SITY LIBRARI IES lllllllllll’lll (”HUN/HM! H I ll. 1293 00067 8387 A..— LluflARY ' Michigan State University This is to certify that the thesis entitled DIET AND LIFESTYL’E VARIABLES AS DETERMINANTS OF SURVIVAL IN FORMER COLLEGE ATHLETES AND NON-ATHLETE CONTROLS presented by Mary Louise Sunman has been accepted towards fulfillment of the requirements for MASTER OF SCIENCE degmmin Department of Food Science and Human Nutrition fiwwemla QhD Major professor Date 5! fi/ng 0-7539 MS U is an Affirmative Action/Equal Opportunity Institution IVIESI_J RETURNING MATERIALS: PTace in book drop to LIBRARJES remove this checkout from W your record. FINES WTII be charged if book is returned after the date stamped beTow. DIET AND LIFESTYLE VARIABLES AS DETERMINANTS OF SURVIVAL IN FORMER COLLEGE ATHLETES AND NON-ATHLETE CONTROLS By Mary Louise Sunman A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of MASTER OF SCIENCE Department of Food Science and Human Nutrition 1987 ABSTRACT DIET AND LIFE-STYLE VARIABLES AS DETERMINANTS OF SURVIVAL IN FORMER COLLEGE ATHLETES AND NON-ATHLETE CONTROLS By Mary Louise Sunman The purpose of this investigation was to determine if former college athletes and their controls differed in body mass index, energy intake, fat intake, smoking habits and aerobic activity, and whether these variables, together with participation in college athletics, predicted years of survival. The sample studied consisted of 336 white male alumni who attended college prior to 1938. The sample was surveyed by mail with five health and life-style questionnaires between 1952 and 1985. Food frequency questionnaires and 24-hour recalls were included in the last three surveys. Results showed that more former athletes than controls were smokers in 1952, and that former athletes consumed more Koala/day than controls in 1985. The Cox Proportional Hazards Regression Model showed that aerobic activity was a significant predictor of survival. Subjects who survived later than 1985 were more aerobically active than those who died at an earlier date. ACKNOWLEDGEMENTS I would like to express my gratitude to Dr.Sharon Hoerr, my advisor and chairperson, for her invaluable help and guidance throughout my two years at Michigan State University. Drs. Chenoweth, Schemmel and Olson provided help and insight in their capacity as committee members. Particular thanks are due to Dr. Homer Sprague for statistical guidance. My studies would not have been possible without two years of support from the Fulbright Commission, and a grant from the British Universities North America Club. I would like to thank Dr. Dutson for arranging financial assistance through the Michigan Agricultural Experiment Station. I am indebted to Ron Carda for sharing his knowledge of the Michigan State longevity study, and to Richard Ivans for help with the Nutrient Database. Jon Alley and Karen Ellerman both contributed by coding dietary recalls, and Katie Richards patiently and accurately transcribed my scribblings. I would like to thank my husband, Roy, who provided moral support and technical assistance. ii TABLE OF CONTENTS LIST OF TABLES ........................................ iii LIST OF FIGURES ....................................... vi Introduction ........................................... 1 Review of the literature ............................... 2 Mortality statistics and trends ................... 2 The relationship between body mass index and mortality .......................... 3 Studies in which no relationship was found between body weight, body fatness and mortality ...................... 4 Studies in which a relationship was found between body weight, body fatness and mortality ...................... 6 Linear relationships ................ 6 J-shaped relationships .............. 7 U-shaped relationships .............. 9 Smoking as a confounder to weight- mortality relationships ........................... 11 The relationship between dietary intake and mortality .............................. 12 Diet quality and mortality ................. 12 Energy intake and mortality ................ 14 Fat intake and mortality ................... 15 Level of physical activity and mortality ......................................... 18 Statement of the problem ............................... 21 The hypotheses ......................................... 25 iii Methods ................................................ 26 Background to the study ........................... 26 The sample ........................................ 27 The questionnaires ................................ 27 Collection of the data ............................ 28 Height and weight .......................... 29 Dietary variables .......................... 29 Alcohol intake ............................. 31 Physical activity .......................... 31 Smoking habits ............................. 32 Years of survival ......................... 33 Management of the data ............................ 33 Validation of the methods ......................... 34 Dietary variables .......................... 34 Body mass index ............................ 35 Physical activity .......................... 38 Statistical analysis .............................. 39 Results and Discussion ................................. 42 Hypothesis one: Results ........................... 43 Comparison of dietary variables BMI and aerobic activity of former athletes and controls ...................... 43 Comparison of smoking habits between former athletes and controls ............... 45 Age as a confounder to dietary variables, BMI and aerobic activity ........ 47 Smoking as a confounder to dietary variables, BMI and aerobic activity ........ 49 Hypothesis one: Summary ........................... 51 Hypothesis one: Discussion ........................ 52 Hypothesis two: Results ........................... 60 Prediction of survival by the Cox Proportional Hazards Regression Model ...... 60 Comparison of the sample by survivors and those dying between the surveys ........ 61 Comparison of the sample by survivors and those dying between surveys and by former athletes and controls ............... 62 iv Hypothesis two: Summary ........................... 63 Hypothesis two: Discussion ........................ 68 Summary and Implications ............................... 71 List of References ..................................... 74 Appendices ............................................. 81 1) The questionnaires ............................ 81 2) Caloric standards for the Michigan State University longevity study activities ................................... 97 3) M80 Nutrient Database computer commands used ................................ 98 LIST OF TABLES. Table 1 Summary of the studies reviewed relating mortality and BMI ..................... 2 Comparison of dietary intakes of males in ‘NHANES II with those in the present study ...... 3 Comparison of dietary intakes of males in the National Food Consumption Survey 1977-78 with those in the present study ........ 4 Comparison of BMI for white males in a national sample 1977-80 with those in the present study ........................... 5 Comparison of dietary variables, BMI and aerobic activity between former athletes and controls in each year surveyed ....................................... 6 Categorization of respondents to each survey by smoking habits in 1952 ............... 7 Percentages of smokers and non-smokers in 1952 by athlete and control for each year surveyed .................................. 8 Percent of respondents to each survey who smoked in 1952 by category of smoking behavior ........................................ 9 Pearson product-moment correlations (r) of dietary variables, physical activity and BMI with age for each year surveyed ........ 10 Comparison of dietary variables, BMI and aerobic activity in those who smoked in 1952 and those who did not smoke in 1952 by year of questionnaire .......... vi Page 36 36 38 44 46 46 47 48 50 11 12 13 14 15 16 17 18 Comparison of aerobic activity (Kcals/week) between athletes, controls, smokers and non-smokers in 1975 ................ Significance levels of dietary variables, BMI and participation in college athletics for prediction of average years of survival from stepwise regression using the Cox Proportional Hazards regression model ........... Numbers and percentages of subjects surviving to 1985, dying between surveys and lost to follow-up by former athlete and control ......... Comparison of dietary variables, BMI and aerobic activity of subjects dead after each survey with survivors who responded to each survey .................... Dietary variables, BMI and aerobic activity of survivors compared by athletes and controls ........................... Dietary variables, BMI and aerobic activity of subjects dead before the subsequent survey compared by athletes and controls .................................... Comparison of aerobic activity (Heals/week) between athletes, controls, survivors and those who died between 1975 and 1985 ........... Comparison of energy intake (Heals/day) between athletes, controls, survivors and those who died between 1975 and 1985 ............ vii 51 60 62 64 65 66 67 67 LIST OF FIGURES Figure Page 1 An illustration of J-shaped and U-shaped relationships between weight and mortality ..................... 8 viii INTRODUCTION Improvements in lifestyle, including nutritional status, are accepted as being influential in the increased longevity that has occurred during the last century. Lifestyle variables are also blamed for the concomitant increase in the incidence of death from degenerative diseases such as cancer and heart disease (Morrison, 1983). This apparent paradox might be explained by human survival curves. It is apparent that although average life expectancy has increased, it is due to more people living longer rather than due to an increase in the maximum life span as is seen in animal models (Yu, Masoro, Murata, Bertrand and Lynd, 1982). In other words, the maximum age obtained in the population does not increase, but the increased numbers of people approaching this maximum age increases average longevity (Morrison, 1983). It is clear that any mechanism hypothesized to explain the relationship between lifestyle and longevity in an animal model will describe different phenomena from that observed in human populations. For example, attempts to 2 explain the relationship between energy intake and longevity in humans must be based on data from human populations and measure also a wide range of possible contributing or confounding lifestyle variables. The aim of this investigation is to assess body mass index, total energy intake, percent kilocalories from fat, smoking habits and participation in college sports and subsequent physical activity as predictors of life expectancy in a sample of male college graduates who have been followed longitudinally for 33 years. REVIEW OF THE LITERATURE WW Data from the United States Bureau of the Census indicate that life expectancy for US residents has been increasing steadily since 1900 (US Department of Commerce, 1984). More specifically, life expectancy for white males has increased from 46.6 years at birth in 1900, to 71.8 years at birth in 1984 (National Center for Health Statistics, 1985). In Michigan, life expectancy for men of all races was 71.2 years at birth in 1985 (Michigan Department of Public Health, 1987). Parallel to the increase in life expectancy has been a decrease in the death rate (number of deaths per 3 1000 of the population, excluding fetal deaths). The death rate for white males in the 0.8. has decreased from 17.7 in 1900 to 10.9 in 1970. The estimated death rate for white males in 1985-86 was 9.6 per 1000 population (National Center for Health Statistics, 1987). Leading causes of death in 1985 were major cardiovascular disease, which accounted for 45.8% of all deaths, and malignant neoplasms, accounting for 22.6% of deaths. Improvements in lifestyle, including nutritional status, are accepted as being influential in both the increased life expectancy and the rise of degenerative diseases such as cancer and heart disease which have occurred over the last century (Morrison, 1983). The following review focuses specifically on the relationship between body weight, selected dietary variables (fat and kilocalories), and physical activity to mortality and longevity. Body weight and body fatness have been used by epidemiologists and actuaries to predict longevity and future occurrence of disease. A comparison of studies relating body weight and body fatness to longevity is confounded by the use of: 4 1) Different methods and analytical techniques: 2) Differences in the weight indicies; 3) Age of the study population: and 4) Endpoints to which body weight was related. Also, the recognition of the confounding effects of covariables such as smoking and physical activity on the relationship between body weight and mortality has led to increasingly sophisticated statistical analyses in recent years and to reworking of the data sets of previous studies, often with new conclusions. A summary of the studies reviewed relating body mass index (BMI) and mortality is presented in Table 1. These are longitudinal studies in which BMI was related directly to mortality in large North American or European populations. These will be discussed according to the weight-mortality relationships found: no relationship or linear and non-linear relationships. 33nd1Qa_1n_Hh1Qh_flQ_BQlai1Qnah12_Ha&_Eand_B§1H§§n_BQd1 W. In only one study, the Seven Countries Study, was no relationship found between BMI and mortality (Keys et al., 1981). However follow-up was restricted to ten years, cigarette smoking was not considered, and the population studied was not from the United States. The Table 1. Summary of the studies reviewed relating mortality and BMI. Study lo.len Io. Deaths loses its ice at Snotins Popsl- lt/nort. follow entry consid. ation relation 09 foundt Seven 11,250 0 1241 Country Build 3.100, 500, 106. 6 B.P. 000 000 000 lanitobs 3963 0 199 Ihite- 10393 0 1122 hall lranins 1911 0 129 has Vanden- 1503 1464 n/e broucke Chicano 1233 0 246 Peoples Gas inerican 316, 419, 101, Cancer 000 000 000 10 40-59 lo European lone 6.6 15-69 Io 0.6. Linear 26 25-14 Io Pilots Linear 10 40-64 Yes British '1' 26 30-62 Yes Original '0' 25 40-65 Yes Dutch '0' 14 40-59 Yes 0.6. '0' 6 >30 Yes 0.5. '0' t [or definitions of J-shaped' 1, and Figure 1. and '0-shaped' relationships, see page 6 National Center for Health Statistics (NCHS) has carried out three major national health surveys, in 1960, 1971 and 1976. Trends of increasing BMI with decreasing mortality in these surveys would appear to refute a positive relationship between these two variables found in other studies. However in the analyses of these surveys by NCHS, the effects of confounders were not considered (Feinleib, 1985). Wm Lineamlationships. In two studies of large samples, linear relationships were found between weight and mortality. In the 1959 Build and Blood Pressure Study, a continuous and direct relationship between body weight and risk of mortality was reported, with lowest mortality occuring at 15% below the average weight of adults for any particular height (Society of Actuaries, 1959). This linear relationship between body weight and mortality was also found in the 1979 Build and Blood Pressure Study (Build Study 1979). In the Manitoba Study Rabkin, Matthewson and Hsu (1978) found that initial measurements of BMI were associated positively with the 26-year incidence of ischemic heart disease (p<0.01) in a sample of 3983 men 7 with a mean age at entry of 30.8 years. BMI was found to be a significant predictor of 390 cases of ischemic heart disease after adjustment for the effects of age and blood pressure (p<0.01). Longevity was associated positively with a below average BMI. The mortality ratio (observed/expected deaths) was 50% in those with a BMI of less than 22.5, but 200% in those with a BMI greater than 27.6. A “J" shaped relationship between weight and mortality is one in which the mortality at low weights is higher than that in the mid-range weights but not as high as that in the upper weights. A "J-shaped" relationship is illustrated in Figure 1. In the Whitehall Study, Jarrett, Shipley and Rose (1982) reported ten-year mortality rates in men aged 40-64 years in relation to BMI at initial examination in 1967. In men aged 40-49 years, mortality increased with increasing BMI. However in men older than 49 years, the highest mortality was seen in those with the lowest BMI. Pooled data from all ages showed a "J-shaped” relationship between mortality and BMI. Increased mortality at low BMI was not explained by cigarette smoking. CO A J'SHAPED CURVE OF MORTALITY AND WEIGHT INDICES i MORTALITY INDEX ~{> HEIGHT INDEX A U-SHAPED CURVE OF MORTALITY AND WEIGHT INDICES X Lu 0 Z ). t .1 ( y.