain of 6.3 kg. Pre-pregnancy weight gain did
not inﬂuence weight gain, but there was a trend of increased weight retention based on
pregnancy-related weight gain. The results also showed women who breastfed and
participated in aerobic exercise had signiﬁcantly lower weight gains 8-10 years
postpartum. The authors concluded that breast-feeding and exercise may be beneﬁcial to
control long-term weight (Rooney and Schauberger, 2002).

The Stockholm Pregnancy and Women’s Nutrition (SPAWN) study is the longest
prospective examination of the relationship between weight gain and retention after
pregnancy (Linne and Rossner, 2003). Five hundred sixty-three of the original 1463
women participated in the study follow-up. Participants ﬁlled out questionnaires and
anthropometric data were collected. The sample was divided into four groups based on
factors such as number of years between pregnancies, pre-pregnancy BMI, weight gain
during pregnancy, and weight change one year postpartum (Linne and Rossner, 2003).
The results showed no differences between weight gain during pregnancy and retention in
women who had more than two years between their ﬁrst and second pregnancies.
Overweight women (BMI > 25.0) did not gain more weight during pregnancy or retain
more weight at one-year after the ﬁrst and second pregnancy than women with lower

BMIS. Women who were high pregnancy weight gainers (14 kg or more) and retainers

65

(1.8 kg or more) gained more weight during the ﬁrst pregnancy and retained that weight
into the second pregnancy. The same held true for those women for weight gain and
retention after the second pregnancy (Linne and Rossner, 2003). Overall, women who
were overweight before pregnancy did not have higher risk of postpartum weight
retention than normal weight women. Weight gain during pregnancy and weight
retention up to one year after delivery was predictive for weight development in the
second pregnancy, but the correlation was not extremely high (Linne and Rossner, 2003).
Based on the same SPAWN data, Linne et al. (2004) evaluated how well pre-
pregnancy BMI, gestational weight gain, and postpartum weight retention predict
retention of weight after 15 years. Results showed overweight women did not gain more
weight than normal weight women at one-year follow-up. High weight gainers during
pregnancy retained more weight at the one and 15-year follow-ups. 56% of the high
weight gainers during pregnancy ended up in the high weight retainers group. Overall,
the data showed that weight retention at the end of the postpartum year predicts future
overweight 15 years later. The authors suggested that Since the women who had not lost
the weight by one year postpartum were also those who retained weight at 15-years after
delivery, one-year postpartum represented a suitable time period for intervention. The
authors noted that small changes in everyday lifestyle are important determinants of long-
terrn weight gain and the effects of those changes may be too small to measure by
questionnaire. In addition, lack of measurement of such factors as consumption of
alcohol, sweets, etc was a methodological weakness (Linne et al., 2004). The idea that
pregnancy-related lifestyle changes are important is supported by a Rossner and Ohlin

study of the factors that may link pregnancy to long-term weight retention and obesity.

66

This study was also based on SPAWN data and suggested that body weight changes,
measured one year after delivery, depend more on lifestyle changes during and after
pregnancy than by maternal body weight before conception (Rossner, and Ohlin, 1995).
Pregnancy and Postpartum E ﬂeets on Lifestyle Behaviors

Investigators can suggest that lifestyle factors such as nutrition, physical activity,
and alcohol consumption play a role in long-term weight change, but they also need to
evaluate what those lifestyle factors are, and what physiological effects they might have
on future weight development. lnforrnation from the Australian Longitudinal Study of
Women’s Health indicates that young women with children are particularly unlikely to be
“adequately active”, which the authors considered deﬁned as at least once a week. Less
than half as many women with children engaged in adequate physical activity compared
to women without children (Brown et al., 2000). These trends have also been
demonstrated in Canadian studies which revealed that mothers with young children are
the least likely population group to exercise (Brown et al., 2000).

Harris, Ellison, and Clement (1999) designed a longitudinal study to determine
the psychosocial and behavioral changes resulting from motherhood and their effects on
long-term weight gain (Harris, Ellison, and Clement, 1999). Seventy-four women
participated in this study (Antenatal Care Project) which was a follow-up of an original
randomized control study of antenatal care in south London, England. The researchers
collected body weight data at the beginning of pregnancy, lifestyle and behavior variables
during pregnancy, antenatal and obstetric care history, postpartum measures of postnatal

depression, and measures of parenting stress. Postpartum height, weight and other

67

lifestyle change information were gathered 30 months postpartum (Harris, Ellison, and
Clement, 1999).

Results showed that weight gains consisted of weight gained during pregnancy
and 2-3 years postpartum. The results showed lifestyle changes that accompany
pregnancy and motherhood increase some women’s vulnerability to eating disorder
psychopathology. Women who felt they did less exercise after pregnancy than before
were at a greater risk of long-term weight gain, as were those with few supportive
individuals (Harris, Ellison, and Clement, 1999).

In 2002, Miller et al. examined the mediators of physical activity behavior change
among women with small children. The investigators used an intervention trial with a
control group, a group given print information about overcoming barriers of physical
activity, and a group that was also invited to discuss development of local area strategies
for the promotion of physical activity among mothers with young children (Miller et al.,
2002). After controlling for baseline physical activity, residualized changes in self-
efﬁcacy and partner support signiﬁcantly predicted whether a woman would meet
physical activity guidelines. The researchers concluded that partner support and self-
efﬁcacy may be mediators of physical activity behavior change (Miller et al., 2002).
Self-eﬂicacy and Physical Activity Behaviors

Self-efﬁcacy is one of the most frequently identiﬁed psychological determinants
of adherence to physical activity (McAuley and Mihalko, 1998). Self-efﬁcacy beliefs are
deﬁned as peoples’ judgments concerning their abilities to perform Speciﬁc tasks
successfully in the face of speciﬁc obstacles to obtain Speciﬁc outcomes. These

judgments are inﬂuenced by diverse sources of efficacy information (Bandura, 1977).

68

Self-efﬁcacy is a dynamic construct that can be inﬂuenced by cognitive, physiological,
behavioral, and social information (McAuley and Mihalko, 1998). Exercise efﬁcacy
refers to measures directed at the assessment of beliefs regarding individuals’ capabilities
to engage successfully in incremental bouts of physical activity. Self-efﬁcacy beliefs for
exercise are important because they are thought to play a signiﬁcant role in an
individual’s decision, effort, and persistence to continue to exercise (F eltz, 1988).

DuCharme and Brawley (1995) examined the inﬂuence of various aspects of self-
efﬁcacy on the exercise attendance of novice participants. Study participants were sixty-
three healthy adult females who enrolled for the ﬁrst time at a large women’s ﬁtness club.
Barrier self-efﬁcacy, percent conﬁdence, scheduling self-efﬁcacy, behavioral intention,
and exercise behavior were assessed during weeks one and nine of a 16-week exercise
program (DuCharme and Brawley, 1995). Results indicated that both barrier and
scheduling self-efﬁcacy Signiﬁcantly predicted behavioral intention throughout the
exercise program. The results of this study support Bandura’s self-efﬁcacy theory in that
efﬁcacy predicted attendance behavior of adherent novice exercisers (DuCharme and
Brawley, 1995).
Self-eﬂicacy and Physical Activity Behaviors During and After Pregnancy

It is not known whether pregnancy alters the role of self-efﬁcacy in a woman’s
exercise behaviors. In 1989, Godin, Vezina, and Leclerc conducted a study to identify
the factors that may inﬂuence a pregnant woman’s decision to exercise after delivery
(Godin, Vezina, and LeClerc, 1989). Ninety-eight pregnant women completed a
questionnaire regarding attitudes, social norms, perceived barriers to exercise, and

intentions to exercise after delivery. Differences were found in intentions to exercise

69

between nulliparous and pluriparous pregnant women. The pluriparous women were
more realistic in deﬁning their intentions to exercise after delivery, emphasizing the
importance of barriers to exercise based on past postpartum experiences (Godin, Vezina,
and LeClerc, 1989). The investigator did not follow the participants after delivery to
analyze the actual behaviors produced by the intentions.

In 2001, Beilock, Feltz, and Pivarnik studied psychosocial variables related to
postpartum exercise behavior in female athletes. Athletes’ (N=26) training patterns
before, during, and after pregnancy were examined. The authors used a recall survey to
examine exercise patterns and psychological variables mediating a return to training and
competition after pregnancy (Beilock et al., 2001). Potential study subjects had to have
at least four years of competitive athletic experience prior to a pregnancy that occurred
within the past ten years. The questionnaires included the Training Patterns
Questionnaire, a perceived barriers questionnaire, the Social Provisions Scale, and a
follow-up telephone interview examining the types of social support needed during the
postpartum period. The number of participants who continued to train during pregnancy
decreased overtime from the ﬁrst to the third trimester, while even fewer competed
postpartum (Beilock et al. 2001). Lack of energy was reported as a primary barrier to
training both during pregnancy and postpartum. Lack of time was perceived as the most
signiﬁcant barrier to training in the postpartum. Overall, the list of perceived ban'iers to
training postpartum included lack of time, energy, interest, motivation, as well as
breastfeeding (sensitive breasts), weight gain, and worry about returning to training too
soon. Participants indicated overwhelmingly that childcare was the form of social

support most needed (Beilock et al., 2001).

70

In 2004, Downs and Hausenblaus performed a retrospective examination of
exercise beliefs and behaviors during pregnancy and postpartum. Women who were one-
year postpartum were recruited from a private practice ofﬁce in New Britain, Connecticut
(Downs and Hausenblaus, 2004). Leisure time activity and exercise beliefs were
collected from 74 women ranging ﬁom six days to ﬁve months postpartum. The most
common beliefs during pregnancy were that exercise improves mood, and physical
limitations (e.g. nausea) obstructed exercise participation. The most common beliefs
during the postpartum period were that exercise controlled weight gain and a lack of time
obstructed exercise participation (Downs and Hausenblaus, 2004).

The role of self-efﬁcacy on exercise behavior is an important factor that plays a
Signiﬁcant role in a woman’s choice to participate in physical activity. Perceived barriers
to exercise may inﬂuence her intentions to as well as actual participation in physical
activity. The question remains of whether this relationship holds for pregnant women
and those who have recently given birth. It is not known whether women who are active
before pregnancy are more likely to remain active during pregnancy and through the
postpartum period or whether activity affects long-term weight changes. More studies
are needed to establish a clear role of self-efficacy and a woman's belief to overcome
barriers to physical activity in actual activity participation beyond gestation.

Summary

Major points of emphasis for this literature review are that physical activity is
safe, beneﬁcial for both mother and fetus, and recommended for women with healthy,
normal, singleton pregnancies. Early animal and human studies focused on the

physiological effects of physical activity on the maternal-fetal unit. Most study ﬁndings

71

suggest that a woman’s physiology adapts during pregnancy in order to better dissipate
heat, handle the stress of physical activity, and improve fetoplacental function. As the
amount of research increased regarding the risks of physical activity during pregnancy,
the ACOG has adapted their guidelines to reﬂect current knowledge. Current
recommendations for a woman with a normal, healthy, uncomplicated singleton
pregnancy strongly encourage moderate physical activity for 30 minutes most days of the
week. This recommendation is similar to that proposed for the general population.

