Search

Since Hippocrates first advised us more than 2000 years ago that exercise-though not too much of it--was good for health, the epidemiology of physical activity has developed apace with the epidemiological method itself. It was only in the mid-20th century that Professor Jeremy N Morris and his associates used quantitative analyses, which dealt with possible selection and confounding biases, to show that vigorous exercise protects against coronary heart disease (CHD). They began by demonstrating an apparent protection against CHD enjoyed by active conductors compared with sedentary drivers of London double-decker buses. In addition, postmen seemed to be protected against CHD like conductors, as opposed to less active government workers. The Morris group pursued the matter further, adapting classical infectious disease epidemiology to the new problems of chronic, non-communicable diseases. Realizing that if physical exercise were to be shown to contribute to the prevention of CHD, it would have to be accomplished through study of leisure-time activities, presumably because of a lack of variability in intensities of physical work. Accordingly, they chose typical sedentary middle-management grade men for study, obtained 5-minute logs of their activities over a 2-day period, and followed them for non-fatal and fatal diseases. In a subsequent study, Morris et al. queried such executive-grade civil servants by detailed mail-back questionnaires on their health habits and health status. They then followed these men for chronic disease occurrence, as in the earlier survey. By 1973 they had distinguished between 'moderately vigorous' and 'vigorous' exercise. In both of these civil service surveys, they demonstrated strong associations between moderately vigorous or vigorous exercise and CHD occurrence, independent of other associations, in age classes 35-64 years. In the last 30 years, with modern-day computers, a large number of epidemiological studies have been conducted in both sexes, in different ethnic groups, in broad age classes, in a variety of social groups, and on most continents of the world. These studies have extended and amplified those of the Morris group, thereby helping to solidify the cause-and-effect evidence that exercise protects against heart disease and averts premature mortality.

Although many studies suggest that consumption of alcohol increases the risk of several site-specific cancers, the evidence remains unclear for prostate cancer. Few data exist on beverage-specific associations as well as lifetime patterns of alcohol consumption and prostate cancer risk.

Physical activity has been proposed to decrease lung cancer risk; however, few data are available. Further, no studies have examined specific kinds and intensities of activities.

The associations of self-reported measures of physical activity from a mail survey with an objective measure of physical fitness were investigated. Respondents to a health status survey (n = 12,225), conducted in 1982 by the Institute for Aerobics Research in Dallas, TX, formed the population. From this group, males who also had a clinical examination within 60 days of the return of their questionnaire served as subjects. The study subjects (n = 375, mean age = 47.1 years) completed a maximal physical fitness assessment using a modified Balke protocol. The questionnaire included a section of inquiries concerning leisure time physical activity participation in which subjects were asked to quantitatively recall exercise participation for varying periods of time. Reported exercise participation values were converted to estimates of energy expenditure and combined into overall indices of physical activity participation. Multiple regression analyses were used to determine the individual contributions of the physical activity indices in predicting maximal treadmill performance (physical fitness). Significant predictors of physical fitness were age (beta = -0.34), an index of running, walking, and jogging participation (beta = 0.31), and the response to a question on frequency of sweating (beta = 0.35). The multiple correlation coefficient for these variables in predicting physical fitness was 0.65. These results indicate that exercise behavior can be accurately estimated in large populations by using simple questions in a mail survey.

While the inverse association between physical activity and coronary heart disease risk is well documented, questions remain regarding the intensity of exercise, the potential for confounding by other risk factors for coronary heart disease, and the role of blood lipids and apolipoproteins. The authors examined these issues in the Boston Area Health Study, a case-control study of 340 patients (266 men, 74 women) who survived a first myocardial infarction between January 1, 1982, and December 31, 1983, and 340 controls matched on sex, age, and residence. The relative risk of myocardial infarction for those in the highest quartile of physical activity, compared with the lowest, was 0.50 (95 percent confidence interval (CI) 0.31-0.80) for men and 1.00 (95 percent CI 0.41-2.43) for women. When subjects were categorized by level of energy expenditure on moderate to vigorous sports alone, men in the most active category had 0.39 (95 percent CI 0.23-0.69) times the risk of those in the least active category, and women, 0.43 (95 percent CI 0.15-1.26) times the risk. Adjustment for body mass index, smoking, alcohol intake, diet, personal and family medical history, and personality type did not substantially change results nor did further adjustment for blood lipids. This was not surprising as total energy expenditure was uncorrelated with blood lipids or apolipoproteins. Moderate to vigorous sporting activity, however, appeared to be directly related to high density lipoprotein (HDL) cholesterol (p = 0.06), especially the HDL2 subfraction (p = 0.10). In these data, findings suggest that physical activity is inversely related to myocardial infarction risk, independently of other risk factors for coronary heart disease.