- m 0 I VEIGHT INDEX ihi_il1iuELraid4ni4af_thuuaeed_juniiflzshmuaai 9 A "U-shaped” relationship between weight and mortality is one in which mortality at low and high weights are equally high, and both are higher than mortality at the middle range of weights. A "U-shaped" relationship is illustrated in Figure 1. In 1980 Sorlie, Gordon and Kannel reported the findings of an analysis of the 24-year, follow-up data from the Framingham Study. They found a "U-shaped" curve of mortality risk against body weight, with minimum mortality at the average weight of the group. The increased risk of mortality with low body weight remained after correction for the significant association of low body weight with high, short-term mortality from chronic disease. The increased proportion of lean persons who smoked did not account entirely for the excess mortality in the leanest group. Despite this finding, three years later Garrison, Feileib, Castelli and McNamara (1983) suggested that the increased mortality in low weight men seen in the Framingham Study was a reflection of the higher proportion of lean men who were cigarette smokers. Vandenbroucke et al. (1984) in a follow-up of 1503 men aged 40-65 years at the start of the study in 1954, also observed high numbers of smokers in the group with the lowest BMI. Twenty-five year mortality plotted against BMI produced a "U-shaped" curve for both smokers and non-smokers. The shape of the mortality curve did not 10 change when mortality from the first five years was excluded to reduce the effect of chronic disease undetected at the srart of the study. In a 14-year longitudinal study, Dyer, Stamler, Berkson and Lindberg (1975) examined data from 1233 white male employees of the Chicago Peoples Gas Company, aged 40- 59 years at the start of the study in 1958. The relationships between BMI and total mortality, cardiovascular/renal deaths, and coronary heart disease deaths were examined. Variables measured were age, systolic blood pressure, serum cholesterol, and number of cigarettes smoked. Deaths from all causes examined, from cardiovascular/renal disease and from coronary heart disease, were found to have significant quadratic relationships to BMI for the total sample. The level of significance was not reported. When smoking was considered as a variable, BMI continued to show a significant quadratic relationship to all causes of death except coronary heart disease amongst non-smokers. Mortality curves were “U-shaped" with lowest risk of mortality at 25- 35% above Ideal Weight (1983 Metropolitan Life Insurance Tables). In 1959, the American Cancer Society began a follow- up study of 750,000 men and women who were free from disease and who had not reported weight loss at the start of the study. This group was followed until 1973, and the mortality findings analyzed according to variations in 11 weight by height from the average weight for the population studied (Lew and Garfinkel, 1979). The lowest mortality rates occurred in people weighing 80-89% of the average weight, and who did not smoke cigarettes. Within each category of smoking habits, the lowest total mortality generally occurred in those persons slightly below average weight. Mortality from cancer showed a "U-shaped" curve when plotted against the relative weight index, even when controlling for the effects of cigarette smoking. no; .- .-; ; ..‘....— . ,- ;. -... ; ; ;_ os:o ..-. In recently published results from the second National Health and Nutrition Examination Survey (NHANES II), smokers have been shown to have a lower mean body mass index than non-smokers in a representative sample of the 0.8 population. Smokers had a mean BMI of 24.610.1, whereas non-smokers had a mean BMI of 25.710.1. This difference was significant at p<.05 (Albanes, Jones, Micizzi and Mattson, 1987). As discussed previously, many studies of weight-mortality relationships have found differences between smokers and non-smokers (Feinleib, Castelli and McNamara, 1983; Sorlie, Gordon and Kannel, 1980; Vandenbroucke et al., 1984). In these studies, smokers tended to be lighter than non-smokers and to have higher mortality rates. Failing to allow for the effect of smoking on the weight-mortality relationship will therefore increase the weight at which lowest mortality occurs, due 12 to excess mortality in low weight smokers. Smoking is an important confounder to weight-mortality relationships. W. The following review focuses on the relationships between overall diet quality, total kilocalories, and fat intake to mortality. The many epidemiological studies on the relationship of macronutrients to cancer and heart disease mortality are not reviewed here. D12i_anlill_and_Minaliix, Only two recent investigations have focused on the effect of overall diet quality (as determined by consumption of certain types of foods) to mortality. In a 21-year follow-up study on 27,530 adult Seventh Day Adventists, Kahn, Phillips, Snowdon and Choi (1984) investigated the association between mortality from all causes and frequency of consumption of 28 specific foods. Food consumption was measured in 1960 by a self- administered food frequency questionnaire. Death certificates were subsequently obtained for subjects who died before 1981. All cause mortality (deaths from all causes except violence or trauma) showed a significant negative association at the 0.01 level with green salad consumption. At the same level, all cause mortality showed a significant positive association with meat and egg 13 consumption. These associations remained significant after adjusting for age, sex, smoking history, chronic disease and age at first exposure to the Adventist church. The total nutrient and calorie composition of the diet was not examined. No further dietary information was obtained after the initial questionnaire, making it impossible to evaluate whether eating habits changed over the follow-up period. In the second study Nube, Kok, Vandenbroucke, Heide-Weissel and Heide (1987) applied a scoring system to the diets of 2,820 middle-aged Dutch civil servants and their spouses. The interviews were part of a health examination survey in the early 1950's. The diet score was derived from consumption frequency data of 10 food items. These were white bread, brown bread, milk, porridge or yogurt, potatoes, vegetables, meat, fish, eggs and fruit. In 1985, 25-year age adjusted survival was calculated and compared to diet score, with higher scores representing a more “prudent“ diet than did low scores. In men, a significant linear trend (p<0.01) was found between diet score and 25-year age adjusted survival. However in women, no relationship of diet score to longevity was observed. Again, no nutrient analysis of the dietary intakes was made, and no data were collected on eating habits subsequent to the initial survey. The items selected as representative of a prudent diet did not cover a wide range of foods. 14 Enerzx_ln1ake_and_Mcrialiix. Total energy intake has been cited as a determinant of longevity in animals, with increased longevity being associated with restricted caloric intake (Yu, Masoro, Murata, Bertrand and Lynd, 1982; “Limited food intake", 1982). However animal models have limited applicability to the complex interrelationships between lifestyle and longevity found in human populations. From human survival curves it is apparent that although average life expectancy has increased, it is due to more people living longer rather than due to an increase in the maximum lifespan as is seen in the animal models. The maximum age obtained in the population does not increase, but the increased numbers of people approaching this maximum age increases average longevity (Morrison, 1983). It is clear that any mechanism hypothesized to explain the relationship between energy intake and longevity in an animal model will describe a different phenomena to that observed in human populations. Attempts to explain the relationship between energy intake and longevity in humans must be based on data from human populations and measure also a wide range of possible contributing or confounding lifestyle variables. Kushi et al. (1981), reporting data from 1001 middle-aged men, found no significant relationship between energy intake in 1959 and subsequent 20-year mortality from ronary heart disease. Dietary information was obtained ‘by diet history and coded as food frequencies. When 15 adjusted for age, cigarette smoking, blood pressure and serum cholesterol, energy intake was not a significant predictor of coronary heart disease mortality. The association of BMI, an indicator of obesity from weight and height measurements alone, with increased mortality risk has been discussed previously. Obesity has been cited as a risk factor for hypertension, hypercholesterolemia and diabetes (Van Itallie, 1985) all of which are conditions considered to contribute to some of the major causes of mortality. However often changes in BMI are assumed to reflect changes in caloric intake, without consideration of other components of energy balance such as the type of energy intake, i.e. fat intake, or the physical activity. This assumption might or might not be valid. Willi. The results of studies relating fat intake to morbidity and mortality should be interpreted with care, considering the high degree of collinearity between fat and energy intake. There is little evidence linking fat intake directly to total mortality. However high intakes of fat have been linked to both the incidence of coronary heart disease and the incidence of cancer in various body sites. Cardiovascular disease is the most frequent cause of death among men in the United States (National Center for Health Statistics, 1985). Cancer is the second most frequent 16 cause of death, accounting for one in three deaths (American Cancer Society, 1985). In 1982, the Committee on Diet, Nutrition and Cancer, National Research Council, concluded that of all dietary components studied, the combined epidemiological and experimental evidence was most suggestive of a causal relationship between dietary fat intake and incidence of cancers of the colon and breast (Committee on Diet, Nutrition and Cancer, National Research Council, 1982). Similar but less consistent correlations have been reported with cancers of the prostate, ovary and endometrium (Armstrong and Doll, 1975). Schenkler (1976) examined the relationship between diet and longevity in 28 elderly women. As daily intake of fat increased by one gram, lifespan decreased by 44 days (r = -0.27). When fat intake was expressed as percent of calories, the negative correlation with lifespan became stronger (r = -0.35). No other dietary components correlated significantly with life span. In 1981, Shekelle et al. reported the results of a 20-year follow-up study of diet, serum cholesterol and death from coronary heart disease in 1900 men aged 40-55 years at the start of the survey in 1957. Scores summarizing each participant's intake of energy, cholesterol, saturated and unsaturated fat were calculated. Using logistic regression, positive associations were found between diet score and 19-year risk 17 of death from coronary heart disease (p<0.01). This finding persisted after adjustments for change in BMI and smoking habits. However these dietary variables (kilocalories, cholesterol, saturated and unsaturated fat), were not related significantly to risk of death from all types of cancer grouped together, or from all other causes of death grouped together. In an analysis of dietary information from the Seven Countries study, Keys et al. (1981) found that both the total death rate from all causes, and coronary heart disease death rate were not related to percent kilocalories as fat in the diet. The study covered 12,763 men in 16 cohorts from seven countries. Dietary data were collected by a variety of means from 24-hour recall questionnaires to seven-day weighed intakes. Although mortality was not related to relative weight, physical activity or total fat content of the diet - blood pressure, serum cholesterol and percent kilocalories from saturated fat were significant predictors of all cause mortality and coronary deaths. Percent kilocalories from saturated fat had a moderately strong positive correlation with mortality from all causes (r = 0.47), and a strong positive correlation with coronary heart disease mortality (r = 0.84). Significance levels were not reported. In 1983, Sidney and Farquahar reported international per-capita nutrient intake and age-adjusted total cancer mortality rates in 20 countries. Total 18 calorie intake was correlated positively with cancer mortality (r = 0.66, p<0.01). Total fat intake also correlated with cancer mortality (r 0.68, p<0.01), as did percent of kilocalories from fat (r 0.67, p<0.01). In 1986, however, Heilbrun, Hankin, Nomura and Stemmerman reported the original fat intake of 99 men who subsequently developed cancer of the colon during 14 years of follow-up. This was compared to the original fat intake of 378 men who remained free of any cancer. Mean fat intake was lower in the men who subsequently developed colon cancer (p = 0.05). Differences in total energy intake were not reported although BMI was found to predict the development of cancer in the same population (Nomura, Heilbrun and Stemmerman, 1985). L21al_Qf_Eh1aiQal_AQ&ixiil_and_MQzlalilz Many early studies on the effect of athletic participation on mortality and longevity failed to compare college athletes with a control group of college graduates. As early as 1926, Greenway and Hiscock observed that college graduates had a higher life expectancy than non-graduates. This finding emphasized the importance of selecting valid controls for college athletes, and of avoiding the use of national statistics for comparison. In 1954, Rock found no difference between the survival rates and average age of death of honors graduates, college athletes and a random sample of students 19 from the 1860-1900 Cambridge University classes. Seven- year, follow-up data of the 1952 Michigan State survey, reported by Montoya, Van Russ and Nevai (1962), indicated that there was no difference in the longevity or cause of death, excluding violent deaths, between former college athletes and their controls. Mean age of death for athletes was 62 years, compared to 64 years for non- athletes. This difference was not statistically different at an alpha of 0.05. It might be expected that participation in college athletics would have a minor influence on health parameters as compared to habitual, long-term activity levels from graduation onwards. In 1984, Paffenbarger, Hyde, Wing and Steinmetz reported that personal athleticism altered trends in lifestyle and coronary heart disease. Analysis of 572 first heart attacks among 16,936 Harvard alumni between 1962 and 1972, and 1,143 total deaths between 1962 and 1978, showed that habitual post-college exercise, but not sports participation in college, predicted low coronary heart disease risk. Sedentary students who became active alumni acquired low risk. Exercise benefit was independent of contrary lifestyle variables such as obesity or cigarette smoking. Total mortality was related inversely to levels of physical activity (p<.001). Further analyses of these data, reported by Paffenbarger et al. in 1986, confirmed this trend. Death rates declined steadily as energy expended in 20 physical activity increased from 500 to 3500 kilocalories/week (p of the trend <0.0001). In a longitudinal study of 2,622 female former athletic alumnae and 2,776 female non-athletes aged 21-70 years, Frisch et a1. (1985) reported a significantly lower relative risk of developing cancer of the breast and reproductive system in former athletes compared to non- athletes. The relative risk for cancer of the reproductive system was 2.53 (95% Confidence Interval 1.17.5.47), and the relative risk of breast cancer 1.86 (95% Confidence Interval 1.00,3.47). The authors concluded that long-term athletic training might lower the risk of breast cancer and cancers of the reproductive system. Death rates from cancers of these sites were not reported. It is not clear whether site specific cancers are more common in male athletes than controls. Olson, Montoye, Sprague, Stephens and Van Huss (1978) reported no significant differences in causes of death between male athletes and non-athletes after 23 years of follow-up in the Michigan State longevity study. However at the present stage of analysis it is impossible to ascertain the incidence of disease in those subjects who did not die. It is not possible, therefore, to comment on the differences in disease incidence between male athletes and controls, only on numbers and causes of death. STATEMENT OF THE PROBLEM. From the literature reviewed it is apparent that although widely cited as major contributors to some of the main causes of morbidity, there is a lack of unequivocal evidence that diet and other lifestyle factors are significant determinants of overall life expectancy. The weaknesses of studies to date may be summarized in the following points: 1) Some study populations have been poorly defined preventing disaggregation into subgroups which might have results significantly different to the results obtained for the sample overall. Examples of such disaggregation are by age, race, socio-economic status and BMI. A significant association between the dependent and independent variables in one subgroup might be masked by an opposite association in another group. The unmasking of different weight mortality relationships in smokers and non- smokers in the Framingham study illustrate this point (Sorlie et al., 1980; Garrison et al., 1983). 