It has been Shown that on average a woman will retain 1 kg, beyond weight
gained due to age, for each pregnancy. However, the range is wide and can exceed 20kg
in some women. The relationship between physical activity during pregnancy and
postpartum weight change has been difficult to establish. Factors such as race, lactation,
and weight gain during pregnancy have also been found to effect postpartum weight
retention. Differences in study design, population, measurement of physical activity and
lactation, and postpartum follow-up period have made clear relationships hard to deﬁne.
Some long-term studies, including 10 years or more of follow-up, have shown that
certain lifestyle factors, such as physical activity have played a role in weight gain and
retention after delivery. Psychosocial factors regarding physical activity and pregnancy
may also play a role in postpartum weight retention. A woman’s self-efficacy has been
shown to effect exercise participation and the ability to overcome barriers that arise
during the postpartum period. The need exists to further assess these relationships
between such lifestyle factors and weight retention. A long-term follow-up that includes
original physical activity, lactation, and other lifestyle data may be able to add to the

current knowledge base. Learning about the barriers to physical activity for postpartum

72

women may also increase our ability to create programs and interventions to increase, if

not maintain physical activity participation.

73

References

American College of Obstetricians and Gynecologists (1985) Exercise during pregnancy
and the postpartum period. Technical Bulletin 58. Washington, DC: ACOG Press.

American College of Obstetricians and Gynecologists (1994) Exercise during pregnancy
and the postpartum period. Technical Bulletin 129. Washington, DC: ACOG Press.

American College of Obstetricians and Gynecologists (2002) Exercise during pregnancy
and the postpartum period. Technical Bulletin 26 7. Washington, DC: ACOG Press.

Astrand P, and Rodahl K (1970) Textbook of work physiology. McGraw-Hill Book
Company: New York.

Artal R, Platt L, Sperling M, Kammula R, Jilek J, Nakarnura R (1981) Maternal
cardiovascular and metabolic responses in normal pregnancy. Am J Obstet Gynecol 140:
123-7.

Artal R, Wisewell R, and Drinkwater B (1991) Exercise in pregnancy. 2nd edition,
Williams and Wilkins: Baltimore.

Bagnall K, Mottola M, and McFadden K (1983) The effects of strenuous exercise on

maternal rates and their developing fetuses. Canadian Journal of Applied Physiology 8:
254-9.

Bandura A (1977) Self-efﬁcacy: toward a unifying theory of behavioral change.
Psychological Review 84: 191-2 15.

Beilock S, F eltz D, Pivarnik J (2001) Training patterns of athletes during pregnancy and
postpartum. Research Quarterly for Exercise and Sport 72: 1-8.

Bell R, Palma S, and Lumley J (1995) The effect of vigorous exercise during pregnancy
on birth-weight. Aust NZ J Obstet Gynecol 35: 46-51.

Berkowitz G, Kelsey J, Holford T, and Berkowitz R (1983) Physical activity and the risk
of Spontaneous preterm delivery. Journal of Reproductive Medicine 28: 581-8.

Boardley D, Sargent R, Coker A, Hussey J, and Sharpe P (1995) The relationship

between diet, activity, and other factors, and postpartum weight change by race. Obstet
Gynecol 86: 834-8.

Brewer M, Bates M, and Vannoy L (1989) Postpartum changes in maternal weight and
body fat depots in lactating vs. nonlactating women. Am J of Clin Nutr 49: 259-65.

74

Brown W, Mishra G, Lee C, and Bauman A (2000) Leisure time physical activity in
Australian women: relationship with well-being and symptoms. Res Q Exerc Sport 71 :
206-16.

Casperson C, Powell K, and Christenson G (1985) Physical activity, exercise, and
physical ﬁtness: deﬁnitions and distinctions for health-related research. Public Health
Reports 100: 126-31.

Clapp J (1987) The effects of exercise on uterine blood ﬂow. In CR Rosenfeld (ed):
Uterine Blood Flow. Perinatology Press: Ithaca, pp.300-10.

Clapp J, and Rizk K (1992) Effect of recreational exercise on mid-trimester placental
growth. Am J of Obstet and Gynecol 159: 1456-60.

Clapp J (1994) A clinical approach to exercise during pregnancy. Clin Sports Med 13:
443.

Clapp J, Lopez B, and Harcar-Sevcik R (1999) Neonatal behavioral proﬁle of the

offspring of women who continued to exercise regularly throughout pregnancy. Am J
Obstet Gynecol 180: 91-4.

Clapp J (2000) Exercise during pregnancy: a clinical update. Clin Sports Med 19: 273-
85.

Clapp J, Kim H, Burciu B, and Lopez B (2000) Beginning regular exercise in early
pregnancy: effect on fetoplacental growth. Am J Obstet Gynecol 183: 1484-8.

Clapp J, Kim J, Burciu B, Schmidt S, Petry K, and Lopez B (2002) Continuing regular
exercise during pregnancy: effect of exercise volume on fetoplacental growth. Am J
Obstet Gynecol 186: 142-7.

Crowell D (1995) Weight change in the postpartum period: a review of the literature.
Journal of Nurse-Midwifery 40: 418-23.

Davies G, Wolfe L, Mottola M, MacKinnon C (2003) Joint SOCG/CSEP clinical practice
guideline: exercise in pregnancy and the postpartum period. Canadian Journal of
Applied Physiology 28: 329-41.

Dewey K, Heinig M, and Nommsen L (1993) Maternal weight-loss patterns during
prolonged lactation. Am J Clin Nutr 58: 162-6.

Downs D, and Hausenblas H (2004) Women’s exercise beliefs and behaviors during their
pregnancy and postpartum. Journal of Midwifery and Women ’s Health 49: 138-44.

DuCharme K, and Brawely L (1995) Predicting the intentions and behavior of exercise
initiates using two forms of self-efﬁcacy. Journal of Behavioral Medicine 18: 479-97.

75

F airbaim J S (1924) Gynecology with obstetrics. Oxford Medical Publications: London.
Fairbaim J S (1926) Obstetrics. Oxford Medical Publications: London.

F eltz D (1988) Self-conﬁdence and sports performance. Exercise and Sport Science
Reviews 16: 423-57.

Greene G, Smiciklas-Wright H, Scholl T, and Karp R (1988) Postpartum weight change:
how much of the weight gained in pregnancy will be lost after delivery? Obstet Gynecol
71: 701-7.

Godin G, Vezina L, and LeClerc O (1989) Factors inﬂuencing intentions of pregnant
women to exercise after giving birth. Public Health Reports 104: 188-95.

Haiek L, Kramer M, Ciampi A, and Tirado R (2001) Postpartum weight loss and infant
feeding. Journal of the American Board of Family Practice 14: 85-94.

Harris H, Ellison G, Holliday M, and Lucassen E (1997) The impact of pregnancy on the
long-term weight gain of primiparous women in England. Int J of Obesity 21: 747-55.

Harris H, Ellison G, and Clement S (1999) Do the psychosocial and behavioral changes
that accompany motherhood inﬂuence the impact of pregnancy on long-term weight
gain? Journal of Psychometric Obstet Gynecol 20: 65-79.

Hatch M, Levin B, Shu X, and Susser M (1998) Maternal leisure-time exercise and
timely delivery. Am J of Public Health 88: 1528-33.

Institute of Medicine (1990) Subcommittee on nutritional status and weight gain during
pregnancy. Nutrition During Pregnancy. National Academy of Sciences: Washington,
DC.

Jackson M, Gott P, Lye S, Ritchie J, and Clapp J (1995) The effects of maternal aerobic
exercise on human placental development: placental volumetric composition and surface

areas. Placenta 16: 179-91.

Janney C, Zhang D, and Sowers M (1997) Lactation and weight retention. Am J Clin
Nutr 66: 1116-24.

Johnston E (1991) Weight changes during pregnancy and the postpartum period.
Progress in Food and Nutrition Science 15: 117-57.

Kardel K, and Kase T (1998) Training in pregnant women: effects on fetal development
and birth. Am J Obstet Gynecol 178: 280-6.

76

Keppel K, and Taffel S (1993) Pregnancy-related weight gain and retention: implications
of the 1990 Institute of Medicine guidelines. Am J Public Health 83: 1100-1103.

Klebanoff M, Shiono P, and Carey J (1990) The effect of physical activity during
pregnancy on preterm delivery and birth weight. Am J Obstet Gynecol 163: 1450-6.

Knuttgen H, and Emerson K (1974) Physiological response to pregnancy at rest and
during exercise. Journal of Applied Physiology 36: 549-53.

Larsen-Myer D (2002) Effect of postpartum exercise on mothers and their offspring: a
review of the literature. Obesity Research 10: 841-53.

Lederman S (2001) Pregnancy weight gain and postpartum loss: avoiding obesity while
optimizing the growth and development of the fetus. JAMWA 5 6: 53-8.

Linne Y, Barkeling B, Rossner S (2002) Long-tenn weight development after pregnancy.
Obesity Reviews 3: 75-83.

Linne Y, and Rossner S (2003) Interrelationships between weight development and
weight retention in subsequent pregnancies: the SPAWN study. Acta Obstet Gynecol
Scand 82: 318-25.

Linne Y, Dye L, Barkeling B, Rossner S (2004) Long-term weight development in
women: a 15-year follow-up of the effects of pregnancy. Obesity Research 12: 1166-78.

Lokey E, Tran Z, Wells C, Myers B, Tran A (1991) Effects of physical activity on
pregnancy outcomes: a meta-analytic review. MSSE 23: 1234-9.

Lotgering F, Gilbert R, and Longo L (1983) Exercise responses in pregnant sheep: blood
gases, temperatures, and fetal cardiovascular system. Journal of Applied Physiology 55 :
842-50.

Lotgering F, Gilbert R, and Longo L (1985) Maternal and fetal responses to exercise
during pregnancy. Physiology Review 65: 1-36.

McAuley E, and Mihalko S (1998) Exercise-related Self-efﬁcacy. In Duda JL (Ed),
Advances in Sport and Exercise Psychology Measurement, (371-90). Fitness Information
Technology Inc: Morgantown, WV.

McKeown T, and Record R(1957) The inﬂuence of reproduction on body weight in
women. Endocrin 15: 393-409.

Miller Y, Trost S, and Brown W (2002) Mediators of physical activity behavior change
among women with young children. Am J Preventative Medicine 23: 98-103.

77

Muscati S, Gray-Donald K, and Koski K (1996) Timing of weight gain during
pregnancy: promoting fetal growth and minimizing maternal weight retention. Int J of
Obesity 20: 526-32.

Ohlin A, and Rossner S (1990) Epidemiology of gestational weight gain and body weight
changes after pregnancy. Epidemiologic Reviews 21: 261-75.