Clinical observations and laboratory experiments have suggested a role for testosterone in the development of prostatic cancer. Since physical training may act to lower testosterone levels, men who are very active may have a reduced risk of this cancer. To test this hypothesis, the authors prospectively followed 17,719 Harvard alumni (aged 30-79 years) from 1965 or 1969 to 1977, and 1980 to 1988 for the occurrence of prostatic cancer (n = 419). Physical activity was assessed in either 1962 or 1966, and again in 1977, based on self-reported stair climbing, walking, and sports played. Alumni who expended greater than 4,000 kcal/week at both assessments were at reduced risk of developing prostatic cancer (age-adjusted rate ratio, 0.12; 95% confidence interval 0.02-0.89) compared with their inactive counterparts who expended less than 1,000 kcal/week at both assessments. These results should be interpreted cautiously, since only one case of prostatic cancer was identified among alumni highly active at both assessments, who contributed 4% of total person-years to the analysis. Among alumni aged 70 years and older, those who extended greater than 4,000 kcal/week at either assessment had about half the risk of those who expended less than 1,000 kcal/week at either assessment (age-adjusted rate ratio, 0.53; 95% confidence interval 0.29-0.95). The authors attribute these findings to the increased precision with which physical activity could be measured when combining two activity assessments. Alternatively, long-term maintenance of a high level of physical activity may be necessary for further reduction of risk. There was no evidence of a dose-response relation.

Little direct evidence exists in the epidemiologic literature to support the widely held belief that an individual's physical activity behavior varies over the years. To provide data, the authors prospectively studied patterns of physical activity among 6,092 Harvard College alumni (average age at the start of follow-up, 43 years) who responded to questionnaires in 1962 or 1966, 1977, and 1988, and who declared themselves free of cardiovascular disease and cancer throughout that period. The median energy expenditure increased between 1962 or 1966 and 1977, but stabilized between 1977 and 1988. However, rank correlation coefficients between estimates of energy expenditure in 1977 and 1988 ranged from 0.36 to 0.41 for the various age groups (compared with 0.81-0.84 for the Quetelet index, and 0.60-0.65 for the current number of cigarettes smoked per day), implying that physical activity at the individual level was not constant. For 4,238 alumni for whom data on collegiate sports participation were available, the rank correlation coefficients between collegiate activity and alumni activity 1-7 decades later were even lower, ranging from 0.05 to 0.17 (compared with 0.31-0.58 for the Quetelet index). Thus, physical activity assessed over a short time period in an individual may not be a valid proxy for activity over the long term. The characteristics that distinguished alumni who consistently expended less than 1,000 kcal/week (in climbing stairs, walking, and playing sports) in 1962 or 1966, 1977, and 1988 from those who consistently expended greater than 2,500 kcal/week were as follows: at baseline, older age, overweight, and cigarette smoking; and during college, not having been a varsity athlete and fewer hours playing sports.

Assessment of habitual physical activity in epidemiologic and health education studies has been difficult. A seven-day physical activity recall interview was developed and administered in a community health survey, a randomized clinical trial, and two worksite health promotion programs during 1979-1982. These studies were conducted in several populations in California, Texas, Pennsylvania, and New Jersey. Energy expenditure estimates from the physical activity recall conformed to expected age- and sex-specific values in the cross-sectional community survey. Estimates of energy expenditure were also congruent with other questions on physical activity and job classification. In a randomized, one-year exercise trial, the physical activity recall detected increases in energy expenditure in the treated group and was positively associated with miles run during training (p less than 0.05). Changes in energy expenditure were associated with changes in maximal oxygen uptake (VO2max (r = 0.33, p less than 0.05) and body fatness (r = -0.50, p less than 0.01) at six months, and in high density lipoprotein-cholesterol (r = 0.31, p less than 0.05) and triglyceride (r = -0.41, p less than 0.01) at one year. The physical activity recall detected significant (p less than 0.01) increases in energy expenditure in treatment groups in two worksite health promotion projects. These data suggest that the physical activity recall provides useful estimates of habitual physical activity for research in epidemiologic and health education studies.