2) Conversely, other studies have examined populations with unique characteristics making extrapolation of results to the general population invalid. An example was the Manitoba study in which 0.8. and Canadian pilots were selected for the study population (Rabkin et al., 1978). 21 22 3) Although in many studies the study populations have been large, usually the length of follow- up has not been long. The results have been based on relatively few deaths. An example was the Chicago Peoples Gas Study, which included a sample size of 1233, but based mortality data on only 246 deaths (Dyer et al., 1975). 4) There is no concensus on which sets of variables form an optimal predictive model for mortality. The variety of variables studied vary from study to study. A variable that is a significant predictor of mortality in one study might not be significant in a more comprehensive model. An example is the different conclusions that are drawn from one data set when cigarette smoking is or is not included in the analysis (Sorlie et al., 1980; Garrison et al., 1983). 5) Many studies used normative data from outside the study population, introducing a possible source of error in evaluation of results. Nationally generated statistics such as life expectancy might not reflect accurately what would be normative in selected samples of the same population. This is particularly true when subjects were not recruited into the study randomly, but were selected on the basis of characteristics such as age, sex and geographical location. More extreme selection criteria such as occupation make it increasingly unlikely that the study population will match the general population from which the normative data were generated. An example 23 is the selection of U.S. and Canadian pilots as a study population (Rabkin et al., 1978). The present study overcomes many of the problems described above, and will therefore make a significant contribution to our knowledge of the role of diet and related lifestyle variables to mortality. More specifically this study has the following advantages: 1) The study group is well defined and homogeneous, yet pertains to a large population (white males from a nonmetropolitan area). In 1970 the total adult white male population of the U.S. was 86,720,987. Of these 24,510,744 or 28.3% were living in rural areas. In 1970 26.3% of the adult white male population completed a college education (U.S. Bureau of the Census, 1980). There is no information on differential levels of education between rural and urban areas. Assuming that 26.3% of adult rural males are college graduates, a total of 6,446,326 males fall into the classification of rural white graduates. This represents 7.5% of the total adult white male population in the U.S. 2) The length of the follow-up has been unusually long. The group has been followed for a total of 35 years to date. Further follow-up is planned. 3) As a consequence of the above two points, the total mortality rate in the sample has been high. By combining expected years of life for those still alive with age at death for those deceased, it is possible to use data 24 from the entire sample to create an independent variable "years of survival" for each subject. This, together with the homogeneity of the sample overcomes the drawback of the relatively small sample size in other studies. 4) Average survival will be derived from mortality data from within the sample, avoiding possible error by comparison with normative data from an incompatible population. 5) Information on a wide range of health and lifestyle variables has been collected longitudinally. This allows many confounders to be examined and controlled for when examining the complex relationships between lifestyle, morbidity, mortality and survival. Only parts of the total data available will be reported here. 6) Dietary data has been collected consistently and in two forms, 24 hour recall and food frequency questionnaire, allowing cross validation during analysis. In summary, the value of this study lies in the extent and length of the longitudinal data, allowing analyses of the relationships between lifestyle variables and mortality in a well defined population of men. hypotheses. THE HYPOTHESES. Data will be used to address the following null There is no difference in average years of survival, body mass index, total energy intake, percent kilocalories from fat, total fat intake, smoking habits and aerobic activity between male, former college athletes and their controls. Body mass index, total energy intake, percent kilocalories from fat, total fat intake, smoking habits, participation in college athletics and levels of aerobic activity are not predictive of average years of survival in college men. 25 METHODS. W. The original impetus for this investigation was a national Phi Epsilon Kappa study of the longevity and morbidity of college athletes that was begun in 1950. Although never completed nationally, the Michigan State University portion of this survey was completed, and the results comparing the life expectancy and cause of death in male athletes and non-athletes were reported by Montoye, Van Huss, Olson, Pierson and Hudec (1957). In 1952, addresses were obtained for 1,129 varsity letter winners who had participated in Michigan State athletics before 1938. A stratified, random sampling technique was used to select a non-athlete control from the student directory for each athlete in the study. The control attended Michigan State in the same year and class as the paired athlete. Athletes and non-athlete controls were therefore age matched. Mean age difference between athletes and controls was 0.05 years (Montoye et al., 1959). Mailed questionnaires were sent to each athlete and control. Of these, 625 athletes and 557 non-athletes returned the information requested. These subjects formed the basis of the longitudinal study. All living respondents were mailed repeat questionnaires in 1960, 1967, 1975 and 1985. It is from these returns that the 26 27 sample for the present study was drawn. W. The present study comprises approximately half of the records which were collected from 1967-1985. The subjects selected were former footballers, former track athletes and their respective controls. This resulted in a sample of 338 subjects, 213 of whom were former athletes and 125 of whom were controls. The difference in the numbers of athletes and controls is as a result of a decreased response rate in the control group (Montoye, 1967). Although age matched at the start of the study, each former athlete was not paired with his original control in this sample. This was because of deaths and non- response in both groups. Footballers and track athletes were selected as they represented two well defined groups of athletes, with different body types typically associated with the two different sports. The remaining athletes not investigated in this study were letter winners in a wide variety of sports such as swimming, hockey, baseball and cross country running. Thuuesiionnaizes. Questionnaires were sent to subjects in 1952, 1960, 1967, 1976 and 1985. Information was requested on a wide range of health and lifestyle variables, including height, weight, history of illness, physical activity, 28 smoking habits, drinking habits and family history (Appendix 1). From 1968 onwards, dietary information was obtained in two forms: a 24-hour recall and a food frequency questionnaire based on thirty foods and food groups. The format for collection of the dietary information was recommended by faculty from the department of Food Science and Human Nutrition under the direction of Dr. Olaf Mickelsen. Wis. For each of the surveys returned in 1967, 1976 and 1985 the following data were extracted for use in the present study: 1) Name, date of birth. Subjects were assigned an identification number which was used in all analyses to ensure anonymity. 2) Date the questionnaire was completed. 3) Year of death, if subject was deceased. 4) Height. 5) Weight. 6) Dietary recall. 7) Food frequency questionnaire. 8) Consumption of alcoholic beverages. 9) Level of physical activity. 10) Smoking habits in 1952. 29 Information on all dietary variables, height and weight and alcohol intake was collected from the original survey returns which are archived in Jenison Fieldhouse, Michigan State University. Level of physical activity and smoking habits in 1952 were taken from a database of results which have been coded previously. A more detailed explanation of the collection of each variable is given below. Hfiizhi_and_nniihl- Self-reported height and weight were collected at each survey. Reports were made in feet and inches for height, and pounds for weight. These were converted to metric equivalents before being recorded. Body mass index (BMI) or Wt(kg)/Ht(m)2 was calculated for each subject in the years they responded to the survey. Dielazx_1aziahles. Subjects were asked to complete one open format 24-hour recall and one food frequency questionnaire in each of the last three surveys (1967, 1975, and 1985). Instructions and a guide to portion sizes was provided in the questionnaire (Appendix 1). A total of 667 24-hour recalls were analyzed for the present study, using the Michigan State University Nutrient Database which is one of the largest in the country and contains nutrient analyses for over 5500 foods (Leveille, Zabik and Morgan, 1983). 30 Computer commands for the program are shown in Appendix 3. All of the dietary information was coded under the supervision of the author, allowing standardization of common portion sizes, coding technique, and constancy of judgment error. The food frequency questionnaires were used for clarification of 24-hour recalls. The most common examples of this clarification were in ascertaining whether sugar was taken in tea and coffee or on breakfast cereal. If these items were not recorded on the 24-hour recall as having sugar added, the response to the items on the food frequency questionnaire "Sugar: on cereal, Daily/Weekly/Never", and "Sugar: in coffee, tea, etcetera, Daily/Weekly/Never“, were examined and the appropriate adjustments made to the recall. Otherwise, results from the food frequency questionnaires are not reported in this study. Analysis was made for over 80 nutrients, from which the following were extracted for use in the present study: 1) Percent kilocalories from protein. 2) Percent kilocalories from fat. 3) Percent kilocalories from carbohydrate. 4) Percent kilocalories from alcohol. 5) Total kilocalories. 6) Total fat in grams. 31 AlQQth_1n&ak§. Questions on alcohol intake were asked separate to the dietary recall and food frequency questionnaire (Appendix 1). Responses were reformulated in terms of equivalent daily intake of beer, wine or liquor and added to the 24-hour recall. If alcoholic beverages were already recorded on the recall, no further alcoholic beverages were added. Consumption of alcoholic beverages "less than once a week" was interpreted as one 100z bottle of beer a week, one 602 glass of medium white table wine a week or one shot of 90% proof whiskey a week for the the categories beer, wine and liquor, respectively. These were selected as they were the alcoholic beverages recorded most frequently on the 24-hour recalls. Ehxsical_actixiix. Physical activity was expressed as kilocalories of aerobic activity per week, as calculated by Quinn (1987). Calculations were made for the 1975 and 1985 surveys only. Information collected from the 1967 questionnaire was considered as lacking sufficient detail to allow for the calculation of activity levels. Quinn adapted the work of Bannister and Brown (1968), and Howley and Glover (1976) in order to arrive at a figure of aerobic activity per week. A table was formed listing the caloric expenditure in kilocalories/min/lb body weight for each of the activities listed on the 1975 and 32 1985 questionnaires. Aerobic activities were considered to be those which "utilized the major muscle groups and were performed at an intensity considered appropriate for conditioning" (Quinn, 1987). They were cycling, jogging, lawn mowing (power or hand), golf (walking), walking, rowing, skating, cross country skiing, snow shoeing, dancing, swimming and calisthenics (Appendix 2). Caloric expenditure was calculated by multiplying the subjects weight in pounds by the number of minutes each activity was performed per month, as reported on the questionnaire (Appendix 1). This figure was multiplied by the caloric expenditure associated with that activity to arrive at caloric expenditure per month. This was repeated for each month, and a mean monthly caloric expenditure calculated. This figure was divided by 4.2 (4.2 weeks/month) to arrive at weekly aerobic activity in kilocalories. Smekinz_hahits. Cigarette smoking as of 1952 was recorded in one of four categories: No cigarette use, light smoker (less than 20 cigarettes per day), moderate smoker (20 to 40 cigarettes per day), and heavy smoker (more than 40 cigarettes per day). Smoking habits from subsequent surveys were not coded for computer analysis at the time of this study. Statements about smoking behavior can only be related to whether a subject was a smoker or non-smoker at 33 the beginning of the survey in 1952. MW. The independent variable, average years of survival, was calculated for each year surveyed. The variable represents the mean number of years survived by the sample at each survey. For each subject, the value for the variable was either: 1) Age at death if the subject had died, 2) Age in the year of the survey if the subject was still alive, or 3) The age of the subject as of 1980 if the subject was lost to follow-up. By calculating years of survival, a value for the independent variable was available for each subject. If age at death alone had been used as an independent variable, inferences about the influence of the dependent variables on survival would have been restricted to those subjects who died prior to 1985. This would have reduced the sample size by approximately 50%. WW. Dietary analysis was performed on the mainframe computer at Michigan State University and results transferred to magnetic tape for storage. Once calculated, all variables used in the present study were copied from magnetic tape to the mainframe computer and assembled into 34 one data file. This was then downloaded to a floppy disk and edited using a word processor. Age (year of birth minus year of questionnaire), and body mass index (weight(kg)/height(m)2) were calculated using a spreadsheet. Edited files were transformed to ASCII format for subsequent statistical analysis on both the mainframe and microcomputer. YalidatiodeatamllectioLmetths. Dietazx_1ariahles. Although 24-hour recalls of food intake do not provide accurate estimates of the usual intakes of individuals, they have been found to be valid for determining the intake of groups (Madden, Goodman and Guthrie, 1978). The larger the sample, the more reliable the estimate, and the smaller the standard deviation for any particular nutrient (Beaton et al., 1979). Validation of the use of 24-hour recalls in this study was made by comparing the dietary intakes of the sample in the present study with the results of two national surveys. The numbers of subjects in the present study were not of the scale seen in national surveys, and consequently standard deviations for some variables are high. In all three surveys dietary intakes were similar to those found in both the National Health and Nutrition Examination Survey, NHANES II (National Center 35 for Health Statistics, 1983) and the Nationwide Food Consumption Survey of 1977-78, NFCS (U.S. Department of Agriculture, 1983). It is pertinent to compare the data from NHANES II with those from the present study, because methods of collecting the dietary information were comparable (24-hour recall and food frequency questionnaire), although the sampling techniques were different. Comparison of the present study with NHANES II is shown in Table 2. NHANES II data were collected for men up to the age of 74 years. In the USDA Nationwide Food Consumption Survey, three day records were used to collect dietary information. Data were reported for men aged 75 years and over. Comparison with the National Food Consumption Survey is shown in Table 3. Because of the comparability of these data, the author is confident that the collection of dietary information in the present study was as accurate as that in the two nationwide surveys cited. This is true despite the fact that