Ohlin A, and Rossner S (1994) Trends in eating patterns, physical activity and
sociodemographic factors in relation to postpartum body weight development. British
Journal of Nutrition 71 : 457-70.

Parker J, and Abrams B (1993) Differences in postpartum weight retention between black
and white mothers. Obstest Gynecol 81: 768-74.

Pate R, Pratt M, Blair S, Haskell W, Macera C, Bouchard C, Buchner D, Ettinger W,
Heath G, King A, et al (1995) Physical activity and public health. A recommendation

from the Centers of Disease Control and Prevention and the American College of Sports
Medicine. JAMA 273 : 402-7.

Pivarnik J (1998) Potential effects of maternal physical activity on birth weight: brief
review. MSSE 30: 400-6.

Pomerance J, Gluck L, and Lynch V (1974) Physical ﬁtness in pregnancy: it’s effect on
pregnancy outcome. Am J of Obstet Gynecol 119: 867-76.

Prenatal Care. Federal Security Agency and Social Security Administration (1949)
Children’s Bureau Publication No. 4, 25-26.

Randell M (1945) Training for childbirth, 3rd edition. Churchill: London.

Rabkin C, Anderson H, Bland J, Brooke O, Chamberlain G, and Peacock J (1990)
Maternal activity and birth weight: a prospective population-based study. Am J of
Epidemiology 131: 522-31.

Rookus M, Rokebrand P, Burema J, and Deurenberg P (1987) The effects of pregnancy
on the body mass index 9 months postpartum in 49 women. IntJ of Obesity 11: 609-18.

Rooney B, and Schauberger C (2002) Excess pregnancy weight gain and long-term
obesity: one decade later. Obstet Gynecol 100: 245-52.

Rossner S, and Ohlin A (1995) Pregnancy as a risk factor for obesity: lessons from the
Stockholm Pregnancy and Weight Development Study. Obesity Research 1995: 2675-
755.

Sampselle C, Seng J, Yeo S, Killion C, and Oakley D (1999) Physical activity and
postpartum well-being. JOGNN 28: 41-49.

78

Schauberger C, Rooney B, and Brirner L (1992) Factors that inﬂuence weight loss in the
puerperium. Obstet Gynecol 79: 424-9.

Schieve L, Cogswell M, and Scanlon K (1998) An empiric evaluation of the Institute of
Medicine’s pregnancy weight gain guidelines by race. Obstet Gynecol 91: 878-84.

Stemfeld B, Quesenberry C, Eskenazi B, and Newman L (1995) Exercise during
pregnancy and pregnancy outcome. MSSE 27 : 634-40.

Stemfeld B (1997) Physical activity and pregnancy outcome: review and
recommendations. Sports Medicine 23 : 33-47.

Walker L, and Freeland-Graves J (1997) Lifestyle factors related to postpartum weight
gain and body image in bottle- and breastfeeding women. JOGNN 27 : 151-60.

79

 

Hm.

CHAPTER 5

The effects of past pregnancy physical activity participation on current physical
activity, barriers to physical activity, and current body size

Introduction

Background

In 1985, the American College of Obstetricians and Gynecologists published its
ﬁrst recommendations for exercise during pregnancy (ACOG, 1985). Based on limited
research at that time, the recommendations were conservative. Major concerns included
maternal temperature regulation issues, as well as exercise increasing the risk of preterm
birth and low birth weight. Recommendations indicated that maternal heart rates should
not exceed 140 beats per minute and vigorous exercise was to be limited to no more than
15 minutes per bout.

In 1994, the ACOG recommendations were revised to include and address
research ﬁndings that had been performed in the nine years following the original
guidelines (Clapp, 2000; Crowell, 1995; Johnston, 1991; Larson-Myer, 2002). Based on
more recent data, many studies indicated that maintaining a physically active lifestyle
during pregnancy could actually enhance the mother’s health and improve the delivery
process (Clapp, 2000; Crowell, 1995). As a result, new recommendations no longer
included heart rate or duration restrictions for healthy mothers. Continuing aerobic
exercise throughout pregnancy was encouraged for mothers who were active before
pregnancy, and individual health and exercise screenings and prescriptions were
suggested for all pregnant women (ACOG, 1994). However, the effects of sedentary
women starting an exercise program during pregnancy and possible detrimental effects of

resistance training throughout gestation had not yet been examined thoroughly.

80

Postpartum women were advised to undertake a slow progression of activity back to pre-
pregnancy levels after the 6-week postpartum check up and medical permission.
However, few details were provided.

In 2002, ACOG released its most recent list of recommendations for exercise
during pregnancy. These follow the current CDC/ ACSM guidelines for the general
population, suggesting that in the absence of medical or obstetric complications, 30
minutes or more of moderate exercise a day on most, if not all, days of the week is
recommended for pregnant women (ACOG, 2002). The beneﬁts of exercise for obese
pregnant women and the possible prevention of gestational diabetes are discussed as well
beneﬁts of exercise for previously inactive women, or those with medical complications.
The guidelines suggest elite and recreational athletes remain active, modifying exercise
volume and intensity as medically indicated (ACOG, 2002). Even with additional
research focusing on the postpartum period, the recommendations for exercise soon after
delivery still suggested the gradual resumption of activities. Although few details are
provided, the newest guidelines suggest that postpartum PA has been associated with
decreased incidence of postpartum depression (ACOG, 2002).

Postpartum body weight changes and physical activity

Postpartum weight retention (PPWR) is the pregnancy related weight retained
after delivery. Some research shows the risks of weight gain, weight retention, and
subsequent obesity are increased during the childbearing years due to pregnancy-related
weight gain (Greene et al., 1988; Harris et al., 1997; Rooney and Schauberger, 2002).
Although most studies show that weight retention is common during the postpartum

period, there is no consensus on the amount of weight typically retained. According to a

81

review by Linne et al. (2002), mean weight increases after pregnancy, but individual
variations are wide. The Stockholm Pregnancy and Women’s Nutrition (SPAWN)
studies report mean weight increases induced by pregnancy to be 0.5 kg in addition to
age-related weight gain, but with a wide individual range (Linne et al., 2002). In 1988,
Greene et a1 attempted to study the effect of pregnancy on body weight by assessing
interpregnancy weight changes. Average weight gain between pregnancies was 0.9 kg,
while one in ten women retained excessive weight of 7 kg or more associated with
pregnancy (Greene et al., 1988).

The Institute of Medicine (IOM) has published guidelines for weight gain during
pregnancy based on pregravid body size. Speciﬁcally, underweight women should gain
between 12.5-18 kg, normal weight women should gain between 11.5-16 kg, overweight
women should gain 7-11.5 kg, and obese women should gain approximately 6 kg (IOM,
1990). Several studies have been conducted to evaluate the effects of women of different
ages, races, and pro-pregnancy weight following these guidelines on PPWR and infant
birth weight (Keppel and Taffel, 1993; Schieve et al., 1998). Results have shown that
IOM recommendations preserve healthy birth weight and limit PPWR in Caucasian
women of normal pre-pregnancy weight, but may not positively affect BW for African
American women (Keppel and Taffel, 1993; Schieve et al., 1998).

Many variables have been examined as potential predictors of PPWR. They
include pre-pregnancy weight or body mass index (BMI), pregnancy-related weight gain,
smoking, lactation, socioeconomic status (SES), marital status, parity, age, and PA (Ohlin
and Rossner, 1989; Harris et al, 1997; Johnston, 1991). The relationship between these

variables and PPWR has yet to be clearly established. Some studies have shown both

82

 

positive and negative effects of these variables on PPWR, and the fact that they are
intercorrelated makes study design and data collection method extremely important.
Differences in outcome variables have often made it difﬁcult to compare the results of
one study to another. In almost all research, the strongest predictor of postpartum weight
retention is gestational weight gain (Linne et al., 2004). However, it should be noted
most studies have included short durations or follow-up periods and have been
retrospective in design.

Results from the few research studies on the effects of PA on PPWR are varied,
possibly due to differences in study design and PA measurement. Studies have followed
women through 6 to 12 weeks postpartum (Muscati et al., 1996; Sampselle et al., 1999),
with some continuing for 6-9 months (Boardley and Sargent, 1995; Brewer et al., 1989;
Haiek et al., 2001; Keppel and Taffel, 1993; Rookus et al., 1987; Soltani and Fraser,
2000). A few studies have involved long term follow-up with information collected on
lifestyle changes due to pregnancy and motherhood. In 2002, Rooney and Schauberger
conducted an 8-10 year follow-up study to estimate the impact of excess pregnancy
weight gain and failure to lose that weight on excess weight later on in life. They found
that excessive weight gain and weight retention were important predictors of long-term
weight changes and high BMI later in life (Rooney and Schauberger, 2002).

The arrival of a child can cause maternal lifestyle changes that may affect
physical activity behaviors. In a prospective study of body weight changes in women,
postpartum weight retention was more affected by lifestyle changes during and after
pregnancy than by antenatal factors (Ohlin and Rossner, 1996). In 1992, Brown et al.

assessed lifetime parity and weight of Iowa women who were part of a longitudinal

83

cohort study examining body mass and cancer risk Data collection included a mailed
questionnaire requiring recall of pregnancy history and weights at ages 18, 30, 40, and 50
years (Brown et al., 1992). BM] for each age was calculated and parity was deﬁned as
the lifetime number of self-reported live births. The researchers found body weight and
BMI increased with age and parity was associated with an increase in body weight of
0.55kg per live birth (Brown et al., 1992). At every age, women with a lifetime parity of
1-2 live births had lower mean body weights and BMI values than women with no live
births or more than three. The authors stated that weight gain with age most likely results
from lifestyle factors that institute a positive caloric balance during the adult years
(Brown et a1, 1992). They also stated that the results of this study reinforced the need to
identify modiﬁable factors related to weight gain with age in women, so that appropriate
interventions could be implemented.

The Australian Longitudinal Study of Women’s Health indicated that 18—22 year
Old women with children are likely to be sedentary or participate in less than one physical
activity bout per week (Brown et al., 2000). Less than half of this group of women
participated in adequate PA compared with more than half of women without children
(Brown et al., 2000). In a study by Miller et al. (2002), three groups of women, with pre-
school aged children, were separated into different intervention groups to assess
mediators of PA behavior change. A signiﬁcant intervention effect was found indicating
partner support and self-efﬁcacy appeared to be mediators of PA behavior change (Miller
et al. 2002). Lifestyle changes, especially the increase in sedentary behavior caused by
the arrival of a new child can have a negative effect on a woman’s PA level that in turn

could affect weight retention, mental well-being, and overall health (Miller et al., 2002).

84

Self-efficacy and barriers to physical activity

Self-efﬁcacy (SE) is a person’s belief in his/her ability to succeed in tasks that
occur in his/her life to obtain a speciﬁc outcome (Bandura, 1977). A woman’s SE toward
PA can affect her PA level. The strength of an individual’s personal efﬁcacy will
determine what tasks are attempted, how much effort will be expended, and how long a
person will persist in the face of adversity (McAuley and Mihalko, 1998).