A case-control analysis of Parkinson's disease and infections in childhood was conducted in a cohort of 50,002 men who attended Harvard College (Cambridge, MA) or the University of Pennsylvania (Philadelphia, PA) between 1916 and 1950 and who were followed in adulthood for morbidity and mortality data. Cases of Parkinson's disease were identified from responses to mailed questionnaires and death certificates through 1978. Four controls from the same population were selected for each case. A reduced risk of Parkinson's disease was associated with most childhood viral infections. The negative association was statistically significant for a history of measles prior to college entrance (exposure odds ratio = 0.53; 95% confidence limits: 0.31, 0.93). The reduced risk of Parkinson's disease among subjects with a positive history of measles in childhood may reflect an adverse effect of measles in adulthood or of subclinical or atypical measles. Furthermore, a negative history of measles, especially if associated with a lack of other common diseases, could be a marker for negative influenza history before 1918 and thus a higher risk of infection during the 1918 influenza epidemic, because of the lack of partial influenza immunity. These data may also suggest a truly protective effect of measles, compatible with some complex interaction between measles virus and the virus of the 1918 influenza epidemic.

Previous measures of physical activity for epidemiologic studies were considered inadequate to meet the needs of a community-based health education trial. Therefore, new methods of quantifying the physical activity habits of communities were developed which are practical for large health surveys, provide information on the distribution of activity habits in the population, can detect changes in activity over time, and can be compared with other epidemiologic studies of physical activity. Independent self-reports of vigorous activity (at least 6 metabolic equivalents (METs) ), moderate activity (3-5 METs), and total energy expenditure (kilocalories per day) are described, and the physical activity practices of samples of California cities are presented. Relationships between physical activity measures and age, education, occupation, ethnicity, marital status, and body mass index are analyzed, and the reliabilities of the three activity indices are reported. The new assessment procedure is contrasted with nine other measures of physical activity used in community surveys.

In a study population of 14,998 Harvard male alumni, 681 hypertensives were first diagnosed during a 6-10-year follow-up beginning 16-50 years after college entrance. The study comprised 105,662 man-years of observation of these men who had entered college in 1916-1950, and who were followed from 1962 or 1966 to 1972. Presence or absence of a background of collegiate sports did not influence risk of hypertension in this study population, nor did stair-climbing, walking, or light sports play by alumni. But, alumni who did not engage in vigorous sports play were at 35% greater risk of hypertension than those who did, and this relationship held at all ages, 35-74 years. Higher levels of body mass index, weight gain since college, history of parental hypertension, and lack of strenuous exercise independently predicted increased risk of hypertension in alumni. Men 20% or more over ideal weight-for-height were at 78% greater risk than lighter men. Those who had gained 25+ lbs (c. 11.5+ kg) since entering college were at 60% greater risk than those who had gained less. Alumni with a hypertensive parent were at 83% higher risk than men without such parentage. Contemporary vigorous exercise was inversely related to hypertension risk, but chiefly among alumni overweight-for-height. In the clinical sense, attributable risk estimates ranged from 30% to nearly 50% for the alumni characteristics of overweight, weight gain, parental hypertension, and lack of vigorous exercise. In the community sense, attributable risk of these same characteristics ranged 13-26%. To sum up, vigorous exercise is associated with lower hypertension incidence, and, without necessarily altering body weight-for-height, avoids or reduces fat and promotes muscle; obesity, rather than excess weight-for-height, is associated with higher hypertension incidence; hence, vigorous exercise is appropriate for use as an intervention regimen in the prevention of hypertension.

In a 22-year followup of 3686 San Francisco longshoremen, the roles of physical activity, cigarette smoking habit, and systolic blood pressure level were evaluated independently in relation to risk of death from a broad range of diseases. Smoking pattern and blood pressure status were established in 1951 and job activity was assessed annually during the followup period. Lower levels of energy expenditure predicted increased risk of fatal heart attack and perhaps of stroke. Heavy cigarette smoking predicted increased risk of death from heart attack, cancer, chronic obstructive respiratory disease, and pneumonia. Higher levels of systolic blood pressure were associated with death from all cardiovascular diseases, diabetes mellitus, and cirrhosis. Tacit to these findings: sedentary living takes its toll largely through heart disease and stroke; the toxicity of cigarette smoking is associated with a broader range of diseases, including heart attack, cancer, and respiratory disease; and higher level of blood pressure related to an even broader range of cardiovascular disease than either of the other characteristics studied.