both of the nationwide surveys collected data through personal interview, whereas in the present study data were self- reported. EQdLmasLindex. Body mass index (BMI) was the index of adiposity selected for use in the present study. BMI has a high correlation (r = 0.666) with the amount of body fat as estimated from body density, particularly when age is taken 36 Table 2. Comparison of dietary intakes of males in NHANES II with those in the present study. NHANES II MICHIGAN STATE STUDY 1977-1980 1967 1975 1985 AGE 55-64 65-74 64.0 68.9 75.7 YRS n=1227 n=1199 n=336 n=213 n=115 Kcals/ 2071 1829 2143 1934 1779 day % Kcals 16.2 16.0 15.8 15.6 16.1 protein % Kcals 39.2 37.9 41.6 41.0 38.0 fat % Kcals 43.8 44.6 40.6 41.0 43.5 carb Fat 86 0 75.0 98 6 86 2 72 2 s/day Table 3. Comparison of dietary intakes of males in the National Food Consumption Survey 1977-78 with those in the present study NFCS MICHIGAN STATE SURVEY 1977-1978 1967 1975 1985 AGE 51-64 65-74 >74 64.0 68.9 75.7 YRS n=2161 n=1049 n=465 n=336 n=213 n=115 Kcals/ 2158 1913 1866 2143 1934 1779 day % Kcals 16.7 14.6 16.1 15.8 15.6 16.1 protein % Kcals 42.8 41.0 41.2 41.6 40.5 38.0 fat % Kcals 39.2 42.2 42.8 40.6 41.0 43.5 carb Fat 102.6 87.1 85.4 98.6 86.2 72.2 s/day 37 into consideration (Norgan and Ferro-Luzzi, 1982). It has a low correlation (r = 0.062) with height, and is generally accepted as the most satisfactory index of adiposity based on weight and height alone (Keys, Fidanzo, Karvonen, Kimura and Taylor, 1972). The recommendation from the 1982 National Health and Nutrition Examination Survey workshop on body weight, health and longevity was to report weight as body mass index in order to facilitate comparability of data ("Body weight", 1985). When compared to national figures for BMI of white males aged 55-74 years, the mean BMI of the total sample in 1967, 1975 and 1985 fell on the 50th percentile (National Center for Health Statistics, 1983). Comparisons with NCHS data are shown in Table 4. In all three surveys of the Michigan State study, the median value for BMI (50th centile), was identical or close to the mean value. This suggests that BMI was normally distributed in the sample as it is in the general population. In the MSU study BMI was calculated from self-reported height and weight. Although the level of accuracy in reporting height and weight has been found to vary from study to study, it is generally agreed that the degree of inaccuracy involved in the self report of these measures is not sufficient to significantly bias results (Palta, Prineas, Berman and Hannan, 1982; Stewart, Jackson, Ford and Beaglehole, 1987; Stunkard and Albaum, 1981). It should be noted, however, that in none of the 38 Table 4. Comparison of BMI for white males in a national sample 1977-80 with those in the present study. NCHS MICHIGAN STATE STUDY 1977-1980 1967 1975 1985 AGE 55-64 65-74 64 0 68 9 75 7 (YRS) n=1086 n=1045 n=336 n=213 n=115 Mean BMI 26 1 25.6 25 7 25 8 25 4 50th 25 8 25.5 25 5 25 8 25 1 centile studies cited were the subjects older than 60 years. The mean age in the 1967 survey of the Michigan State sample was 64 years. The close agreement of national mean and median figures for BMI with those in the present study suggests that the method of data collection used in the Michigan State study was as accurate as those employed by the NCHS. Ehxsical_aciixitx. Information on physical activity levels was obtained in a format based on the Minnesota Leisure Time Physical Activity (LPTA) questionnaire (Taylor, Jacobs and Schucker, 1978). This instrument has been used extensively in clinical and cardiovascular surveys. Folsom, Jacobs, Caspersen, Gomez-Martin and Knudsen (1986) assessed the test-retest reliability of the Minnesota LTPA questionnaire at five-week intervals in 140 adults from a general population sample. They reported a Spearman rank 39 correlation coefficient between the test and retest of r = 0.88 (p<0.001). The Minnesota LPTA questionnaire has also been validated against duration of treadmill exercise (Leon, Jacobs, and DeBacker, 1981). However it is important to note that in the validations cited, total leisure time activity was calculated. In the present study, only aerobic leisure time activity was calculated; anaerobic leisure time activity was not. The LTPA questionnaire has not been validated for aerobic activity alone. HXEQihefiifi_Qne. There is no difference in average years of survival, body mass index, total energy intake, percent kilocalories from fat, total fat intake, smoking habits and aerobic activity between male, former college athletes and their controls. Descriptive statistics were performed on all variables for the years 1967, 1975 and 1985. Within each year the sample was subdivided into former athletes and non- athlete controls. Differences between athletes and non- athletes for each variable were tested using the Student's 4O t-test. A probability level of .05 or less was defined as statistically significant. To investigate the possible effect of age on the variables measured, Pearson product-moment correlations were calculated between age, dietary variables and BMI for each of the years surveyed. To investigate the possible effect of cigarette smoking as a confounder to the variables measured, the sample was divided into those who smoked in 1952, and those who did not smoke in 1952. Descriptive statistics were performed on these groups, and differences tested for with t-tests as above. Smokers and non-smokers were further subdivided into athletes and controls and descriptive statistics and tests of significance were repeated. Hxngihesis_tno. Body mass index, total energy intake, percent kilocalories from fat, total fat intake, smoking habits, participation in college athletics and levels of physical activity are not predictive of years of survival in college men. This hypothesis was tested by the Cox Proportional Hazards Regression Model (Steeland, Beaumont and Horning, 1986, Biomedical Computer Programs, 1982) using the mainframe computer at the University of Michigan, 41 Ann Arbor. This model presumes that death rates may be modelled as log-linear functions of covariates which explain differences in survival. These covariates may be fixed (such as sex), or time dependent (such as years of smoking). The model estimates regression coefficients that relate the effect of each covariable to the survival function. Step-wise regression using the Cox model was used to identify which covariables were significant predictors of years of survival. The Cox model is the regression model of choice when following a sample over time (Steeland, Beaumont and Horning, 1986). Unlike logistic regression, the Cox model allows the incorporation of variables that change over time. It also allows for the simultaneous adjustment for several confounders. These properties make the Cox model ideal when longitudinal data is available on a "hazard" (death), and time dependent covariables (diet, smoking, activity). The assumption inherent in the model is that of proportional hazards. That is, it is assumed that the relative risk of death remains constant over time. For exploratory purposes, the entire sample was subdivided into those subjects who lived throughout the entire length of the study, those who died between the 1967 and 1975 survey, and those who died between the 1975 and 1985 survey. Comparisons were made between the group who remained alive and the groups who died after each survey in order to determine if significant differences existed 42 between those subjects remaining alive and those dying. This was in contrast to the Cox model, in which average years of survival for the entire group was used as the dependent variable. RESULTS AND DISCUSSION. Results are presented by hypothesis. A discussion of the results follows each hypothesis. Under hypothesis one, smoking habits in 1952 and age are examined as potential confounders to the comparisons between former athletes and controls. In addition to those variables specified in the hypotheses, percent kilocalories from protein, carbohydrate and alcohol are presented for completeness. Similarly, height in meters for each of the years surveyed, weight in kilograms for each of the years surveyed and BMI at graduation for the subjects in each year surveyed are presented in addition to BMI at the time of each survey. 43 We We. Detailed results for the breakdown of the sample by year of survey and by former athlete and control are shown in Table 5. There were no significant differences between athletes and controls for any of the variables measured in 1967. Athletes were heavier than controls, but this was not statistically significant. Similarly the mean BMI of athletes at graduation was slightly but not significantly higher than controls. There were no significant differences between athletes and controls with respect to any of the variables measured in 1975. Athletes expended more kilocalories/week as aerobic activity than controls, but the difference was not statistically significant. The large standard deviations seen in aerobic activity are a result of the wide range of values reported; from 0 to 7839 kilocalories per week. In 1975 the athlete respondents were still slightly, but not significantly, heavier than controls. In the 1985 survey, athletes consumed significantly more kilocalories/day than controls. The slight difference in aerobic activity between former athletes and controls seen in 1975 was not seen in 1985. As in 1967 and 1975, former athletes were heavier than controls, but this difference was not statistically significant. 44 Table 5. Comparison of dietary variables, BMI and aerobic activity of athletes and controls in each year surveyed. YEAR OF QUESTIONNAIRE 1967 1975 1985 VARIABLE ATH CON ATH CON ATH CON N2211 N2125 N=147 N=66 N=80 N=35 Age/ Mean 62.8 65.9 68.5 69.8 75.1 77.1 yrs S.D 9.9 10.9 7.7 8.8 5.3 7.1 Kcals/ Mean 2149 2133 1908 1992 1850* 1618* day S.D 856 617 923 708 802 660 % Kcals Mean 15.9 15.6 16.0 14.8 16.7 14.8 protein S.D 4.4 4.2 4.7 3.6 5.5 4.2 % Kcals Mean 42.0 41.1 40.0 41.7 36.8 40.9 fat S.D 11.4 8.9 11.7 10.0 9.5 11.9 % Kcals Mean 40.0 41.6 41.3 40.3 44.0 42.3 carb S.D 13.4 11.5 12 9 11.2 12.4 14.1 % Kcals Mean 3.4 2.9 4.3 4.5 4.0 3.8 alcohol S.D 7.3 6.2 9.6 7.4 7.7 8.8 Fat Mean 98.4 98.9 83.8 91.6 73.7 68.6 g/day S.D 46.8 38.8 45.0 39.2 37.6 27.5 Activity Mean N.A N.A 1625 1327 1667 1600 kcals/wk S.D 1602 1589 1581 1636 Years Mean 75 4 76 6 76.9 77.6 75.0 76.6 survived S.D 9 8 10 4 7.4 8.3 5.5 7.1 Height Mean 1.78 1.76 1.78 1.77 1.79 1.77 (m) S.D 0.06 0.06 0.06 0.06 0.06 0.06 Weight Mean 82.9 77.6 83.8 78.5 82.7 77.5 (kg) S.D 11.1 8.6 11.0 8.1 11.6 11.3 BMI Mean 23.8 22.3 24.1 22.4 23.4 22.6 Graduation S.D 2 6 2.2 2.7 2.1 2.5 2.2 BMI Mean 26.0 25.0 26.1 25.1 25.7 24.8 S.D 2.7 2.6 2.7 2.3 2.7 3.5 * significant difference between values for athletes and controls, 1985. p<0.05 (t-test). The numbers of athletes and controls who were smokers in 1952 at each year surveyed are shown in Table 6. Categories of smoking are described in the methods section. The percentage of smokers and non-smokers in 1952 by athlete and control at each of the years surveyed is shown in Table 7. A higher percentage of former athletes than controls were smokers in 1952. For former athletes the proportions of respondents who were smokers in 1952 remained similar in each year surveyed, although there was a slight decline in the proportion of respondents who were smokers in 1952 surviving in 1985. For controls, there was a larger decrease than for former athletes in the proportion of subjects who were smokers in 1952 who survived to 1985. This would suggest that more subjects who were non-smokers in 1952 survived to 1985 than subjects who were smokers in 1952. However the difference in proportions of respondents who were smokers in 1952 in the years surveyed were not statistically significant. In order to illustrate the distribution of smokers in 1952 among the three categories of cigarette use, the percent of smokers in each category by athlete and control for each year surveyed is shown in Table 8. Numbers are percentages of all smokers in 1952. There was no significant difference between the percentage of respondents who were 46 Table 6. Categorization of respondents to each survey by smoking behavior in 1952. YEAR OF QUESTIONNAIRE 1967 1975 1985 £8681}; """ Air} """ 66:3 """ Iii: """ 66:3 '''''' ATH """ 661T" CATEGORYX 0 58 54 38 29 25 19 1 30 17 27 8 11 3 2 76 30 50 12 25 5 3 37 23 31 9 19 5 6312;} """"""""""""""""""""""""""""""" Smokers 143 70 108 29 55 13 63-25;?"-EBI"""I£2'""'IZE'"""3§ """" 56"""55‘" * 0 = Non-smoker, 1 = <20 cigarettes per day, 2 = 20-40 cigarettes per day, 3 = >40 cigarettes per day. ** Discrepancies between these totals and the total numbers of athletes and controls in each year are due to missing data on smoking habits. Table 7. Percentages of smokers and non-smokers in 1952 by athlete and control for each year surveyed. YEAR OF QUESTIONNAIRE 1967 1975 1985 SMOKING ATH CON ATH CON ATH CON CATEGORY % Smoker 71.1 56.5 74.0 50.0 68.7 40.6 % Non- 28.9 43.5 26.0 50.0 31.3 59.4 Smoker 47 smokers in 1952 in each of the three categories for athletes and controls for any of the years surveyed. However, it appears that there were more former athletes than controls who were moderate smokers in 1952 in all three years surveyed. Table 8. Percent of respondents to each survey who smoked in 1952 in each category of smoking behavior. YEAR OF QUESTIONNAIRE 1967 1975 1985 SMOKING ATH CON ATH CON ATH CON CATEGORY* n=201 n=124 n=146 n=58 n=80 n=32 % in 1 21.0 24.3 25.0 27.6 20.0 23.1 % in 2 53.1 42.9 46.3 41.4 45.5 38.5 % in 3 25.9 32.8 28.7 31 O 34 5 38 5 *1 = <20 cigarettes per day, 2 = 20-40 cigarettes per day, 3 = >40 cigarettes per day. WW1; Milli. In the examination of differences between former athletes and controls the confounding effect of age on the variables measured was assessed. Table 9 shows the Pearson product-moment correlations between age and dietary variables, physical activity and BMI for each year surveyed. 48 Table 9. Pearson product-moment correlations (r) of physical activity and BMI with age for each year surveyed. dietary variables, YEAR OF VARIABLE 1967 n=336 % Kcals .05 protein % Kcals -.16 fat % Kcals .20** carb % Kcals -.11 alcohol Kcals/ .07 day Fat -.01 s/day Activity N.A Kcals/wk BMI -.22** QUESTIONNAIRE 1975 1985 n=213 n=115 -.O4 -.07 .07 -.06 .08 .19 -.15 -.17 -.10 -.10 .07 -.15 -.30** .01 -.15 -.37** * N/A: not available. ** significant (p<.05). 49 There were no moderate or strong correlations between age and any of the variables measured in 1967, 1975 or 1985. There was a weak negative correlation (r = -.30) between age and aerobic activity in 1975, and between age and BMI in 1985 (r = -.37). WW Wu. Comparison of all variables measured in the total sample by smoker in 1952 and non-smoker in 1952 are shown in Table 10. No significant differences were seen between smokers and non-smokers for any of the variables measured in any of the years surveyed, with the exception of aerobic activity in 1975. Those who were smokers in 1952 expended significantly more kilocalories per week as aerobic activity in 1975 than those who were non-smokers in 1952. Participation in college athletics was then included as a variable in the analysis of the difference in aerobic activity between those respondents who smoked in 1952 and those respondents who did not smoke in 1952. The results are shown in Table 11. When former athletes were considered, there were no significant differences with respect to any of the variables measured between those who smoked in 1952 and those who did not. When controls were considered, there was a significant difference between those who smoked in 1952 and those who did not. Those who smoked in 1952 were more aerobically active than non- 50 Table 10. Comparison of dietary variables, BMI and aerobic activity of those who smoked in 1952 and those who did not smoke in 1952 by year of questionnaire. YEAR OF QUESTIONNAIRE 1967 1975 1985 VARIABLE 5* NON S NON S NON N=223 N=112 N=145 N=68 N=70 N=45 Age/ Mean 63.2 65.4 68.9 69.1 74.9 77.1 yrs S.D 9.8 11.3 8.2 7.8 5.5 6.4 Kcals/ Mean 2164 2102 1879 2051 1729 1857 day S.D 817 685 743 1068 802 710 % Kcals Mean 15.9 15.5 15.7 15.6 16.5 15.4 protein S.D 4.5 4.1 4.5 4.4 5.3 5.0 % Kcals Mean 39.1 41.3 39.9 42.0 38.5 37.3 fat S.D 13.3 9.4 11.2 11.3 10.1 11.0 % Kcals Mean 39.1 43.6 40.4 42.3 41.2 47.1 carb S.D 13.3 10.9 12.6 11.9 12.1 13.4 % Kcals Mean 4.3 1.0 5.5 1.8 5.1 2.1 alcohol S.D 7.9 3.2 10.2 4.7 9.5 4.5 Fat Mean 99.9 96.1 83.8 91.3 71.8 72.8 g/day S.D 47.0 37.1 40.0 49.8 38.4 28.6 Activity Mean N.A N.A 1629** 1339** 1716 1541 kcals/wk S.D 1702 1367 1746 1345 Years Mean 75.0 77.5 76.7 77.8 74.6 77.0 survived S.D 10.3 9.2 7.8 7.3 5.8 6.3 Height Mean 1.78 1.76 1.79 1.76 1.79 1.78 (m) S.D 0.06 0.06 0.06 0.06 0.05 0.06 Weight Mean 81.7 78.9 82.9 80.7 82.8 78.4 (kg) S.D 10.3 10.7 10.1 11.1 9.1 14.4 BMI Mean 23.5 22.8 23.9 22.9 23.7 22.4 Graduation S.D 2.6 2.6 2.6 2.6 2.4 2.3 BMI Mean 25.7 25.5 25.9 25.5 25.9 24.7 S.D 2.6 3.0 2.5 2.9 2.5 3.5 * S=smoker in 1952, Non=non-smoker in 1952. ** significant difference between values for smokers and non- smokers, 1975, p<0.05 (t-test). 