Barrier efﬁcacy involves a person’s belief in his/her ability to overcome barriers
that inhibit participation in certain aspects of his/her life (McAuley, 1992). Investigators
who have focused on barriers have demonstrated this form of efficacy to be consistent
with the prediction of exercise behavior (McAuley and Mihalko, 1998). Barriers to tasks
or participation in PA can be measured by listing a reasonable number of barriers and
assessing one’s conﬁdence in overcoming a barrier by using a 0-100% conﬁdence scale
(McAuley and Mihalko, 1998).

The establishment of certain barriers as important factors inhibiting PA behaviors
in women with children could assist with the creation of intervention programs targeting
increases in PA along with the prevention of PPWR. Beilock et a1 (2001) surveyed past
athletes about the effects of pregnancy on training and return to competition. A list of
common barriers to exercise was compiled and then participants indicated their degree of
conﬁdence from 0-100 in their ability to continue participation in training at least three
times a week (Beilock et a1, 2001). Participants indicated overwhelmingly that childcare
was the biggest barrier to training in the postpartum period. Determining common
barriers to exercise could help with the creation of new interventions for PA participation

that target mothers with children.

85

Significance of the problem

According to a review by Mottola (2002), potential beneﬁts of exercise during the
postpartum period may include improved cardiovascular ﬁtness, facilitated weight loss,
increased energy, improved psychosocial well-being, decreased urinary stress
incontinence, and decreased lactation-induced bone loss. While some research has
shown the tendency for the childbearing years to be periods of decreased PA, few studies
have examined the role of PA during pregnancy and the postpartum period to explain the
variance in individual PPWR (Linne et a1, 2002).

Due to lifestyle changes that occur because of pregnancy, factors related to the
arrival of a new born may affect a woman’s ability to exercise during the postpartum
period, therefore affecting PPWR. Determining these factors or barriers can assist the
ﬁtness and PA community in increasing accessibility and motivation to become or remain
physically active after pregnancy.

Few prospective longitudinal studies have been undertaken to examine PPWR and
the lifestyle factors that change with motherhood (Linne et al, 2004; Rooney and
Schauberger, 2002). In two noteworthy long-term studies, possible effects of PA during
pregnancy on weight retention were not directly addressed though both noted that
exercise participation after pregnancy may improve long-term weight control (Linne, et
al., 2004; Rooney and Schauberger, 2002). There is a need for longitudinal prospective
studies that address the long-term effects of PA during pregnancy and the postpartum
period on weight as well as the barriers to PA that change lifestyle behaviors after
pregnancy.

Purpose of the study

86

Due to gaps in current knowledge there is a need for a longitudinal study to assess
long-term effects of PA participation on PPWR, and lifestyle changes on PA
participation. This dissertation will follow a cohort of women (previously studied during
pregnancy) to assess lifestyle factors, such as PA and barriers to exercise on current
anthropometrics. The purpose of this study is to examine PA participation during a past
pregnancy, current weight, anthropometrics, current PA levels, and barriers to exercise of

women who are six years postpartum. Based on this purpose, the research questions are

as follows.
1. Is current physical activity related to current body size?
2. Is physical activity performed during a pregnancy that occurred Six
years ago related to current physical activity?
3. What are the current ban‘iers to exercise and what is the average
percent conﬁdence score?
4. Is a woman’s percent conﬁdence to overcome current barriers to

exercise related to past pregnancy physical activity participation?

Limitations of the study
The study participants were obtained from a homogeneous convenience sample

with respect to age, race, and socioeconomic status. Generalizability of study results to
more diverse groups of previously pregnant women Should be applied with caution.
Breast-feeding data in the original study were collected at 12 weeks postpartum. The
women will be asked as to recall duration of breastfeeding for that pregnancy, and their
responses may be susceptible to recall bias. Recalled pre-pregnancy weight for the

pregnancy six years ago will be a surrogate for pre-pregnancy weight, which was not

87

available. Finally, some study participants may have had additional pregnancies since the
original study.
Methods
Participants

56 women participated in a study designed to determine the reliability and validity
of three commonly used physical activity measurement modalities from 1997-2000
(Pivarnik et al., 2002; Stein et al., 2003). Study participants were recruited from
obstetrical care clinics in the Mid-Michigan area. All participants delivered healthy,
singleton infants. The original MAMA study involved the collection of heart rate
telemetry, accelerometry, and a PAR over two consecutive days at 20 and 32 weeks
gestation, and 12 weeks postpartum. The researcher reviewed the two-day PAR data
with participants during each visit to the lab. When the women were recruited they were
informed that the investigators may wish to obtain some follow-up measures (on them
and/or their offspring) at a later time. Thus, participants were then contacted to
participate in the current study. Email, telephone calls, and letters were used to contact
the 41 of the original 56 participants believed to still live in the area. After many phone
calls, thirty women were located and agreed to participate, with only one reﬁrsal.
Instrumentation

A stadiometer was used to measure participant height in centimeters. A beam
balance scale was used to measure participant weight in kilograms. Bioelectrical
Impedance Analysis (BIA) was used to measure fat free mass (FFM). BIA involves the

passing of a small current through the body (using either hand/foot electrodes or a scale)

88

measuring the resulting tissue resistance and reactance values (Lohman, 1992). Percent
body fat is calculated using FFM and measured weight.

The Modiﬁable Activity Questionnaire (MAQ) was used to recall PA from a past
pregnancy. The MAQ is a PAR questionnaire designed for easy modiﬁcation to
maximize the ability to assess PA levels in different populations (Kriska et al, 1990:
Kriska, 1997). This questionnaire has been used to assess historical as well as past week
PA levels. The MAQ collects information regarding mode, intensity, duration, and
frequency of PA as well as occupational activity values. The compendium of physical

activities can then be used to calculate MET hours and Kcal expenditure for different

 

activities and time periods (Ainsworth et al., 2000). The MAQ was created initially as an
interviewer administered tool and was validated in two studies assessing PA levels and
their effect on diabetes risk in Pima Indians. The MAQ PA data were compared with
Caltrac activity counts using Spearman rank-order correlations, r=0.62. Test-retest
correlations ranged from r=0.88-0.92. The MAQ has also been used to assess the effects
of PA on female adult bone parameters, and glucose tolerance of Mauritius citizens. A
copy of the questionnaire to be used with this study can be found in appendix B.

A Perceived Barriers Efﬁcacy Questionnaire (PBEQ) was used to assess
current barriers to PA (Beilock et al., 2001). The PBEQ requires participants to list their
own perceived barriers to current exercise. Each participant is then asked to assign a
percent conﬁdence in overcoming each barrier. Percent conﬁdence is a percentile that
can range from 0-100, 0 meaning no conﬁdence and 100 meaning total conﬁdence in
overcoming a chosen barrier. After a percent conﬁdence score is given to each listed

barrier, the top three barriers are ranked from ﬁrst, meaning most, to third, meaning least

89

important ban'ier. A percent conﬁdence score is then calculated for each participant by
averaging the scores of the tOp three ranked barriers. This questionnaire was pilot tested
in another group (N=6) of postpartum women and is based on a similar tool used to
assess barriers to exercise in athletic women returning to sport after delivery (Beilock et
al., 2001). A copy of this questionnaire can be found in appendix C.

. Sample size, power analysis

Due to the follow-up design of this study, we could not increase sample size to
enhance statistical power. Therefore, correlation values needed to achieve meaningful
relationships and signiﬁcance in line with previous reports were calculated using the
available sample of 30 women. A review of other PAR tools and MAQ studies assisted
in the determination of correlation signiﬁcance.

Agreement between measured PA and self-reported PA via PAR tools for the
same time periods have been established in studies comparing direct measurement
techniques such as Vitalog and Caltrac accelerometers, to exercise logs and recall (Stein
et al., 2003; Taylor et al., 1984). Correlations have ranged from 0.28 to 0.90 in studies
assessing the validity and reliability of PAR tools (Kriska, 1997; Norman et al., 2001;
Vuillemin et al., 2000; Wareham et al., 2002). Based on our sample size of 30,
conversion tables indicated that an F036 would be signiﬁcant for an alpha level of
P<0.05, while an r=0.46 would signiﬁcant for an alpha level of P<0.01.

Data Collection Procedures

In the original MAMA study, actual PA data for all the three previously

mentioned time points were collected using a two-day exercise PAR. During those two

days, women wore watches with alarms that sounded on the hour. This alarm served as a

90

reminder that they should record any activity performed within the previous hour. In
addition, the study coordinator debriefed each participant to be sure that she recalled her
previous day’s activities as accurately as possible (Stein et al., 2003). During the follow-
up study participants came to the Human Energy Research Lab at Michigan State
University for one 45-60 minute follow-up visit. Upon arrival the participant was
reminded of the study aims and protocols, after which written informed consent forms
were completed.

Anthropometrics

Height was measured to the nearest in 0.1 cm by using a stadiometer. Body
weight was measured to the nearest 0.1 kg on a beam-balance. Both instruments were
calibrated with known standards before use. FFM was estimated by BIA, and % fat was
calculated from FFM and total body weight measurements. A single trained investigator
performed these measures. Appendix 1 shows the anthropometric equations used in this
study.

Surveys

The ﬁrst questionnaire collected demographic information including age, parity,
time back to work ﬁ'om pregnancy, smoking, marital status, and race. The MAQ
assessed current PA and the PBQ assessed current barriers to PA. The previous sample
was very homogeneous( predominantly Caucasian, college educated, and married),
therefore, SES and education were not addressed speciﬁcally.

The MAQ was used to assess PA, at 20 weeks and 32 weeks gestation and 12
weeks postpartum during a pregnancy six years ago, and last week. The details for

administering the MAQ were discussed previously in chapter 3 of this dissertation.

91

Overall, for each time period of interest, participants were asked to circle all activities
(from a pre-determined list generated from pilot studies) in which they had participated.
Refer to appendix B for examples of this activity list for the MAQ. The frequencies, in
days per week, and duration, in hours per bout, were established for each chosen activity.
The MET value for all listed activities was determined using the compendium for
physical activities developed by Ainsworth et al. (2000). Estimated energy expenditure
for each activity was calculated and multiplied by the hours per week in order to calculate
a weekly value, for example Kcal/week, (MET*WT)(dur/60)*freq=Kcal/wk (Kriska,

1997). The estimated energy costs of each activity were derived from the most recent

 

compendium of physical activities (Ainsworth et al., 2000). After the average weekly
energy expenditure values were calculated, those data were divided by seven to derive
daily energy expenditure, then divided by the maternal body weights for each study
period in order to obtain energy expenditure relative to weight
([Kcal/day]/kg=Kcal/day/kg). Please refer to appendix A for the energy expenditure
equations used with the MAQ described in this section.

Current barriers to PA were then assessed via the PBEQ. The participants were
asked to list their perceived barriers to PA, rank order the top 3 barriers and indicate
percent conﬁdence in overcoming the listed barriers to physical activity on a scale of 0-
100%. An overall conﬁdence score was then calculated by averaging the scores of the
top 3 barriers.