Risk of first heart attack was found to be related inversely to energy expenditure reported by 16,936 Harvard male alumni, aged 35-74 years, of whom 572 experienced heart attacks in 117,680 person-years of followup. Stairs climbed, blocks walked, strenuous sports played, and a composite physical activity index all opposed risk. Men with index below 2000 kilocalories per week were at 64% higher risk than classmates with higher index. Adult exercise was independent of other influences on heart attack risk, and peak exertion as strenuous sports play enhanced the effect of total energy expenditure. Notably, alumni physical activity supplanted student athleticism assessed in college 16-50 years earlier. If it is postulated that varsity athlete status implies selective cardiovascular fitness, such selection alone is insufficient to explain lower heart attack risk in later adult years. Ex-varsity athletes retained lower risk only if they maintained a high physical activity index as alumni.

The relationship between sociocultural mobility and subsequent coronary heart disease (CHD) and hypertension (HT) was studied in a cohort of 13,728 male former Harvard University students examined in 1939-1950. All 13,728 were followed for CHD mortality, while 8852 returned self-administered mail questionnaires in 1962 or 1966 and in 1972 which queried for doctor-diagnosed myocardial infarction (MI), angina pectoris (AP) and HT. For each of 98 cases of fatal CHD, 78 cases of MI and 48 cases of AP, four controls were randomly selected. One control was selected for each of 319 HT cases. Significant negative associations between father's occupational status and risk of combined fatal CHD and MI and between geographic mobility and risk of HT were noted in univariate analysis. These associations persisted with stratification by individual confounding factors and by a multivariate confounder-summarizing score. Intergenerational mobility, as indicated by occupational status of the father, was associated with a 1.5 times increased risk of fatal CHD and MI. Intragenerational geographic mobility was associated with a slightly reduced risk of HT.

Pike et al. (Nature 1983;303:767-70) and Moolgavkar et al. (JNCI 1980;65:59-69) proposed quantitative theories for the effect on age-specific breast cancer risk of ages at menarche, first childbirth, and menopause. Here the incidence rate functions predicted by these theories are fit to data for 1,884 women of all ages with breast cancer and 3,432 matched controls admitted to San Francisco Bay area hospitals in 1970-1977. A third function describing age-specific breast cancer risk based on the timing of childbearing and menstrual events is presented, and its fit to the data is compared with that of the functions of Pike et al. and Moolgavkar et al. None of the three fully accounted for the protective effects of early age at first childbirth in premenopausal women or of early age at menopause in parous postmenopausal women. To account for the effects of total parity and body mass (Quetelet) index on risk of breast cancer occurrence, the authors developed a fourth incidence rate function by extending the third. Goodness of fit to the data of the fourth function is demonstrated. Age-specific relative risks of breast cancer according to childbearing, menstrual events, and body size are estimated from the fourth function. The main qualitative findings are that 1) the protective effects of late menarche and of early first full-term pregnancy are greater in premenopausal than in postmenopausal women; 2) first full-term pregnancy initially boosts the level of risk, but incidence rates increase with age more slowly thereafter; 3) among the parous, multiparity is protective both in premenopausal and postmenopausal women, regardless of age at first full-term pregnancy; 4) both nulliparous and lean women are more protected by early menopause than are parous and overweight women; 5) increased body mass index is protective before, but detrimental after, menopause; and 6) postmenopausal incidence rates increase with age more rapidly among overweight than among lean women.

Vaginal exposures to talc and other particulates may play an etiologic role in epithelial ovarian cancer. Surgical sterilization may protect against ovarian cancer by blocking entry of such particulates into the peritoneal cavity. The authors assessed histories of talcum powder use, tubal sterilization, and hysterectomy with ovarian conservation in 188 women in the San Francisco Bay Area with epithelial ovarian cancers diagnosed in 1983-1985 and in 539 control women. To investigate the roles of blood-borne environmental exposures on ovarian cancer risk, they assessed lifetime consumption of coffee, tobacco, and alcohol in these women. Of the 539 controls, 280 were hospitalized women without overt cancer, and 259 were chosen from the general population by random digit telephone dialing. Ninety-seven (52%) of the cancer patients habitually used talcum powder on the perineum, compared with 247 (46%) of the controls. Adjusted for parity, the relative risk (RR) = 1.40, p = 0.06. There were no statistically significant trends with increasing frequency or duration of talc use, and patients did not differ from controls in use of talc on sanitary pads and/or contraceptive diaphragms. Fewer ovarian cancer patients (7%) than controls (13%) reported prior fallopian tube ligation (RR, adjusted for parity, = 0.56, p = 0.06), and fewer patients (20%) than controls (28%) reported prior hysterectomy (RR = 0.66, p = 0.05). The protective effect of hysterectomy was confined to those who underwent this surgery 10 or more years prior to interview and to those who had not undergone prior tubal sterilization. Consumption of cigarettes and alcohol did not differ between cases and controls. By contrast, 11 (6%) cases never regularly consumed coffee, compared with 31 (11%) hospital controls and 26 (10%) population controls (RR, adjusted for smoking, = 2.2, p = 0.03, for the comparison using all controls). Overall, ovarian cancer risk among women who had drunk coffee for more than 40 years was 3.4 times that of women who had never regularly consumed coffee (p less than 0.01). However, the data exhibited no clear trends in risk with increasing consumption. Although risk ratios relating duration of coffee drinking to ovarian cancer were unaffected by adjustment for several characteristics, further study is needed to exclude potential confounding by other unmeasured characteristics.