51 Table 11. Comparison of aerobic activity (Kcals/week) in 1975 between athletes, controls, those who smoked in 1952 and those who did not smoke in 1952. ATHLETE CONTROL SIG Smoker 1651 1563 n s 1952 11697 11744 (n=108) (n=29) Non- 1557 1074 p< 05 Smoker 11327 11391 1952 (n=38) (n=29) SIG n.s p< 05 smokers. These differences were not apparent in 1985. Athletes who were smokers in 1952, athletes who were not smokers in 1952 and controls who were smokers in 1952 all expended similar amounts of energy per week in aerobic activity in 1975. Only controls who were not smokers in 1952 had a significantly lower aerobic activity level in 1975. There were no significant differences between former athletes and controls for any of the variables measured in 1967 or 1975. In 1985, former athletes consumed significantly more kilocalories per day than controls. More former athletes than controls were smokers in 1952. There were no differences between those who smoked in 1952 and those who did not smoke in 1952 with respect to any of the dietary variables measured or body 52 mass index for any of the years surveyed. Although the mean age of the sample increased with each subsequent survey, there were no moderate or strong correlations for age with any of the dietary variables measured, body mass index, or aerobic activity. W. The finding that in 1985 former athletes consumed significantly more kilocalories than controls should be interpreted with care considering the large standard deviations associated with this variable and the small number of controls (35) in the 1985 survey. A difference of 232 kilocalories, although statistically significant, probably has little practical significance for health and longevity. The composition of the diet with respect to protein, fat, carbohydrate and alcohol was the same in the two groups. The difference in energy intake is therefore the result of eating less of the same diet rather than of eating less of one particular macro-nutrient such as fat. The finding that BMI at graduation was higher in former athletes than controls is in agreement with the findings of Montoye, Van Huss, Olson, Peirson and Hudec (1957), that athletes were heavier than non-athletes in 1952. Although it might be expected that taller and heavier men would be recruited for college sports (particularly football), there was no statistically 53 significant difference in the heights or weights of former athletes and controls throughout the survey. Despite the recommendation that BMI be adopted as the adiposity index of choice in reporting results, there are limitations to the interpretation of BMI, particularly when no other anthropometric measurements are available. Although highly correlated with body fat, BMI is not independent of stature. Garn, Leonard and Hawthorne (1986) concluded that BMI as an index of adiposity is confounded by relative sitting height (sitting height/standing height). That is, the correlation of BMI with body fat is reliant on particular proportions of upper and lower body lengths. Although the majority of the population would be expected to fall within a narrow range of relative sitting height, those with unusually short or long legs in relation to their total height would have a lower correlation of BMI to adiposity. Athletes are one group of individuals who are often selected on the basis of physical attributes such as “long legs". Although it is impossible to quantify relative sitting height from the information collected in this study, it remains possibile that in some athletes, BMI was confounded by sitting height, and was not as accurate a measure of adiposity as in controls. Wilson (1986) found that former athletes in the Michigan State study were significantly more mesomorphic and less ectomorphic than non-athletes (p<.05). Furthermore, Wilson found that somatotype was a 54 significant predictor of life expectancy (p =.001), with endomorphs surviving fewer years than mesomorphs and ectomorphs. However, differences in somatotype between former athletes and controls cannot be revealed by comparison of BMI alone. The mean aerobic activity level of former athletes was higher than that of controls, but there was a large variance in energy expended as aerobic activity in both groups and the difference between them was not significant. Quinn (1987) also found in a sample of former athletes and controls drawn from the respondents used from the Michigan State study, higher aerobic leisure time aerobic energy expenditure in former athletes compared to controls. In his sample, the difference was significant (p<.05). Increased activity in former athletes might reflect the continuation of patterns of aerobic activity established in school. The historical recall of activity required in the questionnaire might result in a more accurate record for those with an established and regular pattern of exercise as opposed to those who exercise less consistently or in many different ways. There is the possibility that former athletes are more aware of aerobic activity levels and are therefore more accurate and more thorough in recording them. No consideration of non-aerobic leisure time pursuits was made in this study. Although these activities were not of sufficient length or duration to be aerobic, 55 they do require energy. Also, leisure time activity alone was measured. In manual jobs, vocational activity might be the major component of energy expenditure. However it is likely that most of the subjects in the present study were either retired or in sedentary jobs. There is no reason to believe that vocational activity is not normally distributed in the sample, and would therefore contribute equally to the energy expenditure of former athletes and controls. The lack of consideration of vocational activity, and the contribution of non-aerobic activity to energy balance make it unlikely that energy balance is described completely by the measurements in this study. In the 1985 survey the difference between the aerobic activity level of former athletes and controls is smaller than that seen in the 1975 survey. It appears that surviving controls had aerobic activity levels comparable to those of former athletes, although it should be emphasized that the difference in 1975 was not statistically significant. If a difference in the activity levels of former athletes and controls does exist, it is possible that controls increased their aerobic activity level in response to the increased public interest in fitness and exercise which started in the mid 19703. If this were the case, the figures suggest that former athletes were already exercising aerobically before the ”fitness boom“, and controls started to exercise after 1975, resulting in no difference between the groups by 56 1985. It would be interesting to analyze the type of activities undertaken to determine if the increase seen in the control group was accounted for entirely by an increase in calisthenics and jogging. Alternatively (or additionally), the increased public awareness of the importance of exercise may have made the controls more aware of their activity patterns in and more accurate in completing their questionnaires in 1985 than 1975. It is unfortunate that activity levels have not been calculated for the 1967 survey. They would be useful in determining whether the increased activity in controls occured soley in the mid-seventies, or if differences were apparent at an earlier date. There was no difference in average years of survival between former athletes and controls in this study after 35 years of follow-up. This confirms the findings of Montoye, Van Huss and Nevai (1962) after seven years of follow—up of the same sample, and of Wilson (1986) that participation in college athletics was not a significant predictor of life expectancy. Paffenbarger (1986) also found no relationship between survival and participation in college athletics. In all surveys, the average number of years survived by both former athletes and controls were greater than the life expectancy for white males in the U.S. and in Michigan in 1984. The difference between this sample and the population of Michigan is particularly significant, 57 because it eliminates regional differences in life expectancy as a confounder to the comparison with the sample in the present study. One possible explanation for the longer average survival in the present sample when compared to Michigan statistics is that Michigan statistics include data from metropolitan areas such as Detroit. The mean life expectancy in Detroit is lower than that in the rest of the State (Michigan Department of Public Health, 1987). The sample in the present study was predominantly rural. Also, there is evidence to suggest that college graduates have a higher life expectancy than non-graduates (Greenaway and Hiscock, 1926). It is not clear why this should be so, although a high level of education might result in increased knowledge of health risks and their _ avoidance. Perhaps college graduates attain a higher socio- economic status than non-graduates. Life expectancy increases with increasing income (U.S. Department of Commerce Bureau of the Census, 1978), and socio-economic status (Haan, Kaplan and Camocho, 1987). The lack of correlation of age with any of the variables measured implies that age is not a confounder to any of the relationships between them. This does not eliminate the possibility of increased innaccuracy of reporting dietary data with increasing age. It would be necessary to conduct detailed weighed intakes on a sub- population of the subjects in order to quantify the discrepancies between reported and actual intake and relate these discrepancies to age. The reasons why approximately twice as many former athletes as controls were smokers in 1952 are not clear. The reasons for this are unclear. Montoye, Van Huss, Olson, Peirson and Hudec (1957) suggested that more former athletes smoked as a reaction against the restrictions imposed by training regimes in college. More former athletes were moderate smokers than were controls, and more controls were heavy smokers than were former athletes. There was, however, no significant difference in the level of cigarette use between those former athletes and controls who were smokers. The high percentage of smokers in both former athletes and controls in 1952 might reflect the increased social acceptability of smoking at this time, and the lack of evidence for the associated health hazards. Although Albanes, Jones, Miccizzi and Mattson, 1987) found that smokers were leaner than non-smokers in the NHANES II, no difference in BMI was found between smokers in 1952 and non-smokers in 1952 in this study. The interaction between aerobic activity and smoking in 1952 found in this study might be responsible, however the finding is difficult to explain. Perhaps the non-smoking controls did not feel exercise was necessary as they did not consider themselves at risk. An interesting finding that arose from the comparison of the smoking habits of former athletes and 59 their controls in 1952 was that the percentage of subjects who were smokers in 1952 who survived to 1985 decreased more in the control group than in former athletes. This suggests that in the control group, more of those who were smokers in 1952 than non-smokers in 1952 died. This difference was not apparent in the former athletes. Apart from smoking habits in 1952, participation in athletics in college and level of aerobic activity were the only variables that distinguished former athletes and controls. If more smokers in 1952 in the control group than smokers in 1952 in the former athlete group had died before 1985, athletic participation in college and/or higher levels of aerobic activity might later have been protective against the deleterious effects of smoking. It is possible that any survival advantage gained by increased aerobic activity may be negated by the deleterious effects of smoking, resulting in no significant difference in years of survival between former athletes and controls. However this cannot be concluded without further information on smoking habits prior and subsequent to 1952. Also, the change in proportions of smokers in 1952 and non-smokers in 1952 over the three surveys was not statistically significant. 60 Wanna Wail. Stepwise regression using the Cox model demonstrated that in the total sample aerobic activity was the only significant predictor of average years of survival of the variables measured. The variables used in the Cox model, together with the significance levels for prediction of average years of survival are shown in Table 12. As a Table 12. Significance levels of dietary variables, BMI and participation in college athletics for prediction of average years of survival from stepwise regression using the Cox Proportional Hazards regression model. VARIABLE P-VALUE RSQIQEQ; """""" 6158"” % Kcals protein 0.940 % Kcals fat 0.531 % Kcals carb 0.677 Fat g/day 0.373 Alcohol Y/N 0.748 BMI 0.360 Athlete Y/N 0.059 Activity* 0.025 Smoker Y/N 0.551 * significant at p<.05 61 result of the stepwise procedure, the significance level reported for all variables other than aerobic activity reflect an adjustment to control for differences in aerobic activity between the subjects. Wearing W. Descriptive statistics were used for exploratory purposes, to determine if differences existed in any of the variables measured between those subjects who remained alive throughout the three surveys (survivors), those who died between the 1967 and 1975 surveys, and those who died between the 1975 and 1985 surveys. This analysis attempted to describe the dichotomy survived/died rather than the continuous variable years of survival which was used in the Cox model. Numbers and percentages of those surviving and those dying or lost to follow-up between the surveys are shown in Table 13. The difference in sample size of the survivors between 1967 and 1975 is accounted for by six subjects who did not return 1975 questionnaires, but subsequently returned 1985 questionnaires. There were no statistically significant differences with respect to any of the variables measured between those who died between 1967 and 1975 and those who remained alive throughout all surveys. Similarly, there were no significant differences with respect to any of the variables measured between those 62 who died between 1975 and 1985 and who lived throughout the surveys. Those subjects who remained alive throughout the three surveys were younger than those who died. However these differences were not statistically significant. Table 13. Numbers and percentages of subjects surviving to 1985, dying between surveys and lost to follow-up by former athlete and control. YEAR OF QUESTIONNAIRE 1967 1975 ATHLETE CONTROL ATHLETE CONTROL n=211 n=125 n=147 n=66 Survived to 75 33 71 35 1985 35.5% 26.4% 48.3% 53.0% Died before 50 44 59 24 next survey 23.7% 35.2% 40.1% 36.4% Lost to 6 13 10 8 follow-up 2.8% 10.4% 6.8% 12.1% We. In the total sample there were no differences with respect to any of the variables measured between those subjects who survived and those who died between each of the surveys. These results are shown in Table 14. Participation in college athletics was then included as a variable in the comparison of those who survived and those who died. Comparisons of those who survived by former 63 athlete and control are shown in Table 15. Comparisons of those who died between the surveys by former athlete and control are shown in Table 16. It was found that former athletes who survived had a significantly higher leisure time aerobic activity level than controls who survived. When those controls who survived were compared to those controls who died between 1975 and 1985, significant differences in total kilocalories and aerobic activity were seen. Controls who survived had significantly higher aerobic activity levels than those who died between the 1975 and 1985 surveys (Table 17). Controls who survived had a significantly higher kilocalorie intake than those who died between 1975 and 1985 (Table 18). The Cox Proportional Hazards Regression model indicated that aerobic activity was the only significant predictor of years of survival. Descriptive statistics showed no significant differences with respect to any of the variables measured between those who survived and those who died between the surveys. Significant differences were seen between controls who lived throughout the three surveys, and controls who died between the 1975 and 1985 surveys. Controls who survived had a significantly higher energy expenditure as aerobic activity, and a significantly higher energy intake than did controls who subsequently died. This pattern was not seen in former athletes. 