Data Analysis
Descriptive characteristics and selected physiological measures were calculated

and presented in tabular form. Measures included participants’ current age, parity,

92

weight, height, and body composition. PPWR was calculated by subtracting previously
determined pre-pregnancy weights from current measured weights. In the original
MAMA study breast-feeding data were collected at 12 weeks postpartum only. In the
current study we queried the participants about how many weeks they breast fed
exclusively. Based on the updated breastfeeding data, the median value of both exclusive
and total breastfeeding were noted. Women were then grouped by whether they breastfed
less or equal to or greater than the median for exclusive breastfeeding (24 weeks).

Statistical signiﬁcance for all analytical procedures was set at an alpha level of
P=S0.05. The statistical program of JMP version 4 was used for all statistical procedures
(Sall, Lehman, and Creighton, 2001). During the data analysis process it was discovered
that there were two participants whose PA at all time periods was 2-3 standard deviations
above the sample norm. Closer inspection of the data showed the women, though more
active than the rest, had values within the regression line and therefore, remained part of
the analyses. Some analyses also involved comparisons based on current PA levels.
Therefore the participants were separated into two groups based on those who met the
current CDC/ACSM PA recommendations of 20 MET/wk or more, and those who did
not.
Research Question One: Is current physical activity related to current body size?

Data analysis included Pearson correlations between current leisure time PA
measures (kcal/day/kg) and the body size measures of BMI and % fat.
Research Question Two: Is physical activity that was performed during a pregnancy six

years ago related to current physical activity?

93

Data analysis included Pearson correlations between the current leisure time PA
measure (kcal/day/kg) and the previously recorded data from the three past time periods
of interest (20 and 32 weeks gestation and 12 weeks postpartum).

Research Question Three: What are the current barriers to exercise and what is the
average percent confidence score?

Data analysis included the frequencies of all barriers listed in tabular form with
the top three barriers to exercise for all participants noted as such. The average percent
conﬁdence score for all study participants was also calculated along with standard
deviations.

Research Question F our: Is a woman’s percent conﬁdence in overcoming current
barriers to exercise related to past leisure time physical activity levels during a
pregnancy six years ago?

Data analysis included Pearson correlations between the percent conﬁdence
scores and energy expenditure data (kcal/day/kg) from all time periods of interest (20 and
32 weeks gestation, 12 weeks postpartum). Both previously measured (via PAR) and
historically recalled (via MAQ) PA during pregnancy were compared with percent
conﬁdence scores.

Results

Descriptive characteristics and other select physiological measures of the entire
sample can be found in Table 1. The average BMI for the sample is considered normal
and the % fat is between the 10th ad 20th percentile according to the ACSM Guidelines for
Testing and Prescription. Anthropometric and descriptive data were grouped by current

PA levels (220 MET/wk, or not) and can also be found in Table 1. Nine women had no

94

further children after the one in the MAMA study, while one woman had three. The
median for exclusive breastfeeding was 24 weeks with a range from 0-104 weeks. The
median value for total breastfeeding with the entire sample was 52 weeks with a range
from 3-182 weeks. As previously described in the data analysis section, the women were
grouped by exclusive breastfeeding duration. One-way AN OVAS were run comparing
the means between the breast feeding groups for current weight (1.07, P=0.30), PPWR
(1.10, P=0.29), and percent conﬁdence score (2.47, P=0. 12). There were no signiﬁcant
differences found.

When the anthropometric data were separated into current PA level groups

 

(women meeting or exceeding 20 MET/wk of activity), the more active women were the
same age, height, and had the same number of children as the less active women. One-
way AN OVAS were run and found mean differences based on current PA. At a
signiﬁcance of P5005, the women who met or exceeded the CDC/ACSM
recommendations for PA showed lower values for weight, BMI, %fat, and PPWR than
those who did not. The median value for PPWR was 2.8 kg with the highest value being
14.9 kg and the lowest of -8.4kg. Ten of the thirty participants were at or below their
pre-pregnancy weight of six years ago. Correlations were run to determine the
relationships between PPWR and age (r=0.02), number of children (r=0.02), and energy
expenditures [20 (r=0.02) and 32 weeks gestation (F004), 12 weeks postpartum (Ir—0.1),
and current (r=0.16)]. No signiﬁcant relationships were found. The rest of the results for
Part II of this dissertation are organized by research question.

Research question one stated: Is current physical activity related to current boay

size? Results of Pearson correlations showed that current leisure time energy expenditure

95

(kcal/day/kg) were inversely related to BMI (r=-0.55; P<0.01) and % body fat (r=-0.70;
P<0.01) for the entire sample. Further analysis also showed past pregnancy PA, both
measured and recalled, inversely related to BMI and % body fat. Figures 4-15 contain
the correlation data for past pregnancy PA and body size.

Research question two stated: Is physical activity that was performed during a
pregnancy six years ago related to current physical activity? Results of Pearson
correlations showed that current leisure time energy expenditure (kcal/day/kg) for all
subjects was related to measured leisure time energy expenditure at 20 weeks gestation
(r=0.49; P< 0.01), 32 weeks gestation (r=0.71; P<0.01), and 12 weeks postpartum
(r=0.73; P<0.01). Pearson correlations were also run between the recalled (MAQ) past
pregnancy PA and current PA with the following outcomes: at 20 weeks gestation
(r=0.85; P<0.01), 32 weeks gestation (r=0.85; P<0.01), and 12 weeks postpartum (r=0.80;
P<0.01). Table 4 contains the correlation data for research question 2.

Research question three stated: What are the current barriers to exercise and
what is the average percent conﬁdence score? The results of all Of the ﬁ'equencies for all
listed barriers to exercise are displayed in Table 5. The top three barriers, with
subsequent frequencies are as follows: The number one barrier was time (22), second was
motivation (7), and third was childcare (6). The average percent conﬁdence score was
65.6 % ($24.5). Further analysis showed that percent conﬁdence score was positively
correlated with current PA (r=0.53; P<0.01).

Research question four stated: Is a woman’s percent conﬁdence to overcome
current barriers to exercise related to past physical activity levels during a pregnancy six

years ago? Results of Pearson correlations on the entire sample showed the percent

96

conﬁdence scores were related to measured (PAR) leisure time energy expenditure at
only 12weeks postpartum (r=0.42; P<0.05) when the original PAR leisure time energy
expenditure data were used. In contrast, signiﬁcant relationships were found at all study
times when the MAQ recalled leisure time energy expenditure values were used.
Speciﬁcally, signiﬁcant correlations were found at 20 weeks (r=0.51; P<0.01), 32 weeks
gestation (r= 0.49; P<0.01) and 12 weeks postpartum (r=0.52; P<0.01). Table 6 and
Figures 16-18 consist of research question 4 correlation results.

Discussion

The purpose of this study was to examine PA participation during a past
pregnancy, current weight, anthropometrics, current PA levels, and barriers to exercise of
women who were six years postpartum. Based on this purpose, research questions
included: was current physical activity related to current body size; was physical activity
performed during a pregnancy that occurred six years ago related to current physical
activity; what were the current barriers to exercise and what was the average percent
conﬁdence score; and was a woman’s percent conﬁdence to overcome current barriers to
exercise related to past pregnancy physical activity participation.

The overall descriptive results of this study showed a mean PPWR of 2.8 kg ($5.2
kg), while values ranged from -8.4 to 14.9 kg. The mean PPWR in this sample is higher
than that found by other studies (0.9kg, Greene et al., 1988; and 1.51 kg, Ohlin and
Rossner, 1990). The range of PPWR is wide, but is similar to the range found by a
similar study (-12.3 to 26.5 kg, Ohlin and Rossner, 1990). Differences in pre-pregnancy
weight measurement are common in PPWR studies. Whether medical records or subject

recall are used, the PPWR value will differ. Also, differences in duration of follow-up,

97

number of pregnancies between weight measurements, and cultural differences between
sample populations may cause variations in PPWR.

The results of the MAMA follow-up showed a signiﬁcant inverse relationship
between current PA levels and current body size values (BMI and % fat). This infers that
current activity impacts postpartum body weight and fatness. Though no previous study
examined this relationship speciﬁcally, Ohlin and Rossner (1994) indicated that physical
activity was related to postpartum body weight. Their results showed that women who
retained 25 kg at 1 year postpartum were seldom physically active (as deﬁned by
performing ~ 4-6 hours of light activity/week) in their leisure time (Ohlin and Rossner,
1994)

A statistically signiﬁcant relationship was also found between past pregnancy PA
and current PA, all time periods of interest (20 and 32 weeks gestation and 12 weeks
postpartum) were signiﬁcant at PS 0.05. Refer to Table 4 for these data.

Based on these analyses, past pregnancy PA is correlated with current PA, which
is inversely related to current body size. When correlations were run between past PA (at
all time periods of interest) and current body size, Signiﬁcance was found. This indicates
that past pregnancy PA was inversely correlated with current body size variables. In
order to establish how this relationship may work, further analyses were conducted to
determine if current PA is a confounder, mediator, or modiﬁer with in the past pregnancy
PA and current body size relationship. When current PA was added into the model, the
relationship between past pregnancy PA and current body size was erased, indicating that
current PA is the proximate cause of current body size. Research has shown that past

performance or PA participation is related to current PA participation, which the results

98

of this study also support. So, current PA is the main predictor of current body size, past
pregnancy PA behavior may have a direct relationship to current PA. Refer to Figure 19
for a visual representation of these relationships.

The current barriers to PA found in this study are displayed in Table 5. Those
listed in this study are similar to those found by Beilock et al. (2001). The top barrier
response for both studies was lack of time, and showed much higher frequencies that any
other barriers listed. Other barriers such as lack of energy and motivation were also
mentioned in both studies. The Beilock et al. participants were all former athletes who
were asked to list barriers to training for competition during the postpartum period, while
we queried our participants about current barriers to physical activity in general, not
competition. Though a few participants in the present study may compete from time to
time in road races, a majority of the sample was not competitive. Also, unlike the
longitudinal design of this study, the Beilock et al. study was retrospective, asking for the
recollection of barriers to training postpartum within 10 years past (Beilock et al., 2001).
Unlike the current study, the Beilock et a1. study did not assess current percent
conﬁdence in overcoming barriers to PA; therefore direct comparison with the present
study is not possible. Beilock et al. found that postpartum PA was more predictive of
current PA. The present study also found a signiﬁcant correlation between measured past
pregnancy PA at 12 weeks postpartum and percent conﬁdence, possibly supporting the
theory that postpartum activity and is more predictive of current activity and self-
regulatory measures that activity during pregnancy.