64 Table 14. Comparison of dietary variables, BMI and aerobic activity of subjects dead after each survey and survivors who responded to each survey. YEAR OF QUESTIONNAIRE 1967 1975 VARIABLE DIED* LIVED** DIED LIVED N294 N=110 N283 N=104 Age/ Mean 70.8 57 7 72 6 65.9 yrs S D 11.5 5 9 8 5 5.6 Kcals/ Mean 2125 2142 1868 2021 day S D 585 915 913 855 % Kcals Mean 15.7 16.0 15.8 15.3 protein S.D 4.2 5.1 4.5 4.6 % Kcals Mean 42.0 42.7 41.4 39.6 fat S.D 9.0 11.5 11.6 10.6 % Kcals Mean 40.4 38.3 40.7 41.4 carb S.D 12.5 12.4 13.9 11.8 % Kcals Mean 3.1 4.2 3.4 5.3 alcohol S.D 6.2 8.3 8.0 9.9 Fat Mean 98.7 99.6 84.7 88.8 S/day S.D 35.1 50.3 44.9 44.8 Activity Mean N.A N.A 1384 1624 kcals/wk S.D 1735 1542 Years Mean 74.3 74.5 79.4 75.7 survived S.D 13.1 6.1 8.4 5.9 Height Mean 1.76 1.78 1.78 1.78 (m) S.D 0.06 0.06 0.06 0.06 Weight Mean 79.1 82.1 81.8 82.2 (kg) S.D 10.1 10.0 9.6 10.1 BMI Mean 23.0 23.2 23.9 23.2 Graduation S.D 2.3 2.5 2.8 2.5 BMI Mean 25.6 25.7 25.8 25.8 S.D 2.8 2.6 2.5 2.6 * Died before the subsequent questionnaire was issued. ** Lived throughout all three surveys. No differences were statistically significant. 65 Table 15. Dietary variables, BMI and aerobic activity of survivors compared by athletes and controls. YEAR OF QUESTIONNAIRE 1967 1975 VARIABLE ATHLETE CONTROL ATHLETE CONTROL N275 N235 N271 N235 Age/ Mean 57.0 59 1 65 5 68 8 yrs 5 D 5.1 7 2 5 2 7 1 Kcals/ Mean 2165 2093 1930 2218 day S D 1039 573 937 612 % Kcals Mean 16.1 15.9 16.0 14.2 protein S.D 5.3 4.8 5.0 3.2 % Kcals Mean 42.7 42.5 39.0 40.9 fat S.D 12.5 9.3 11.4 8.6 % Kcals Mean 37.8 39.5 40.6 42.9 carb S.D 13.3 10.6 12.5 10.1 % Kcals Mean 4.7 3.2 6.0 3.6 alcohol S.D 9.1 6.3 11.2 5.9 Fat Mean 99.6 99.7 83.2 101.1 B/day S.D 56.0 36 0 46.8 38.1 Activity Mean N.A N.A 1727* 1414* kcals/wk S.D 1609 1396 Years Mean 74.8 76.8 75.4 76.5 survived S.D 5.4 7.1 5.34 7.0 Height Mean 1.79 1.77 1.79 1.77 (m) S.D 0.05 0.06 0.06 0.06 Weight Mean 83.5 79.4 83.7 79.1 (kg) S.D 10.5 8.2 10.7 8.1 BMI Mean 23.6 22.6 23.6 22.5 Graduation S.D 2.5 2.2 2.5 2.2 BMI Mean 25.9 25.4 26.0 25.3 S.D 2.5 2.6 2.6 2.7 * significant difference between athlete and control p<.05 (t- test). 66 Table 16. Dietary variables, BMI and aerobic activity of subjects dead before the subsequent survey compared by athletes and controls. YEAR OF QUESTIONNAIRE 1967 1975 VARIABLE ATHLETE CONTROL ATHLETE CONTROL N250 N244 N259 N224 Age/ Mean 70.3 71.4 72.2 73.5 yrs S.D 11.4 11.6 8.5 8.5 Kcals/ Mean 2154 2093 1874 1854 day S.D 630 534 939 864 % Kcals Mean 15.4 15.9 16.0 15.3 protein S.D 3.7 4.8 4.7 3.8 % Kcals Mean 43.8 40.0 40.3 43.9 fat S.D 9.1 8.6 11.9 10.5 % Kcals Mean 38.8 42.1 42.9 35.5 carb S.D 12.8 12.0 14.2 12.0 % Kcals Mean 3.0 3.3 2.2 6.1 alcohol S.D 5.7 6.8 7.1 9.2 Fat Mean 103.2 93.7 83.3 88.2 g/day S.D 35.0 34.9 44.5 46.5 Activity Mean N.A N.A 1405 1331 kcals/wk S.D 1596 2089 Years Mean 72 9 75 9 79.2 80.0 survived S.D 13 9 12 2 8.5 8.3 Height Mean 1.77 1.75 1.78 1.78 (m) S.D 0.07 0.06 0.06 0.06 Weight Mean 80.8 77.1 83.1 78.8 (kg) S.D 11.6 7.7 9.9 8.1 BMI Mean 23.4 22.6 24.6 22.3 Graduation S.D 2.3 2.4 2.8 2.2 BMI Mean 25.8 25.3 26.1 25.0 S.D 3.2 2.4 2.6 2.0 No differences were statistically significant. 67 Table 17. Comparison of aerobic activity (Kcals/week) between athletes, controls, survivors and those who died between 1975 and 1985. ATHLETE CONTROL SIG Survived 1727 1414 p<.05* past 1985 £1609 £1396 (n271) (n235) Died 1975 1405 1331 n.s -1985 11596 12089 (n259) (n224) Sig n.s n 8 Table 18. Comparison of energy intake (Kcals/day) between athletes, controls, survivors and those who died between 1975 and 1985. ATHLETE CONTROL SIG Survived 1930 2218 n.s past 1985 1937 1612 (n271) (n235) Died 1975 1874 1854 n.s -1985 1939 i864 (n259) (n224) Sig n.s p< 05* * t-test. 68 The relationship between aerobic activity and survival demonstrated in this study applies to the last ten years of life. This was at a time when the mean age of the sample was 68 years. The finding that aerobic activity is lower in those who have the lowest survival is not unexpected in this timeframe. As the subjects aged and died, they became less aerobically active. If conclusions are to be made about the influence of lifelong aerobic activity on survival it would be desirable to have data from a longer timespan than has been analyzed here. Analysis of the 1967 activity data would increase the time span over which information relating activity and mortality is available. In view of the higher aerobic activity levels of former athletes, and the significant relationship between activity and survival, it might be expected that participation in college athletics would be related to survival. Although the higher aerobic activity of controls than former athletes was not statistically significant in this study, Quinn (1987) showed a statistically significant difference between the aerobic activity of former athletes and controls taken from the same sample as that used in this study. Former athletes expended more energy per week as aerobic activity than controls. It seems that participation in college activity results in higher activity levels after college, and this in turn is related 69 to survival. The observation by Paffenbarger (1984, 1986) that post-graduate level of activity rather than college participation in athletics is important in determining survival was not completely supported by the results of this study. Post-graduate activity levels were significant in determining survival, but appear also to be related to participation in college athletics. In this study none of the dietary variables measured were significantly related to survival. This is contrary to the findings of Schlenker (1976) that total fat predicted longevity, but would support the findings of Keys (1981) that total fat and energy intake did not predict mortality. The lack of a relationship between BMI and survival in this study supports the findings of the Seven Countries study (Keys et al. 1981). However the limitations of the data collection in the present study and the small sample size in comparison to the Seven Countries Study should be considered in comparison of these studies. It appears that there was not a differential death rate in those with a very high or very low BMI, because there were no significant differences between the BMIs of those who survived throughout the three surveys and the BMIs of those who died between the surveys. It is possible that more subjects with both very high and very low BMIs died between the surveys than lived throughout. This would not be reflected in mean values. Equivalent 70 mean BMIs could result from the averaging of extreme values, and the averaging of values in the middle of the range. If this were the case, and those with very high and very low BMIs were dying, the standard deviation for this group would be higher than for the group who survived. This was not the case. There was no differential death rate for those at the ends of the range for BMI. It is possible that many of the subjects who died between the surveys (that is, less than ten years after being surveyed), were chronically ill at the time they completed their last questionnaire. Those who lived throughout the time period, by virtue of their survival, appear to be in better health. This assumption is stronger if the comparison is made between those who died between 1967 and 1975 and those who survived beyond 1985. That is, when those who died are compared to those who lived at least ten years longer. It is possible those who died were divided into those who were chronically ill at the time of completing the questionnare, and those who were well but subsequently became ill and died, significant differences between the groups would be unmasked. Lower energy intake and expenditure as aerobic activity was characteristic of the controls who died between 1975 and 1985, but not of former athletes who died in the same time. If low activity and caloric intake is associated with chronic illness, perhaps more controls than athletes died from chronic disease. However it is not 71 possible to conclude this with the data available. The standard deviations associated with aerobic activity are very large and the numbers in each group relatively small. It is possible that many of those who survived to 1985 died or will die shortly after the survey. They too may have been chronically ill in 1985. The comparison of those who died between 1975 and 1985 with those who were alive in 1985 might, in retrospect, be a comparison of two groups who died within a short time of each other. This cannot be resolved until all subjects are followed to death. SUMMARY AND IMPLICATIONS This study demonstrated that the only difference between former athletes and controls in this sample with respect to the variables measured was in smoking habits. Former athletes were more likely to smoke than controls. No other differences were either practically or statistically significant. Although not statistically significant in this study, former athletes had higher aerobic activity levels than controls, and as discussed previously, this difference has been found to be significant in a larger sample (Quinn, 1987). For the total sample, the Cox model demonstrated that leisure time aerobic activity was a significant 72 predictor of years of survival. Subjects expending more energy as aerobic activity survived longer than those with lower levels of aerobic activity. Level of aerobic activity also differentiated subjects who survived beyond 1985 and those who died between 1975 and 1985. The dietary variables measured in this study did not predict years of survival, or differentiate subjects who lived from those who died. Although participation in college athletics was not directly related to survival, there was an indirect link. Health behaviors established at an early age seemed to be carried into middle and old age. Former college athletes tended to be more aerobically active than controls, and increased aerobic activity was related to survival. In this study it appeared that diet and lifestyle variables in the last ten years of life were poorly related to survival. Survival benefits obtained from changing health behaviors therefore accrued over the lifetime of the subjects. It is difficult to conclude from this study that diet and BMI do not influence survival. The relationship of lifestyle to survival is complex. This study might not have demonstrated relationships that exist in the sample, or have examined the optimal set of variables to predict survival. The relative importance of each variable in determining survival might vary from subject to subject, or within one subject with age. The standard deviation 73 associated with many of the variables was large, and the sample although large for experimental purposes, was small for an epidemiological study. Also, data on diet was collected only in the later years of life. If the relationship of lifestyle to survival is to be fully characterized, longitudinal studies that monitor a wide range of variables over long periods in large numbers of subjects will be necessary. Such studies are expensive and do not produce results for many years. However this is the only way in which definitive conclusions on the effects of lifestyle on morbidity and mortality in humans can be reached. Steps should be taken to ensure that this research is supported. LIST OF REFERENCES LIST OF REFERENCES Albanes D., Jones Y., Micizzi M., Mattson M. Associations between smoking and body weight in the US population. Aug Jiznhlig_flealth; 77: 4, 439-444, 1987. American Cancer Society. Cancer facts and figures. American Cancer Society. New York 1985. Andres R., Elahi D., Tobin J.D., Miller D.C., Brant L. Impact of age on weight goals. Annilntfiued. 103:1030- 1033,1985. Anonymous. Body weight, health and longevity: conclusions and recommendations of the workshop. Nutifiey. 43:61- 63,1985. Armstrong 3., Doll R. Environmental factors and cancer incidence and mortality in different countries, with special reference to dietary practices. Iniiliflnncnn. 15:617-631,1975. Bannister, W.E, Brown S.R. The relative energy requirements of physical activity. In H.B.Falls (Ed.) Exercise Physiology. New York: Academic Press, 1968. Beaton G.H., Milner J., Corey P., McGuire V., Cousins M., Stewart E., Ramos M de., Hewitt D., Grambsch P.V., Kassim N., Little J.A. Sources of variance in 24-hour dietary recall data: implications for nutrition study design and interpretation. AnilifllinLNut. 32:2546-2559,1979. Biomedical computer programs. Berkeley, California. University of California Press 1982. Build Study 1979. Society of Actuaries and Association of Life Insurance Medical Directors. Chicago 1980. 74 75 Committee on Diet, Nutrition and Cancer. Diet, Nutrition and Cancer. Assembley of Life Sciences, National Research Council, National Academy Press, Washington D.C., 1982 Dyer A. R. , Stamler J. Berkson D. M. Lindberg H. A. Relationship of relative weight and body mass to 14- -year mortality in the Chicago Peoples Gas Company. Jgflhrgninis. 28: 109-123,1975. Feinleib M. Epidemiology of obesity in relation to health hazards. Annglntined, 103:1019-1024,1985. Folsom A.R., Jacobs D.R., Caspersen C.J., Gomez-Marin 0., Knudsen J. Test-retest reliability of the Minnesota leisure time physical activity questionnaire. liflhrgngnis. 39:505-511,1986. Frisch R.E., Wyshak G., Albright N.L., Albright T.E., Schiff I., Jones K.P., Kitschi J., Shiang E., Koff E., Marguglio M. Lower prevalence of breast cancer and cancers of the reproductive system among former college athletes compared to non-athletes. Brgiiflangen. 52:885-891,1985. Garn S.M., Leonard W.R., Hawthorne V.M. Three limitations of the body mass index. Amilgflliniflut. 44:996-997,1986. Garrison R.J., Feinleib M., Castelli W.P., McNamara P.M. Cigarette smoking as a confounder of the relationship between relative weight and long-term mortality. The Framingham heart study. liAgugA 249:2199-2203,1983. Greenway J.C., Hiscock I.V. Mortality among Yale men. Editorial. liAiMiA 87:175,1926. Haan M., Kaplan G.A., Camacho T. Poverty and health: prospective evidence from the Alameda county study. AniliEnid. 125:989-998,1987. Heilbrun L.K., Hankin J.H., Nomura A.M., Stemmerman G.N. Colon cancer and dietary fat, phosphorous, and calcium in Hawaiian-Japanese men. AmIJAQlingflut, 43 306-309,1986. Howley E. T. Glover M. H. The caloric cost of running ond walking one mile for men and women. Mediand_522rlfi_figi. 6: 235, 1974. 76 Iverso G.R., Norpoth H. Analysis of variance. Sage University Paper series 1, Quantitative Applications in the Social Sciences. Sage Publications. 1982. Jarrett R.J., Shipley M.J., Rose G. Weight and mortality in the Whitehall study. BLMLJ 285:535-537, 1982. Kahn H.A., Phillips R.L., Snowdon D.A., Choi W. Association between reported diet and all-cause mortality. Twenty-one year follow-up on 27,530 adult seventh day adventists. Amilngid. 119, 775-787,1984. Keys A. Seven Countries: A multivariate analysis of death and coronary heart disease. Cambridge: Harvard University Press, 1980. Keys A., Aravanis C., van Buchem F.S.P., Blackburn H., Buzina R., Djordjevic B.S., Dontas A.S., Fidanza A.S., Karvonen M.J., Kimura N., Menotti A., Nedelijokovic S., Puddu V., Punsar S., Taylor H.L. The diet and all-cause death rate in the Seven Countries Study. Lancet, 2, 58-61, 1981. Keys A., Fidanzo F., Karvonen M.J., Kimura N., Taylor H.L. Indicies of relative weight and obesity. lgflhzgninis. 25:329-343,1972 Kushi L.H., Lew R.A., Stare F.J., Ellison C.R., El Lozy M., Bourke G., Daly L., Graham I., Hickey N., Mulcahy R., Kevany J. Diet and 20-year mortality from coronary heart disease. The Ireland-Boston Heart Study. Neg_Engl+J*Med. 312,811-818,1985. Leveille G., Zabik M.E., Morgan K. Nutrients in foods. Cambridge Massachusetts: The Nutrition Guild, 1983. Leon A.S., Jacobs D.R., DeBacker G. Relationship of physical characteristics and life habits to treadmill exercise capacity. Amplifipidemigl.113:653-660,1981. Lew E.A., Garfinkel L. Variations in mortality by weight among 750,000 men and women. J+Chrgninis. 32:563-576, 1979. 77 Limited food intake and longevity. Nutiflex. 40:314- 316.1982. Madden J.P., Goodman S.J., Guthrie H.A. Validity of the 24- hour recall. JiAminiet+Ass. 73:48-55,1978. Michigan Department of Public Health. Personal communication, Statistical Services, Department of Vital Statistics. July 1987. Montoye H.J., Van Huss W.D., Nevai J.W. Longevity and morbidity of college athletes: A seven-year follow-up study. WW. 2:133.1962. Montoye H.J., Van Huss W.D., Olson H.W., Peirson W.R., Hudec A.J. The longevity and morbidity of college athletes. Phi Epsilon Kappa Fraternity. 1957. Morrison S.D. Nutrition and longevity. Nutiflex. 41:133- 142,1983. National Center for Health Statistics. Body mass index for males 18-74 years of age by race and age, 1976-1980. Unpublished figures, 1983. National Center for Health Statistics. Carroll M.D., Abraham 8., Dresser C.M. Dietary intake source data: United States 1976-80. 1ital_and_Healih_Siaiistics Series II No. 231. DHHS Pubn. No. (PHS) 83-1681. Public Health Service, Washington. U.S. Government Printing Office, March 1983. National Center for Health Statistics. Health, United States, 1985. DHHS Pub. No. (PHS) 86-1232 Public Health Service. Washington D.C. US Government Printing Office, 1985. National Center for Health Statistics. Monthly vital statistics report. Volume 1, April 1987. Nomura A., Heibrun L.H., Stemmerman G.N. Body mass index. A good predictor of cancer in men. JLNatigangez_lnst. 74(2):319-323,1985. 78 Norgan N. G. Fero- Luzzi A. Weight- height indicies as estimators of fatness in men. Humn+flut_flliniflut 36: 363- 372.1982. Nube M., Kok F.J., Vandenbroucke J. P. , van der Heide- Wessel C., van der Heide R. M. Scoring of prudent dietary habits and its relation to 25- -year survival. JLAm+Diet+Afis. 87, 2. 171-175,1987. Olson H.W., Montoye H.J., Sprague H., Stephens K.E., Van Huss W.D. The longevity and morbidity of college athletes. WM. 6:62.1978. Paffenbarger R.S., Hyde R.T., Wing A.L., Hsieh C.C. Physical activity, all-cause mortality and longevity of college alumni. NgEnglilined. 314:605-613,1986. Paffenbarger R.S., Hyde R.T., Wing A.L., Steinmetz C.H. A natural history of athleticism and cardiovascular health. liAiMiA 252:491-495,1984. Palta M., Prineas R.J., Berman R., Hannan P. Comparison of self-reported and measured height and weight. Amil+Epid. 115:223-230,1982. Quinn T.J., The relationship between caloric expenditure and longevity among Michigan State University athletes and non-athletes. Unpublished Ph.D Dissertation, School of Health Education, Counselling Psychology and Human Performance, Michigan State University, 1987. Rabkin S.W., Matthewson F.A.L., Hsu P.H. Relation of body weight to development of ischemic heart disease in a cohort of young North American men after a 26-year observation period: The Manitoba study. Angliflandigl. 39:452-458, 1978. Rook A. An investigation into the longevity of Cambridge sportsmen. Egnil 1:773-777,1954. Schlenker E.D. Nutritional status of older women. Unpublished Ph.D Dissertation, Department of Food Science and Human Nutrition, Michigan State University, 1978. 79 Shekelle R.B., Shryock A.M., Paul 0., Lepper M., Stamler J., Liu S., Raynor W.J. Diet, serum cholesterol and death from coronary heart disease: The Western Electric Study. Nifinzlillnfid. 304:65-70,1981. Sidney S., Farquhar J.W. Cholesterol, cancer, and public health policy. Amilgned. 75:494-508,1983. Society of Actuaries. Build and Blood Pressure Study. 1959 Chicago: Society of Actuaries 1959. Sorlie P., Gordon T., Kannel W.B. Body build and mortality. The Framingham study. 1*A*M*A_243:1828- 1831,1980 Steeland K., Beaumont J., Horning R. The use of regression analysis in a cohort mortality study of welders. J+thgninia. 39:287-294,1986. Stewart A.W., Jackson R.T., Ford M.A., Beaglehole R. Underestimation of relative weight by use of self-reported height and weight. Anilifinid. 125,1,122-126,1987. Stunkard A.J., Albaum J.M. The accuracy of self-reported weights. AW. 34,1593-1599, 1981. Taylor H.L., Jacobs D.R., Schucker B. A questionnaire for the assessment of leisure time physical activities. We. 31:741-755,1978. U.S. Department of Agriculture. Nutrient intakes: Individuals in 48 states, year 1977-78. Nationwide Food Consumption Survey 1977-78. Report No. I-2, Consumer Nutrition Division, Human Nutrition Information Service, U.S.D.A. U.S. Government Printing Office, August 1983. U.S. Department of Commerce Bureau of the Census. Historical statistics of the United States. Colonial times to 1970. Part 1. 93rd Congress lst Session. House document No. 93-78, 1978. Vandenbrouke J.P., Mauritz B.J., deBruin A., Verheesen J.H.H., Heide-Wessel C. van der, Heide R.M. van der. Weight, smoking and mortality. J+A+M+A 252:2895-2860,1984. 80 Van Itallie T.B. Health implications of overweight and obesity in the United States. AnnilniiMed. 103:983- 988,1985. Wilson B.R.A. Somatotype, mortality, and morbidity of former Michigan State University athletes and non- athletes. Unpublished Ph.D Dissertation, School of Health Education, Counselling Psychology and Human Performance, Michigan State University, 1986. Yu B.P., Masoro F.J., Murata I., Bertrand H.A., Lynd F.T. Life span study of SPF Fischer male rats fed ad libitum or restricted diets: Longevity, growth, lean body mass and disease. Jlfiexgnigl, 37:130-141,1982. 1983 Metropolitan Height and Weight Tables. Statistical Bulletin, Metropolitan Life Insurance Company. 64:3,1983. APPENDIX 1 THE QUESTIONNAIRES 1952, 1967, 1975, 1985 81 NATIONAL STUDY OF LONGEVITY AND MORBIDITY 0F ATHLETE IN COLLEGES AND UNIVERSITIES mAMh-hhm'h-ndndhehmu-numm W-Mbuh-umflw-Mfld his fishnet-harm» Wain-'- Tuna-manual. wmnuvnss HAMIM flu-Infi- has“ in. “immune-“s M Ind-fl h. Ind-7 ”pad-uni“ I-uh-h-I‘n-In—flym—d (Gut-ts u and w_“__ Iii-“MQHM m-) '- Afic- L'IM ~— u yuan men yrs.- in mmmm.£au_afimmmhsm “unloved-n1“ nu-nudu—«mm (My-nan“) A. Vb- lnI-un ans men us. in. in. h in. in in. In. Ul- men in In. Us. Ul- ynse in in in. h amt-.75.".- “d” A. ”I. 1'!!- mummies-snvm—m am I-uinmunh‘dfinm.-thdhumndhfldnhmdm mm MStnmolflo-e FreubrlyClnfldlnoodUpwud WA—AMW 0"- I (chat all) (sh-st a.) W— m _ (07H) 82 warn-7 All-HEN? 1. huh-IMDh-(hh-odmi- xmmm~_h_ MI‘- “bu—.— d-s—d- In“ “- h H ”a“ ”L H“ T‘s—MM.— “haven-Hen unfli- “.mhm._ S.W(uW‘Ifl-D "Wham-daamdqfi- do...) that“ Elfin—hym— blinks-74“ M—MV"h-_ tmhtflm-wuud Cd“): aging-amnes- hmflm—nfifi—fi— “think- h‘ Nanak-7 mmmmmumumh ital-madam Mmflaflddmfln~- (HWMMCMmhI-HyJH-n‘bo) mum-“um.“ MM- ”Maui‘s-“hm I.“ 'fldmmu-Ispln-l nu~nmfihh3mmumu sin-dash” ‘Wumh—IH h'd-r-Ism-n-nn'fiuagun“ hutch-3.: 83 “T101051. STUDY OF monozvr'n' AND MORBIDI‘I'Y 0!" sum: GRADUATE OF COLLEGES AND umvmsr'mx hIMMhh—IhflH-‘nbflhm “Mb“M-mdw-hflfl h hdAl-Obufl) Y-dh “standout-Celt. I’m.“ Imam but “d‘ “I. C-duflcu-u Ps-y h—n-fl h hand-1 n—n-d-au-‘h‘ l-whhuh—hn-fl—d (“-13 he and hid—Ids— haiku-Huh 4“.) he Afl- M ‘GunlSpennflinuqulm-m a. bi“ u Cd. Anon-us I” mmmmwmmrmhsm usual-dummi- “sandman-in («yuan—n.s» A. Vb.— I-I-nnn Ill m. in in. it 5‘ mu yes. h I!» .- wane pm. be. h h yeah ”I. is h h- mums-vie. I‘d“ “*m. an. wmmtu):m_m III I-eQ-mh‘dhmJ—buh-ndhfl-hhn-‘dm m M MMndeeFmWyWUp-ud A New and Desist: an a AM Calm Yong (shad .) (ah-k -) m __ m ...___.. UM __ "my .—__ m1 I“ lbs-7 “WENT 1 W‘Mhm-dmh none—n u Mun—O..— In“ his“ I'D). his“ mun-‘— ”*m-qflne “-n‘h‘. ”.mm‘m._ SW (“fink-flu“..- "mm“mmn—Ifl- d“) CAI-5* lwm—h—pm_ kind“ mm—wvuh— duh-Mmchhfiud Q‘s—.9: sung-amuse hfi Mb WEI-nay hmmMMumuuhuda-Lh halal-Indus? Mulwdmms~_m (IWMMCMu—shhflmflmu) Museum-window" m- "Maui‘s-flaming... hush-ls: QMthd-lh- ”Mdbnflhhmhfl'fl umwmm-lnumnn-I-nd Mum"—' Name of Alumnus Street 85 Serial No. SECOND FOLLOW-UP OF THE LONCEVITY AND MOBBIDITY OF MALE GRADUATES OF MICHIGAN STATE UNIVERSITY t Date City State PERSONAL INFORMATION 1. Have there been any changes in your marital status since 1960 (our previous follow up)? Yes D No D (If yes to question I. answer A: if no. move on to question 2) A. Please Explain 2. Present weight __ lbs. A. Have you lost 15 lbs. or more since 1%0? Yes D No D (If yes to question A. answer 1 and 2; if no. move on to question 3) I. How many times did you lose this much weight? 1-2 times D 3 or more tine-s D 2. Any specific reason for these weight fluctuations? 3. Height (in inches,I 4. Which of these body type classification do you feel is closest to your body build? Stocky D Medium B Slender D MCUPATIONAL INFORMATION 5. Are you presently working (job or self employed)? Yes D No C] (If no. answer A: ifyes, move on to question 6) A. Have you had a job or been self employed at any time since 1&0? Yes O Na 0 (If no. skip to question 7; if yes, move on to question 6) 6. Answer the following questions about your present occupation or the last job you have had since 1950, A. B. van n no 0 ._..—.1 9—a . About how much walking getting to and from your job? Blocks . How many hours a week do you work on your job? . How much tension in your job? Great Deal D Some D Very Little D None 0 What kind of work (for example. engineer. teacher. doctorl About how much time on the job is spent sitting? Practically all I] More than half D About half D ' Almost none D About how much time on the job is spent walking? Practically all U More than half D About half D Almost none D Miles _ What type of transportation do you use to and from your job (check all that apply) Subway 0 Bus D Car [:1 Bicycle 0 Others (Please describe)— How often do you have to lift heavy weights or carry heavy things on the job? Frequently 0 Sometimes 0 Very infrequently (or never) D (Hours per week) Any responsibility for supervising other workers on the job? Yes D No D (If yes. answer I: if no. move on to I) l. About how many on the average do you supervise? J When did you start on this job? Year . Just before this job were you doing the same type of work? Yes. did the same type of work D I was on that job ._ years. No. this was my first 10130, No. did different type of work Q. If you check this item. please answer the following questions. . 2, 3. and 4: l I. How long did you do this different type of work? __ years. 2. What kind of work was it? 3. On this job did you spend more or less time sitting than your present job? More D Less D Same U 4. Was there more or less walking on this earlier job than on your present (or last) job? More 0 Less 0 Same D 86 LEISURE TIME ACTIVITIES 7. How often do you do the following? (For each activity listed. please check whether you do it frequently. sometimes. or very infrequently.) Frequently Sometimes Very Infrequently (Or Never) A. Take walk in good weather D D D B. Work around the house or apartment 0 D D (painting. repairing. etc.) C. Gardening in spring or summer D D D D. Take part in sports during season D D D E. If you take part in sports. please indicate what kind of sports and frequency either by the week or year. Frequency Frequency SPORT Per Wk. or Per Yr. SPORT Per Wk. or Per Yr. D Angling (fishing) 0 Judo - D Archery D Lawn Bowling _ — G Badminton U Mountain Climbing _ .— 0 Baseball [3 Paddle Tennis — __ D Basketball ' D Polo (horse) _ .— D Bicycling D Polo (water) _ — D Bob-Sledding C Rowing & Sculling — .— 0 Bowling (exclude lawn bowlinghere) U Shuffleboard __ — 0 Boxing C Skating (ice) __ __ D Canoeing C Skating (roller) __ __ D Codeball C Skiing __ __ D Cricket C Snow Shoeing .___.. ._.__ B Cross Couury l: Squash Rackets __ __ D Curling E Swimming _— __ D Fencing C; Table Tennis __ .— 0 Football [3 Tennis __ _..___ 5' Golf D Track & Field _— .__.._ E Gymnastics 0 Trapping — __ D Handball G Volleyball __ — 0 Hiking D Weight l‘ilting ._..__. __ C Hockey (field) D Wrestling — — E Hockey (ice) : Horseback Riding Others. E: Horseshoe Pitching D __ C Hunting C] __ 1: Ice Boating D __.... D Jan Alan D __ F. Have you been using an exercise plan at any time during or since 1%0? Yes D No D (If yes to question F. answer I and 2; if no. answer question C) 1. Please check how often you used this plan. Frequently E] Sometimes D Very infrequently D 2. Give a brief explanation of the exercises and amounts of time spent. G. Up till the time you graduated from high school did you live mostly on the farm? 0 How many years?__ Or did you live in the city? 0 How many years? DIET RECALL 8. List the things you ate and drank yesterday (this should preferably be a week day). When possible. give the specific name of the item. e.g.. Fresca or Coca Cola. rather than soft drink; McDonald's hamburger; whole milk. skim milk. half and half. rather than just milk. Indicate the amount you ate or drank in terms of cups (200 ml). tablespoons. teaspoons. ounces. numbers and approximate size. e.g.. small. large. medium for fruits. vegetables. etc. You may list meats either in ounces or size of pieces: one hamburger patty (3" diameter x 1" thick) weighs 3 02.; an average serving of steak (3" x 3" x '42") weighs 3 02. Be sure to include everything you ate or drank yesterday -candy. liquor. coffee (list sugar and cream. if used). popcorn. potato chips. etc.. as well as your regular meals. To help you estimate sizes. a rule is marked off on the edge of this page. 87 Breakfast Morning Snacks Amount or Amount or _I.t:in ' _I.tm SL— J-Imh Wk: Amount or Amount or _ltain ‘ Dinner Evening Snacks Amount or Amount or .Jim Ains— JnnH Sin_. A . Check date of diet record: Sung Mona Tues.D Wed.D Thoma Fri.D Sat.D B. Did yesterday's meals include any special or unusual event. e.g.. party. birthday. anniversary, picnic. etc.? Yes C] No D I. If yes. what was it? C. Does the above represent your usual day's food intake? Yes D No D I. If no. how did it difier from your usual intake? D 9. Do A. . Check the column which indicates the approximate frequency with which you consume each food. l etc. you drink coffee? Yes D No D (If yes. answer question a: if no. go on to question 10) What is the average number of cups per day? 1-3 B 4-6 B 7-9 C] more D . SMOKING HABITS It). Do you smoke at the present time? Yes C] No D (If yes to question 10. answer A and B) A. About how old were you when you first began to smoke? Yrs. old. B. What is the average number of cigarettes cigars pipefuls you smoke per day. (continue on to question 11) (If no to question IO. answer C) LC. Did you ever smoke regularly? Yes D No D ] (lfyes to C. answer I. 2. and 3; if no. move on to question ll) 88 About how old were you when you started smoking? Yrs. old. About how old were you when you stopped smoking? Yrs. old. When you were smoking. what was the average number of cigarettes that you smoked per day? DRINKING HABITS ll. Doyou drink at the present time? Yes D No D (If yes to question 11. answer A) Pl?!" cigars pipefuls— A. Please check the amounts you usually drink. Beer Wine Whiskey (gin. etc.) [3 Occasional bottle D Occasional glass other than for religious use Occasional glass D l to 3 bottles per day D Daily. but less than “a bottle 3 to 6 shots per day U over 3 bottles per day I] Over V: bottle per day ‘ over 6 shots per day (continue on to question 12) (If no to question ll. answer B) I B. Didyou ever drink regularly? Yes D No D (If yes to question B. answer I and 2: if no. go on to question 12) 1. Please p've the number of years that you drank regularly before you quit Yrs.. and why you quit 2. Please check the amounts you usually drank. Beer Wine Whiskey (gin. etc.) D Occasional bottle D Occasional glass other than for religious use Occasional glass lto 3 bottles per day Q Daily. but less than '2 bottle 3 to 6 shots per day over 3 bottles per day 0 Over '.~ bottle per day over 6 shots per day HEREDI'I'ARY HISTORY 12. If there are any changes in this history since 1960. will you please bring this information up to date. and make any additions or corrections in the data listed below. A. Father's occupation MEDICAL HISTORY 13. If you have had any of these diseases since 1960. will you please bring this information up to date. Make any correction or addition in the data we listed below. Age at Are you still Are you taking Ailmeiit Onset troubled with medication or this condition? treatment for it.‘. Yes .