Based on the idea of self-regulatory beliefs effecting PA behavior, we

investigated current percent conﬁdence score as an effect modiﬁer in the past pregnancy

99

PA and current PA relationship. A stepwise analysis of past pregnancy and current PA,
adding percent conﬁdence as an interaction term, showed a signiﬁcant (PS 0.01) increase
in regression coefﬁcient when the percent conﬁdence was added as an interaction term.
Speciﬁcally r2 values increased from 0.24 to 0.64 (at 20 weeks gestation), 0.48 to 0.64 (at
32 weeks gestation), and 0.53 to 0.64 (at 12 weeks postpartum). This infers that high
current conﬁdence to overcome ban’iers to PA enhances the effect of past exercise on
current activity behavior. The average percent conﬁdence score in this study was 65.5 %,
which infers a high conﬁdence in overcoming current barriers to PA with this particular
study population. This is supported by a 60% current PA participation rate, showing high
PA participation and high percent conﬁdence in overcoming current barrier to PA
participation.

The connection between percent conﬁdence score and past pregnancy PA was
analyzed using original and recalled past pregnancy PA values. The only signiﬁcant
relationship between percent conﬁdence scores and the original PA values occurred at 12
weeks postpartum, although the value at 32 weeks gestation approached signiﬁcance
(r=0.30). It is interesting that recalled PA was Signiﬁcantly correlated with percent
conﬁdence at all time points (20 and 32 weeks gestation and 12 weeks postpartum). This
may be due to that fact that both measurements, recalled PA and percent conﬁdence,
were personal perceptions. The percent conﬁdence score was a perception of how
conﬁdent the subject was in overcoming barriers to current PA and the recalled PA value
was their perception of how much activity they participated in during a past pregnancy.
Therefore, a woman’s perception (rather than reality) of past physical activity history is

more meaningful in determining her conﬁdence in overcoming barriers to being active.

100

We found no studies that compared both actual and recalled PA with current
barriers to physical activity. However, Godin, Vezina, and Leclerc (1989) investigated
the factors inﬂuencing intentions of pregnant women to exercise after giving birth. Their
results Showed that intention to exercise was strongly correlated with attitude towards the
act, exercise habit, and perceived barriers to exercise (Godin, Vezina, and Leclerc, 1989).

An individuals’ perception of ability to perform certain physical activities has
been investigated using male post myocardial infarction patients. Ewart et al. (1983).
discussed the relationships between pre-exercise testing self-efﬁcacy and treadmill
performance as well as the effect of post-test self-efﬁcacy on subsequent PA participation
in a normal environment (Ewart et al., 1983). Results showed that a patient’s perception
of his physical activity capacity and actual patterns of subsequent physical activity is
inﬂuenced by early treadmill testing performance. The authors believed self-efﬁcacy
judgment to be an important link between functional status and physical performance
(Ewart et al., 1983). Ewart et al. (1986) evaluated a similar population of men with
coronary artery disease, to examine the relationship between self-efﬁcacy and strength
gains during a circuit weight training program (Ewart et al., 1986). Results stated that
pre-training self-efﬁcacy judgments predicted post-test strength gains (Ewart et al.,
1986). These two papers support the notion that one’s self-efﬁcacy, or belief in ability to
perform a task, can affect performance. The present study did not examine this
relationship speciﬁcally, but linked perceived conﬁdence with recalled physical activity.
A common thread between the Ewart et al. papers and the present study is that an

individual’s self-perception and conﬁdence can affect his/her ability to perform a task

101

and/or overcome a barrier, whether it be physical activity participation, exercise testing,
or the recall of past pregnancy PA.
Limitations

Limitations of this longitudinal study include the small sample size and
homogeneous participant population. Many discrepancies found between the outcomes
from this and other studies may be due to the noted study limitations. Study designs and
measurement of PA and barriers also differed, lending to some of the discrepancies in the
ﬁndings.

Even with the acknowledged study limitations, results of this MAMA follow-up
showed, Similar to other prominent PPWR studies, most women will retain weight during
the postpartum period. Women who were active within the current recommended
CDC/ACSM PA guidelines retained less weight six years postpartum than those who
were not. Therefore, PA during a pregnancy and the postpartum period could help
decrease PPWR and long-term weight retention. This MAMA follow-up also showed a
possible relationship between past pregnancy physical activity and current body size.
The existence of this relationship could greatly add to validity of promoting PA during
pregnancy. Long-terrn weight management may be aided if women remain, or become,
physically active during pregnancy and the postpartum.

Future Directions

Based on the results of this and other long-term postpartum weight retention
studies, more research with large, heterogeneous samples are needed to further establish
the role of pregnancy related PA on PPWR. Investigators should attempt to gather

accurate objective pregnancy PA and energy expenditure values at study onset using

102

measurement tools that have already been examined and found to be reliable and valid, in
order to compare outcomes with other studies. If possible, duration of follow-up should
include, but not end at, one year postpartum. Five to ten years Should be the follow-up

goal in order to assess long-term effects of PA during pregnancy.

103

References

Ainsworth B, Haskell W, Whitt M, Irwin M, Swartz A, Strath S, et a1 (2000)
Compendium of physical activities: an update of activity codes and MET intensities.
MSSE supp]: S498-516.

American College of Obstetricians and Gynecologists (1985) Exercise during pregnancy
and the postpartum period. Technical Bulletin 58. Washington, DC: ACOG Press.

American College of Obstetricians and Gynecologists (1994) Exercise during pregnancy
and the postpartum period. Technical Bulletin 129. Washington, DC: ACOG Press.

American College of Obstetricians and Gynecologists (2002) Exercise during pregnancy
and the postpartum period. Technical Bulletin 26 7. Washington, DC: ACOG Press.

Bandura A (1977) Self-efﬁcacy: toward a unifying theory of behavioral change.
Psychological Review 84: 191-215.

Beilock S, Feltz D, Pivarnik J (2001) Training patterns of athletes during pregnancy and
postpartum. Res Q Exerc Sport 72: 1-8.

Boardley D, Sargent R, Coker A, Hussey J, and Sharpe P (1995) The relationship
between diet, activity, and other factors, and postpartum weight change by race. Obstet
Gynecol 86: 834-8.

Brewer M, Bates M, and Vannoy L (1989) Postpartum changes in maternal weight and
body fat depots in lactating vs. nonlactating women. Am J Clin Nutr 49: 259-65.

Brown J, Kaye S, and Folsom A (1992) Parity-related weight change in women. Am J of
Obesity 10: 627-9.

Brown W, Mishra G, Lee C, and Bauman A (2000) Leisure time physical activity in
Australian women: relationship with well-being and symptoms. Research Quarterly for
Exercise and Sport 71: 206-16.

Chasen-Taber L, Erickson J, McBride J, Nasca P, Chasen-Taber S, and Freedson P
(2002) Reproducibility of self-administered lifetime physical activity questionnaire
among female college alumnae. Am J of Epidemiology 155: 282-9.

Clapp J (2000) Exercise during pregnancy: a clinical update. Clinics in Sports Medicine,
19: 273-85.

Crowell D (1995) Weight change in the postpartum period: a review of the literature.
Journal of Nurse-Midwifery 40: 418-23.

104

Downs D, and Hausenblas H (2004) Women’s exercise beliefs and behaviors during their
pregnancy and postpartum. Journal of Midwifery and Women ’s Health 49: 138-44.

Ewart C, Taylor C, Reese L, and DeBusk R (1983) Effects of early post myocardial

infarction exercise testing on self-perception and subsequent physical activity. Am J of
Cardiology 51: 1076-80.

Ewart C, Stewart K, Gillian R, and Kelemen M (1986) Self-efﬁcacy mediates strength
gains during circuit weight training in men with coronary artery disease. MSSE 18: 531-
40.

Feltz D (1988) Self-conﬁdence and sports performance. Exercise and Sport Science
Reviews 16: 423-57.

Godin G, Vezina L, and LeClerc O (1989) Factors inﬂuencing intentions of pregnant
women to exercise after giving birth. Public Health Reports 104: 188-95.

Greene G, Smiciklas-Wright H, Scholl T, and Karp R (1988) Postpartum weight change:
how much of the weight gained in pregnancy will be lost after delivery? Obstet Gynecol
71: 701-7.

Haiek L, Kramer M, Ciampi A, and Tirado R (2001) Postpartum weight loss and infant
feeding. Journal of the American Board of Family Practice 14: 85-94.

Harris H, Ellison G, Holliday M, and Lucassen E (1997) The impact of pregnancy on the
long-term weight gain of primiparous women in England. Int J of Obesity 21 : 747-55.

Heyward V, and Stolarczyk L (1996) Applied Boay Composition Assessment.
Champaign, IL: Human Kinetics.

Institute of Medicine ( 1990) Subcommittee on nutritional status and weight gain during
pregnancy. Nutrition During Pregnancy. National Academy of Sciences: Washington,
DC.

Johnston E (1991) Weight changes during pregnancy and the postpartum period.
Progress in Food and Nutrition Science 15: 117-57.

Keppel K, and Taffel S (1993) Pregnancy-related weight gain and retention: implications
of the 1990 Institute of Medicine guidelines. American Journal of Public Health 83:
1100-1103. '

Kriska A, Knowler W, LaPorte R, Drash A, Wing R, Blair S, et a1 (1990) Development
of questionnaire to examine relationship of physical activity and diabetes in Pima
Indians. Diabetes Care 13: 401-11.

Kriska A (1997) Modiﬁable Activity Questionnaire. MSSE 29: $73-78.

105

.Larsen-Myer D (2002) Effect of postpartum exercise on mothers and their offspring: a
review of the literature. Obesity Research 10: 841-53.

Linne Y, Barkeling B, Rossner S (2002) Long-term weight development after pregnancy.
Obesity Reviews 3: 75-83.

Linne Y, Dye L, Barkeling B, Rossner S (2004) Long-term weight development in
women: a lS-year follow-up of the effects of pregnancy. Obesity Research 12: 1166-78.

Lohman T (1992) Advances in boay composition assessment. Current issues in exercise
science series. Monograph No. 3. Champaign, IL: Human Kinetics.

McAuley E, and Mihalko S (1998) Exercise-related Self-efficacy. In Duda, JL (Ed),
Advances in Sport and Exercise Psychology Measurement. Fitness Information
Technology Inc: Morgantown, WV.

McAuley E (1992) The role of efficacy cognitions in the prediction of exercise behavior
in middle-aged adults. Journal of Behavioral Medicine 16: 103-13.

Miller Y, Trost S, and Brown W (2002) Mediators of physical activity behavior change
among women with young children. American Journal of Preventative Medicine 23: 98-
103.

Mottola M (2002) Exercise in the postpartum period: practical applications. Current
Sports Medicine Reports 1: 362-8.

Muscati S, Gray-Donald K, and Koski K (1996) Timing of weight gain during
pregnancy: promoting fetal growth and minimizing maternal weight retention. Int J of
Obesity 20: 526-32.

Norman A, Bellocco R, Bergstrom A, and Wolk A (2001) Validity and reproducibility of
self-reported total physical activity-differences by relative weight. Int J of Obesity 25 :
682-88.

Ohlin A, and Rossner S (1989) Maternal body weight development after pregnancy. Int J
of Obesity 14: 159-173.