\'u \h \n High Blood Pressure Angina Pectriris Stroke (Cerebral Thrombosis) Heart Attack (Coronary Thrombosis) Rhumtit' Heart Disease Cancer Diabetes Tuberculosis l'lcer Lin-r Ailment Arthritis (Qtittl ()tlit-t E E 89 Serial No. THIRD FOLLOW-UP OF THE LONGEVITY AND MORBIDITY OF MALE GRADUATES OF MICHIGAN STATE UNIVERSITY Name of Alumnus Date Street City State Social Security Number PERSONAL INFORMATION 1. Have there been any changes in your marital status since 1968 (our previous follow-up)? Yes D No D (If yes to question I. answer A; if no. move on to question 2) A. Please Explain 2. Present weight—.lbs. A. Have you lost 15 lbs. or more since 1968? Yes D No Cl OCCUPATIONAL INFORMATION 3. Are you presently working (job or self employed)? Yes C) No D (If no. answer A; if yes. move on to question 4) A. Have you had a job or been self employed at any time since 1968? Yes D No D (If no. skip to question 5; if yes. move on to question 4) 4. Is this the same job you reported on the 1968 questionnaire? Yes D No D (If yes. move on to question 5; if no. answer the following questions A through J. A. What kind of work (for example, engineer. teacher. doctor) B. About how much time on the job is spent sitting? Practically all El More than half El About half 0 ' Almost none Cl C. About how much time on the job is spent walking? Practically all 0 More than half 0 About half 0 Almost none 0 D. Do you ever walk to or from work? Yes D No D If yes. how far do you walk? Blocks Miles How many times a year Do you ever bicycle to and from work? Yes Cl No Cl If yes. how far do you cycle (both ways)? Blocks Miles Number of times per year E. What type of transportation do you use to and from your job (check all that apply)? Subway Cl Bus Cl Car Cl Bicycle El Walking 0 Others (Please describe) F. How often do you have to lift heavy weights or carry heavy things on the job? Frequently Cl Sometimes [3 Very infrequently (or never) 0 G. How many hours a week do you work on your job? (Hours per week) H. How much tension in your job? Great deal El Some El Very little D None D 1. Any responsibility for supervising other workers on the job? Yes (:1 N00 (If yes, answer 1; if no. move on to J) I. About how many on the average do you supervise? J. When did you start on this job? Year LEISURE TIME ACTIVITIES 5. How many hours a month do you do the following activities and which months? (List number of hours involved in each activity under the monthia) you participate. Leave blank where not involved.) ACTIVITY 90 JII. Feb. "'1 June My Aug. Sept. Oct. Flsnlng - bank. boat. ice Fishing - wading Arcnery. target Badminton Baseball - hard. solt Basketball Bicycling - pleasure Tobegganing. sledding Bowling. including lawn Canoeing or rowing 4°99)” Curling Fencing Gardening Lawn mowing . riding Lawn mowmg - power mower Lawn mowing - hand mower Snow shoveling Golt . walking Goll - power can Handball. including paddleball. racket and squash Walking - beck packing Walking - cross country Walking - mountain climbing Walking - pleasure Home workshop (carpentry) Horseback riding Horseshoe pitching Hunting - bow and gun Selling - ice and water Judo. including karate Paddle tennis Flowing. skulling Shullleboard (not hand) Skating - ice. roll-r Skiing - downhill Skiing - cross country Skiing - water Snowshoeing Dancing - ballroom Dancing - square Swimming - pleasure Swlmrnlng - exercrse Table tennis Tennis - singles Tennis - doubles Volleyball Weight lilting Cellsthenics - home Calisthenics - Health Club Others: llllllllll, 91 6. If you have been routinely exercising under a home exercise plan or Health Club plan (commercial. Y.M.C.A.. Athletic Club. etc.) answer the following questions: A. Number of hours per month .which months (circle): Jan.. F eb.. Mar.. Apr.. May. June. July. Aug.. Sept.. Oct.. Nov.. Dec. B. What type of exercises? DIET RECALL 7. List the things you ate and drank yesterday (this should preferably be a week day). When possible. give the specific name of the item. e.g.. F reeca or Coca Cola. rather than soft drink; McDonald's hamburger; whole milk. skim milk. half and half. rather than just milk. Indicate the amount you ate or drank in terms of cups (200 ml). tablespoons. teaspoons. ounces. numbers and approximate size. e.g.. small. large. medium for fruits. vegetables. etc. You may list meats either in ounces or size of pieces: one hamburger patty (3" diameter x 1" thick) weighs 3 01.; an average serving of steak (3" x 3" x Vi") weighs 3 oz. Be sure to include everything you ate or drank yesterday — candy. liquor. cofl'ee (list sugar and cream. if used). popcorn, potato chips. etc.. as well as your regular meals. To help you estimate sizes. a rule is marked off on the edge of this page. Breakfast Morning Snacks Amount or Amount or Item §ize lt_e_m ' Luncn Afternoon Snacks Amount or Amount or item _§in [tin Size Dinner . Evenlng Snacks Amount or Amount or Item Size Item Size b A. Check date of diet record: Sun. 0 Mon. [3 Tues. 0 Wed. D Thurs. D Fri. (3 Sat. D B. Did yesterday 's meals include any special or unusual event. e.g.. party. birthday. anniversary. picnic. etc.? Yes D No D 1. If yes. what was it? C. Does the above represent your usual day's food intake? Yes D No D 1. If no. how did it differ from your usual intake? D. Check the column which indicates the approximate frequency with which you consume each food. ‘NOM ice cream (not lcs milk) Cream or custard other than . in cottee. tea. etc. . on cereal . on cakes. brownies. sweet rolls. etc. then around meat low or non-calorie Jelly. lam. preserves. marmalade on . . etc, 92 E. Do you drink coffee? Yes D No D (If yes. answer questionA; if no. go on to question 8) A. What is the average number of cups per day? 1-3 0 4-6 B 7-9 (:1 more D SMOKING HABITS 8. Do you smoke at the present time? Yes D No D (If yes to question 8 answer A and B; if no. answer C) A. What is the average number of cigarettes __ . cigars__ . and/or pipefulls... you smoke per day? B. Have you stopped at any time between I968 and now? Yes D No D If yes. how long did you stop ? C. Did you smoke regularly any time between 1968 and now? YesCl NoCl If no. go on to question 9. If yes. how long? How many cigarettes__. cigars._.. pipefulls._..did you smoke per day? DRINKING HABITS 9. Do you drink alcoholic beverages at the present time? Yes D No D (If yes to question 9. answer AandB:ifno.answerC) ' A. Please check the amounts you usually drink. leer Wine Liquor D Occasional bottle 0 Occasional glass other than ior religious use 0 Occasional glass CI 1 to 3 bottles per day El Daily. but less than V: bottle D 3 to 6 shots per day 0 over 3 bottles per day 0 Over Vi bottle per day 0 over 6 shots per day B. Had you stopped drinking at any time between 1968 and now? YesEJ NoD If no. go on to question 10. If yes. for how long a period 'did you stop? C. Did you drink regularly at any time between 1968 and now? Yes D No D Ifno. goon toquestion 10. If yes. for how long a period did you drink? How much? (Please check the amounts.) Beer Wine Liquor U Occasional bottle El Occasional g'lass other than ior religious use D Occasional glass 0 1to3bottles per day UDaily. but less than‘abottle D 3toeshotsperdsy 0 over 3 bottles per day D Over is bottle per day 0 over 6 shots per day HEREDITARY HISTORY 10. As of 1968. the individuals listed were still alive. Will RELATIONSHIP this information A. Father's occupation (when working) MEDICAL HISTORY 11. In 1968 you indicated you had the following conditions. Will you please bring this information up-to-date. Make any correction or addition in the data we listed below. Are you still Are you taking Aliment Age at troubled with medlcstlon or Onset this condition? treatment ior It? Yes No Yes Ho High Blood Pressure Angina Pectorls Stroke (Cerebral Thrombosis) Heart Attack (Coronary Thrombosis) Rheumatic Heart Disease Cancer Diabetes Tuberculosis Ulcer Liver Aliment Arthrlils Gaul Other [3000000000000 0000000000000 0000000000000 [3000000000000 93 Serial No. _______-- -_. l-‘Ol’liTlI FOLLOW-LT? OF THE LONCEYITY AND .xlolililni'n' ()F MALE GRADUATES or MICHIGAN STATE L'.\'l\'F.llSl’l'\' Name of Alumnus Date Street City State Social Security Number PERSONAL INFORMATION 1. Have there been any changes in your marital status since 1976 (our previous follow-up)? Yes D No D (If yes to question I. answer A; if no. move on to question 2) A. Please Explain 2. Present weight—lbs. Have you lost 15 lbs. or more since 1976? Yes D No D OCCUPATIONAL INFORMATION 3. Are you presently working (job or self employed)? Yes D No D (If no. answer A; if yes. move on to question 4) A. Have you had a job or been self employed at any time since 1976? YesCl No D (If no. answer A: if yes. more on to question 4) -l. Is this the same job you reported on the 1976 questionnaire? Yes D No Cl (If yes. move on to question 5; if no. answer the following questions A through J. A. What kind of work (for example. engineer. teacher. doctor) B. About how much time on the Job is spent sitting? Practically all El More than half Cl About half 0 Almost none 0 C. About how much time on the job is spent walking? Practically all 0 More than half El About half 0 Almost none D D. Do you ever walk to or from work? Yes D No D If yes. how far do you walk? Blocks Miles How many times a year Do you ever bicycle to and from work? Yes D No Cl If yes, how far do you cycle (both ways)? Blocks Miles Number of times per year E. What type of transportation do you use to and from your job (check all that apply)? SubwayD Bus Cl Car D Bicycle Cl Walking 0 Others (Please describe) F. How often do you have to lift heavy weights or carry heavy things on the job? Frequently [3 Sometimes 0 Very infrequently (or never) 0 G. How many hours a week do you work on your job? (Hours per week) H. How much tension in your job? Great deal 0 Some D Very little 0 None I] I. Any responsibility for supervising other workers on the job? Yes D NOD (If yes. answer 1; if no. move on to J) l. About how many on the average do you supervise? J. When did you start on this job? Year LEISURE TIME ACTIVITIES 5. How many hours a month do you do the following activities and which months? (List number of hours involved in each activity under the monthls) you participate. Leave blank where not involved.) 94 ACTIVITY Jan. Feb. Am «or June July Aug. Sept. Oct. Fishing - bank. boat. ice Fishing - wading Archery. target Badminton Baseball - hard. soit Basketball Bicycling - pleasure Tobegganing. sledding Bowling. including lawn Canoeing or rowmg Jossino Curling Fencing Gardening Lawn mowing - riding Lawn mowing . power mower Lawn mowing - hand mower Snow shoveling Golt - walking Goli - power cart Handball. including paddleball. racket and squash Walking - back packing Walking - cross country Walking - mountain climbing Walking - pleasure Home workshop (carpentry) Horseback riding Horseshoe pitching Hunting - bow and gun Sailing - ice and water Judo. including karate Paddle tennis Rowing. skulling Shuttleboard (not hand) Skating - ice. roller Skiing - downhill Skiing - cross country Skiing - water Snowshoeing Dancing - ballroom Dancing - square Swimming - pleasure Swimming - exercise Table tennis Tennis - singles Tennis - doubles Volleyball Weight lilting Calisthsnics - home Calisthenics - Health Club Others: I lllllllll-l-ll 95 6. If you have been routinely exercising under a home exercise plan or Health Club plan (commercial. Y.M.C.A., Athletic Club. etc.) answer the following questions: A. Number of hours per month .which months (circle): Jan.. F eb.. Mar.. Apr.. May. June. July. Aug.. Sept.. Oct.. Nov.. Dec. B. What type of exercises? DIET RECALL 7. List the things you ate and drank yesterday (this should preferably be a week day). When possible. give the specific name of the item. e.g.. F resca or Coca Cola. rather than soft drink; McDonald's hamburger: whole milk. skim milk. half and half . rather than just milk. Indicate the amount you ate or drank in terms of cups (200 ml). tablespoons. teaspoons. ounces. numbers and approximate size. e.g.. small. large. medium for fruits. vegetables. etc. You may list meats either in ounces or size of pieces: one hamburger patty (3" diameter x 1" thick) weighs 3 02.; an average serving of steak (3" x 3" x Vi") weighs 3 oz. Be sure to include everything you ate or drank yesterday — candy. liquor. cofl'ee (list sugar and cream. if used). popcorn. potato chips. etc.. as well as your regular meals. To help you estimate sizes. a rule is marked off on the edge of this page. l—Breakfast Mornifl Snacks Amount or I Amount or item §i_ze ltgm . §_ize l Lunch Afternoon Snacks Amount or T Amount 0r M . sz_e ltlm Size Dinner Evening SnaCks Amount or Amount or i Item Size ltem Size .i ’ l. . t . i A. Check date of diet record: Sun. El Mon. D Tues. El Wed. 0 Thurs. Cl Fri. 0 Sat. CI - l B. Did yesterday's meals include any special or unusual event. e.g.. party. birthday. anniversary. picnic. ' etc.? Yes Cl No D I. If yes. what was it? . C. Does the above represent your usual day's food intake? Yes Cl No D I. If no. how did it differ from your usual intake? D. Check the column which indicates the approximate frequency with which you consume each food. a l Never ‘ Cream or half ice cream (not ice milk Cream or custard Other than . in coffee. tea. etc. 0 0t On cereal . On trolls not lOW calorie rosted cakes. brownies. sweet rolls. etc (other around meat low or non-calorie) Jelly. ism. preserves. marmalade potatoes ups (on we . etc i ried meal. , Molasses lat 96 E. Do you drink coffee? Yes D No D (If yes. answer questionA; if no. go on to question 8) A. What is the average number of cups per day? 13 Cl 4-6 C] 79 Cl more Cl SMOKING HABITS 8. Do you smoke at the present time? Yes D No D (If yes to question 8 answer A and B; if no. answer C) A. What is the average number of cigarettes__. cigare_ . and/or pipefulls... you smoke per day? B. Have you stopped at any time between I976 and now? Yes D No D If yes. how long did you stop ? C. Did you smoke regularly any time between 1976 and now? YesEl NoD If no. go on to question 9. If yes. how long? How many cigarettes_. cigare_. pipefulls_did you smoke per day? DRINKING HABITS 9. Do you drink alcoholic beverages at the present time? Yes D No D (If yes to question 9. answer Aand B:ifno.answerC) A. Please check the amounts you usually drink. leer Wile Liquor D Occasional bottle 0 Occasional glass other than for religious use 0 Occasional glass Cl 1 to 3 bottles per day D Daily. but less than Vi bottle Cl 3 to 6 shots per day 0 over 3 bottles per day 0 Over Vi bottle per day D over 6 shots per day B. Had you stopped drinking at any time between 1976 and now? Yes El NOD If no. go on to question 10. If yes. for how long a period did you stop? C. Did you drink regularly at any time between 1976 and now? Yes D No D If no. go ontoqueetion 10. If yes. for how long a period did you drink? How much? (Please check the amounts.) leer Wine Liquor D Occasional bottle 0 Occasional glass other than for religious use 0 Occasional glass Cl 1 to 3 bottles per day 0 Daily. but less than is bottle 0 s to 6 shots per day D over 3 bottles per day 0 Over Vi bottle per day 0 over 6 shots per day HEREDITARY HISTORY 10. As of l the individuals listed were still alive. Will vou RELATIONSHIP this information A. F ather's occupation (when working) MEDICAL HISTORY II. In 1976 you indicated you had the following conditions. Will you please bring this information up-to-date. Make any correction or addition in the data we listed below. Are you still Are you taking Alment Age at troubled with medication or Onset this condition? treatment for it? Yes No Yes Ho High Blood Pressure 0 Cl C D Angina Psctoris D C] U 0 Stroke (Cerebral Thrombosis) 0 Cl 0 0 Heart Attack (Coronary Thrombosis) D U U D Rheumatic Heart Disease 0 D U 0 Cancer 0 D U D Diabetes 0 U U D Tuberculosis 0 Cl 0 D Ulcer U D U 0 Liver Aliment D D U 0 Arthritis El El Cl Cl Gout D D D D Other 0 D D D APPENDIX 2 CALORIC STANDARDS FOR THE MICHIGAN STATE UNIVERSITY LONGEVITY STUDY ACTIVITIES 97 CALORIC STANDARDS FOR THE MICHIGAN STATE UNIVERSITY LONGEVITY STUDY ACTIVITIES. (QUINN, 1987) Kcal/min/lb 1) Fishing (bankiboat.ice) ............ .020 2) Fishing (wading) ................... .028 3) Archery ............................ .029 4) Badmington ......................... .040 5) Baseball (hard,soft) ............... .031 6) Basketball ......................... .045 7) Bicycle (pleasure,5 mph)* .......... .029 8) Tobogganing (sled) ................. .025 9) Bowling ............................ .029 10) Canoeing (rowing,leisure)* ......... .020 11) Joggingk ........................... .074 12) Chrling ............................ .020 13) Fencing ............................ ' .033 14) Gardening .......................... .039 15) Lawn mowing (riding) ............... .017 16) Lawn mowing (power mower)* ......... .051 17) Lawn mowing (hand mower)* .......... .055 18) Snow shovelling .................... .039 19) Golf (walking)* .................... .039 20) Golf (power cart) .................. .020 21) Handball ........................... .080 22) Walking (backpacking)* ............. .050 23) Walking (cross country)* ........... .044 24) Walking (mountain climbing)* ....... .055 25) Walking (pleasure)* ................ .036 26) Heme workshop (carpentry) .......... .023 27) Herseback riding (trotting).; ...... .045 28) Hcrseshoe pitching ................. .023 29) Hunting (bow and gun) .............. .040 30) Sailing (ice and water) ............ .020 31) Judo ............................... .089 32) Paddle tennis ...................... .033 33) Rowing (sculling)* ................. .029 34) Shuffleboard ....................... .020 35) Skating (ice,roller)* .............. .038 36) Skiing (downhill) .................. .064 37) Skiing (cross country)* ............ .080 38) Skiing (water) ..................... .052 39) Snowshoeing (2.3 mph)* ............. .060 40) Dancing (ballroom)* ................ .029 41) Dancing (squam)* .................. .045 42) Swiimiing (pleaslarewg ............... .045 43) Swiming (exercise)* ............... .058 44) Table tennis ....................... .031 45) Tennis (singles) ................... .050 46) Tennis (doubles) ................... .045 47) VOlleyball ......................... .022 48) weight lifting ..................... .049 49) Calisthenics (home)* ............... .033 50) Calisthenics (health club)* ........ .040 * ACTIVITIES UTILIZED IN THE AEROBIC CALORIC EXPENDITURE CALCULATIONS. APPENDIX 3 MSU NUTRIENT DATABASE C(MPUTER COMMANDS USED 98 MSU NUTRIENT DATA BASE COMMAND STATEMENTS USED *JOBCARD*,RGI,JC10000,L1000,GE1,PE1,CM60000. HAL.LIB.UNSUP. APLIB9,TT3368,R*DEMORAN=DEMOIN,R*EOODUPN=FOODTBL. APLIB9,TT3279,*889704. ATTACH,FOODIN,FINALDIETS. OOFYBR,INPUT,PARAM. FILE,REPORTO,FO=SQ,BFS=2054. FILE,FOODIN,F0=SQ,BFS=2054. SWITCH,6,0N. SWITCH.5.0FF. 889704. *CATALOG,RDAOUT,FINALRDA,ID=APUMH,RP=1. *CATALOG,NUTOUT,FINALNUT,ID=APUMH,RP=1. *CATALOG,PERCENT,FINALPERCENT,RP=1,ID=APUMH. CATALOG,REPORTO,FINALREPORT.RP=1,ID=APUMH. COPYBF,REPORTO. EXIT. CATALOG,REPORTO,FINALREPORT,RP=1,ID=APUMH. REWIND,REPORTO. OOPYBF,REPORTO. *EOR OUT YYYNYYY99999 NUTl NB I! l )3)!!! I! YYNYNYYNNNYY NUTZ YNNNNNNN *EOR MEALCARDS GO