Ohlin A, and Rossner S (1990) Epidemiology of gestational weight gain and body weight
changes after pregnancy. Epidemiologic Reviews 21: 261-75.

Ohlin A, and Rossner S (1994) Trends in eating patterns, physical activity and

sociodemographic factors in relation to postpartum body weight development. British
Journal of Nutrition 71: 45 7-70.

106

Ohlin A, and Rossner S (1996) Factors related to body weight changes during and after
pregnancy: the Stockholm Pregnancy and Weight Development Study. Obesity Research
4: 271-6.

Pivarnik J, Stein A, Rivera J (2002) Effect of pregnancy on heart rate/oxygen
consumption calibration curves. MSSE 34: 750-5.

Rookus M, Rokebrand P, Burema J, and Deurenberg P (1987) The effects of pregnancy
on the body mass index 9 months postpartum in 49 women. IntJ of Obesity 11: 609-18.

Rooney B, and Schauberger C (2002) Excess pregnancy weight gain and long-term
obesity: one decade later. Obstet Gynecol 100: 245-52.

Sampselle C, Seng J, Yeo S, Killion C, and Oakley D (1999) Physical activity and
postpartum well-being. JOGNN 28: 41-49.

Schieve L, Cogswell M, and Scanlon K (1998) An empiric evaluation of the Institute of
Medicine’s pregnancy weight gain guidelines by race. Obstet Gynecol 91: 878-84.

Sall J, Lehman A, and Creighton L (2001) JMP Start Statistics: a guide to statistics and
data analysis using JMP and JMP IN software, 2nd edition, Canada: Duxury.

Soltani H, and Fraser R (2000) A longitudinal study of maternal anthropometric changes
in normal weight, overweight and obese women during pregnancy and postpartum.
British Journal of Nutrition 84: 95-101.

Stein A, Rivera J, Pivarnik J (2003) Measuring energy expenditure in habitually active
and sedentary pregnant women. MSSE 35 : 1441-6.

Taylor C, Coffey T, Berra K, Iaffaldano R, Casey K, and Haskell L (1984) Seven-day
activity and self-report compared to a direct measure of physical activity. Am J of
Epidemiology 120: 818-24.

Torgen M, Alfredsson L, Koster M, Wiktorin C, Smith K, and Kilbom A (1997)
Reproducibility of a questionnaire for assessment of present and past physical activities.
Int Arch Occup Environ Health 70: 107-18.

Vuillemin A, Oppert J, Guillemin F, Essermeant L, F ontvieille A, Galan P, et al (2000)

Self-administered questionnaire compared with interview to assess past-year physical
activity. MSSE 32: 1119-24.

Wareham N, Jakes R, Rennie K, Mitchell J, Hennings S, and Day N (2002) Validity and

repeatability of EPCI-Norfolk physical activity questionnaire. Int J of Epidemiology 31:
168-74.

107

Hes—e r US$326 9338 AZHmUV o». 88— 3320. 8a magma 3. 952: @398: mania. ANNo Zmﬂéwv

 

 

 

 

 

 

 

 

 

 

 

 

 

mm .1293 <5 33 8.3. . 2: ES gm: 32%
Se ease $-53 we 3

Hon: $an wok Hamb 0mm NM Nw.m 333 w H .m Pm
sumo 8.3 3.3 Cwbv 3.03 3.3 3.3 3.3 3.3
NNO zmdér wab 3&3 27m.» PM Nab.» hub N33,. and.»
B"; 3.3 3.3 CF3 3&3 6.3 3.3 3.: 3.3
Iwo zmdéw wmb 39w 4PM.» Nd Ngma Rab uqb» 9N4
HS. 3. 5 3.3 Qua: 3.33 6.3 3.3 3.3 0.3

 

.. Famoﬁam ﬁnances. $.58 m" w Mobm

108

 

Table 2: Average physical activity values (MiSD) at all time periods of interest (20 and
32 weeks gestation, 12 weeks postpartum, and current)

0 weeks E2 weeks 12 weeks urrent PA
ET/wk) MET/wk) ET/wk) MET/wk)

 

 

 

Total sample 24.9 17.3 21.6 35.6
“=30 (37.1) (22.8) (25.6) (35.1)
220 MET/wk 53.4 48.6 49.8 54.4

(45.6) (23.0) (26.2) (33.6)

 

_.20 MET/wk 5.9 6.0 7.5 7.4
(7.1) (6.1) (7.0) (7.6)

 

 

 

 

 

 

 

109

Table 3: One-way ANOVA of recall ability (MAQ-PAR) (MiSD) at the past pregnancy
time points of interest (20 and 32 weeks gestations and 12 weeks postpartum) by current
PA group (220 MET/wk)

 

 

 

 

Ekecall Ability lF-ratio lP-value

kcal/kg/day)

20 weeks 0.10 3.00 0.09
(2.9)

32 weeks 0.02 0.87 0.36
(1.5)

12 weeks —0.47 0.03 0.87
(2.0)

 

 

 

 

 

 

None of the values were signiﬁcant.

110

 

Table 4: Correlations between measured (PAR) and recalled (MAQ) past pregnancy PA
and current PA at all three time periods of interest (20 and 32 weeks gestation and 12
weeks postpartum)

 

 

 

 

 

urrent PA and PAR urrent PA and MAQ
kcal/kg/day) kcal/kg/day)

20 weeks 0.49* 0.85*
i(kcal/kg/day)

2 weeks 0.71* 0.85*
kcal/kg/day)

12 weeks 0.73* 0.80*
cal/kg/day)

 

 

 

 

 

*Denotes signiﬁcance at P-value of $0.05

111

Table 5: Listed current barriers to exercise in rank order.

Top Barriers Frequency

 

Time 22
Motivation

Childcare

Weather

Work/worktime

Scheduling

Tiredness

Wanting to spend time with children
Injuo'

Lack of sleep

Kid prep time
Laziness/excuses

Lack of alone time

Family demands/housework
Illness

Accessibility

Money

Boredom

Guilt

Apathy/lethargy

Energy

Hr—‘Hv—v—‘Hv—NNNNNNWW-h-b-PONQ

112

Table 6: Correlations between percent conﬁdence scores and MAQ (recalled) energy
expenditures at time periods of interest (20 and 32 weeks gestation and 12 weeks

 

 

 

 

 

postpartum).
/o conﬁdence score A) conﬁdence score
nd PAR (measured) nd MAQ (recalled)
20 wk PA 0.16 0.51*
Ichal/kg/day)
2 wk PA 0.30 0.49*
kcal/kg/day)
12 pp PA 0.42* 0.52*
kcal/kg/day)

 

 

 

*Indicates signiﬁcant values at p50.05

113

 

Figure 1: Correlations (r) between past measured (PAR) and recalled (MAQ) physical
activity at 20 weeks gestation

 

 

  
     

 

 

20
>.
U
E
532 15'
d)
:5
W52
1"“! 10.
«E
33’
8'6
©§ 50
<
&

r-O.57
o . . .
0 5 10 15 20

PA @ 20 Weeks Pregnancy
(kcal/kg/day; MAQ)

114

Figure 2: Correlations (r) between past measured (PAR) and recalled (MAQ) physical
activity at 32 weeks gestation

 

20

15‘

10'

 

PA @ 32 Weeks Pregnancy
(kcal/kg/day, PAR)

 

r-O.85

 

 

0 5 1'0 1'5 20
PA @ 32 Weeks Pregnancy
(kcal/kg/day; MAQ)

115

Figure 3: Correlations (r) between past measured (PAR) and recalled (MAQ) physical
activity at 12 weeks postpartum

 

 

 

 

20
E
3
E
3E 15‘ . .
89.
22%
ﬁg 10‘
on:
23‘
ﬁg 5‘ '0‘ 0.
© ‘ 0
<
Q ~ r-0.86
o . . .
0 5 10 15 20

PAR @ 12 Weeks Postpartum
(kcal/kg/day; MAQ)

116

 

Figure 4: Correlations (r) between past measured (PAR) physical activity and BMI at 20
weeks gestation

40

 

 

Current BMI

 

 

r--0.32

 

 

0 5 1'0 1'5 20
PA @ 20 Weeks Pregnancy
(kcal/kg/day; PAR)

117

 

Figure 5: Correlations (r) between past measured (PAR) physical activity and BMI at 32
weeks gestation

 

Current BMI

 

r-~O.42

 

 

 

0 5 1'0 1'5 20
PA @ 32 Weeks Pregnancy
(kcal/kg/day; PAR)

118

 

Figure 6: Correlations (r) between past measured (PAR) physical activity and BMI at 12
weeks postpartum.

 

 

Current BMI

10*

r--O.37

 

 

0

 

0 5 1'0 1'5 20
PA @ 12 Weeks Postpartum
(kcal/kg/day; PAR)

119

 

Figure 7: Correlations (r) between past measured (PAR) physical activity and percent
body fat at 20 weeks gestation

 

 

 

 

50*
O
O
40 ' .
O
z;
u- 30 .. . .
s y . .
.9
9:3 20; o
L.
a o
10‘ r--0.50
O
o I l T
O 5 1O 15 20

PA @ 20 Weeks Pregnancy
(kcal/kg/day; PAR)

120

Figure 8: Correlations (r) between past measured (PAR) physical activity and percent
body fat at 32 weeks gestation

 

Current 96 Fat

10‘ r--O.66

 

 

 

0 5 1'0 115 20
PA @ 32 Weeks Pregnancy
(kcal/kg/day; PAR)

12]

Figure 9: Correlations (r) between past measured (PAR) physical activity and percent
body fat at 12 weeks postpartum

 

Current 96 Fat

 

r--O.73

 

 

 

0 5 1'0 1'5 20
PA @ 12 Weeks Postpartum
(kcal/kg/day; PAR)

122

Figure 10: Correlations (r) between past recalled (MAQ) physical activity and BMI at 20
weeks gestation

 

Current BMI

 

r--O.56

 

 

o s . .
0 5 10 15 20
PA @ 20 Weeks Pregnancy

(kcal/kg/day; MAQ)

 

123

Figure 11: Correlations (r) between past recalled (MAQ) physical activity and BMI at 32
weeks gestation

4O

 

 

Current BMI

 

r--0.46

 

 

 

0 5 1'0 1'5 20
PA @ 32 Weeks Pregnancy
(kcal/kg/day; MAQ)

124

 

Figure 12: Correlations (r) between past recalled (MAQ) physical activity and BMI at 12
weeks postpartum

40

 

 

Current BMI

 

r--0.45

 

 

 

0 5 10 115 20
PA @ 12 Weeks Postpartum
(kcal/kg/day; MAQ)

125

Figure 13: Correlations (r) between past recalled (MAQ) physical activity and percent fat
at 20 weeks gestation

 

Current 96 Fat

 

 

 

 

 

0 5 1'0 1'5 20
PA @ 20 Weeks Pregnancy
(kcal/kg/day; MAQ)

126

Figure 14: Correlations (r) between past recalled (MAQ) physical activity and percent fat
at 32 weeks gestation

Current 96 Fat

 

 

 

 

 

50
O
40 .
O
' :r
so .' -.
20~ ' . ;
1 .
10* r--0.73
.
o . . .
0 5 10 1 5 20

PA @ 32 Weeks Pregnancy
(kcal/kg/day; MAQ)

127

Figure 15: Correlations (r) between past recalled (MAQ) physical activity and percent fat
at 12 weeks postpartum

50

 

30 f" '

Current %FAT

10‘ ra-OJO

 

 

 

0 5 1'0 1'5 20
PA @ 12 Weeks Postpartum
(kcal/kg/day; MAQ)

128

 

Figure 16: Correlations (r) between percent conﬁdence and past physical activity, both
measured (PAR) and recalled (MAQ) at 20 weeks gestation

 

 

 

 

 

 

 

20
o PAR-20
g o MAQ-20
g 15 T
gm
5 13‘
U)
155? 10
o .5
3 8
8 5
5.4
©
<
o.
o .

 

 

0 . 20 40 60 80 100

Overcoming Barriers (96 Confidence)

129

Figure 17: Correlations (1') between percent conﬁdence and past physical activity, both
measured (PAR) and recalled (MAQ) at 32 weeks gestation

PA @ 32 Weeks Pregnancy

(kcal/kg/day)

 

20

15“

_|
O

0

 

O PAR-32

 

 

O MAQ-32

 

 

O 20

40

60

O

 

 

O é.
O 00 0
_ do .0 863085

80 100

Overcoming Barriers (96 Confidence)

130

' .l ’5‘ 12'

 

W1...

Figure 18: Correlations (r) between percent conﬁdence and past physical activity, both
measured (PAR) and recalled (MAQ) at 12 weeks postpartum

 

 

 

 

 

 

 

 

20
E OPAR-PP
é o wmﬁ>
N 15
1?... 0
85‘
in?
5.2 10 I
0::
33.3
«1C. 0
" 5« :8
© 0 o
o
< ' o
o. 0 0
OJ .1'0-0588'.8: E

O 20 40 60 80 100

Overcoming Barriers (96 Confidence)

131

Figure 19: Visual representation of the past physical activity and current body size

relationship

Past Pregnancy PA

 

Current PA ’ Current Body Size
Past Pregnancy PA ’ Current PA
Past Pregnancy PA ' Current Body Size

How does current PA ﬁt into this relationship?

 

 

, Current PA , Current Body Size

1

Behavior inﬂuence Proximate cause

132

Appendix A

Anthropometric Equations

BMI (Body Mass Index)
BMI=kg-m2

BIA (Bioelectrical Impedance)
FF M (kg)= 0.493 (HT2/R)+0.141 (BW)+11.59 (Lohman,
1992)

% Body Fat
FM= BW—FFM
%BF= (FM/BW) X
100

*Equations were obtained from Applied Body composition Assessment by Heyward and Stolarczyk

MAQ Leisure Time Energy Expenditure Equations

Kcal/wk
(MET*WT)(dur/60)*freq=Kcal/wk

Kcal/day/kg
(Kcal/wk)/7=Kcal/day
(Kcal/day)/kg=Kcal/day/kg

*Equation obtained from Ainsworth et a1, Compendium of Physical Activities, 1993

Equation Key

FF M= fat free mass (kg) MET= metabolic equivalent

HT= height (cm) WT= weight of the subject (kg)

R= resistance (ohms) Dur= duration of the activity bout (min)
Freq: frequency of activity each week

BW=body weight (kg) (days/wk)

FM= fat mass (kg)

133

Appendix B
Subject ID#:

 

Modiﬁable Activity Questionnaire

Please circle all activities listed below that you have ever participated in more than 10 times
in your life:

Jogging (outdoor, treadmill) 1 Football/Soccer 14 Stair Master 27
Swimming (laps, snorkeling) 2 Racquetball/Handball/Squash 15 Fencing 28
Bicycling (indoor, outdoor) 3 Horseback riding 16 Hiking 29
SoftbalVBaseball 4 Hunting 17 Tennis 30
Volleyball 5 Fishing 18 Golf 31
Bowling 6 Aerobic Dance/Step Aerobic 19 Canoeing/Rowing/Kayaking 32
Basketball 7 Water Aerobics 20 Water skiing 33
Skating (roller, ice, blading) 8 Dancing (Square,Line,Ballrm) 21 Jumping rope 34
Martial Arts (karate, judo) 9 Gardening or Yardwork 22 Snow skiing (X-C/Nondic trk) 35
Tai Chi 10 Badminton 23 Snow skiing (downhill) 36
Calisthenicsfl’oning

exercises 11 Strength/Weight training 24 Snow shoeing 37
Wood Chopping 12 Rock climbing 25 Yoga 38
Water/coal hauling 13 Scuba Diving 26 Other 39
Walking for exercise 40

Please list the names of the activities chosen below:

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

134

List each activity from those circled, that you participated in during
the fall before you became pregnant ﬁve years ago. What was the
frequency, duration, and intensity of each activity listed?

Activity Frequency Duration Intensity

 

 

 

 

 

 

 

 

 

 

 

 

 

MET/wk KCAka

List each activity from those circled, that you participated in during
the second trimester of your pregnancy ﬁve years ago. What was the
frequency, duration, and intensity of each activity listed?

Activity Frequency Duration Intensity

 

 

 

 

 

 

 

 

 

 

 

 

 

MET/wk KCAka

135

List each activity from those circled, that you participated in during
the third trimester of your pregnancy ﬁve years ago. What was the
frequency, duration, and intensity of each activity listed?

Activity Frequency Duration Intensity

 

 

 

 

 

 

 

 

 

 

 

 

 

MET/wk KCALIwk

List each activity from those circled, that you participated in at 12
weeks postpartum from your pregnancy ﬁve years ago. What was
the frequency, duration, and intensity of each activity listed?

 

Activity Frequency ‘ Duration Intensity

l

 

 

 

 

 

 

 

 

 

 

 

 

MET/wk KCAka

136

List each activity from those circled, that you currently participate in
during a typical week. What was the frequency, duration, and
intensity of each activity listed?

Activity Frequency Duration Intensity

 

 

 

 

 

 

 

 

 

 

 

 

 

MET/wk KCAka

In general, how many HOURS each NIGHT do you sleep?

In general, how many HOURS per DAY do you usually spend watching television/using the computer?

 

Did you ever compete in an individual or team sport (not including any time spent in sports performed

during school physical education classes)?
Yes/No

If yes, how many total years did you participate in competitive sports?

137

Appendix C
Perceived Barriers Efﬁcacy Questionnaire

DIRECTIONS: List at least THREE barriers that you think interfere with your ability to
participate in regular physical activity for the next three months. Next to each barrier,
indicate the level of conﬁdence (ranging from 0% to 100%) that you could overcome that
barrier if it was present. Rank each barrier, 1 is perceived to be the hardest, 2 not as hard,
and so on down the list. Please feel free to elaborate on why you chose the barriers in the
space provided below.

Barriers % Conﬁdence Rank

 

 

 

 

 

 

 

 

 

 

 

 

Elaborate:

138

Appendix D

The Effect of Physical Activity During a Past Pregnancy on Current Physical
Activity, Barriers to Physical Activity, and Body Size; and Validation of a Historical
Recall Tool

Statement of Informed Consent

Participant’s Name:

 

Purpose of the Study:

The purpose of this follow-up study is to assess current body size, physical activity
levels, and barriers to physical activity, as well as to validate a historical physical activity
recall tool.

Your participation in this study is voluntary. You may decline to participate in given
activities if you so choose, refuse to answer speciﬁc survey questions, or withdraw from
the study at any time without penalty.

Description of the Study:

You are invited to participate in this follow-up project which involves coming into the
Human Energy Research lab (HERL) at Michigan State University (MSU) for one, two
hour follow-up visit approximately ﬁve years after the pregnancy which was examined in
the previous Maternal Activity Measurement and Assessment (MAMA) study. The one
visit will consist of body size measurement and data collection via survey instruments.
The details about the measurement and data collection are listed below.

Anthropometrics (body size measurements)

You will arrive at the HERL (Human Energy Research Lab) at Michigan State University
in comfortable clothing meant for movement. You will have your height and weight
measured, then we will measure your bioelectrical impedance (BIA) while you lay down
with electrodes on your hands and ankles. BIA will measure your fat free mass by
sending a small current through you body. BIA is based on the fact that water and
muscle conduct currents while fat tissue does not. Current body composition will then be
calculated from the measured fat free mass value and we will compare these values to
those collected in the study 5 years ago.

Data Collection

After the anthropometric data is collected, we will administer surveys to assess current
and past physical activity as well as current barriers to physical activity. Past physical
activity will consist of you remembering how active you were before, at 20 weeks, 32
weeks gestation, and 12 weeks postpartum for the pregnancy that was examined 5 years
ago. Current physical activity will be assessed by asking you to describe your activity for
the past week. The barriers to physical activity will be assessed by asking you to list your
current perceived barriers to physical activity participation, having your rank at least the

139

top three barriers chosen then have you determine your percent conﬁdence for
overcoming the listed barriers.

In addition to the data collection mentioned above, your medical records may be
examined for information regarding infant birth weight, initial pregnancy visit weight,
and gestational age.

Side Effects and Risks:
There are no physical risks involved in this study, therefore there should be no physical
side effects involved in study participation.

Beneﬁts:

For your participation you will receive a monetary award of $25. You will also learn
your current weight and body composition. Also, the information obtained in this
research will help future studies dealing with women, especially pregnant and postpartum
women and how their bodies change for the long-term after a pregnancy.

Conﬁdentiality:

All of your records from this study will be considered conﬁdential and your name will
not be identiﬁed in any publication or report from this study. Your privacy will be
protected to the maximum extent allowable by law.

Cost:
Your participation in this research study will not involve any additional costs to you or
your health care insurer.

Your signature below acknowledges your voluntary participation in this research study.
Such participation does not release the investigators, institutions, sponsor, or granting
agency from their professional and ethical responsibility to you. If you have any
questions about the study, please contact the investigator Jim Pivarnik Ph.D. by phone:
(517) 353-3520, by email: jimpiv@msu.edu. or regular mail: room 3 IM Sport Circle,
East Lansing, MI 48824. If you have questions or concerns regarding your rights as a
human subject, or are dissatisﬁed at any time with any aspect of the study, you may
contact, anonymously, if you wish, Peter Vasilenko, Ph.D., Chair of the University
Committee on Research Involving Human Subjects (UCRIHS) by phone: (517) 355-
2180, fax: (517) 432-4503, email: ucrihs@msu.edu. or regular mail: 202 Olds Hall, East
Lansing, MI 48824.

 

 

 

Signature of Study Participant Date

 

 

 

Address Home Work

 

Signature of Investigator

140

IIIIIIIIIIIIIIIIIIIIIIIIIIIIII

lllllllllllllllllllllllllIllllllllllllllllllllllllllllllllll