Numerous studies consistently have shown that higher levels of physical activity are associated with decreased risks of coronary heart disease, cerebrovascular disease, hypertension, non-insulin-dependent diabetes mellitus, colon and, possibly, breast cancer, as well as osteoporosis. The biological processes proposed to explain these inverse associations are highly plausible. If physical activity does reduce the risk of developing these chronic diseases, we also would expect physical activity to delay mortality and enhance longevity. In this article, we review the major epidemiological studies worldwide that have examined the association between physical activity or physical fitness and all-cause mortality. The data from these studies indicate that physical activity is effective in postponing mortality and enhancing longevity. Public health professionals worldwide should emphasize the need to increase activity levels during leisure time, as well as the need to incorporate physical activity into the daily activities of life.

In two case-control studies conducted in the six-county San Francisco Bay Area, 111 women diagnosed with epithelial ovarian carcinoma in 1974-1977 and 188 women diagnosed in 1983-1985 were interviewed concerning their menstrual, sexual, and reproductive histories. For comparison, interviews were conducted with 752 control women admitted to the same hospitals within six months of the cases; for cases diagnosed in the later period, interviews were also conducted with an additional 259 population-based controls selected by random digit dialing. Controls were matched to cases by age and race. Qualitative and quantitative findings were similar for the two studies. In the combined data, cases were more likely than their matched controls to have been nulliparous, to have undergone menarche at an early age, and to have refrained from using oral contraceptives. Menopause occurred slightly later for cases than for controls, but the differences were not statistically significant. Neither age at first term pregnancy (20 or more weeks gestation) nor number of term pregnancies was predictive of ovarian cancer risk. The protection afforded by oral contraceptive use was independent of parity, and it increased with increasing duration of use. There were no trends in risk with time since last oral contraceptive use or with time since first use, after adjustment for duration of use. These observations suggest that oral contraceptive use decreases risk for ovarian cancer, rather than merely indicates fertility, which may itself decrease risk of developing the disease. The authors combined reproductive characteristics and oral contraceptive use to estimate a woman's total duration of ovulation, which was positively associated with ovarian cancer risk (p less than 0.001 for trend). These observations support the concept that the greater the duration of ovulation or accompanying endocrinologic phenomena, the greater a woman's risk for ovarian cancer.

In a case-control study of 1868 breast cancer patients and 3391 control patients we searched for characteristics that predicted risk of breast cancer diagnosed before and after menopause. Common to increased risk of this disease in both periods of womanhood were: early menarche and late menopause; delayed marriage and first childbirth; more nulliparity or reduced gravidity and parity; reduced frequency of abortions; shorter overall child-bearing interval; more advanced education, higher socioeconomic status, and more contraceptive usage; and familial tendencies toward the disease. Breast cancer patients diagnosed before menopause were leaner than controls at age 20 and at time of diagnosis, but breast cancer risk in the postmenopausal period was related to increased weight-for-height at diagnosis and greater weight-for-height at diagnosis and greater weight gain since age 20. Postmenopausal breast cancer patients had a longer interval between first and second childbirths. Frequency and duration of the gravid state, inversely related to breast cancer risk, were largely dependent on contraceptive practices rather than unexplained infertility per se. Whether the breast cancer reaches diagnosis before or after menopause, the bulk of evidence examined here supports the view that it has a common cause and is subject to modifying influences over the long period of cancer latency.

A group of 3975 San Francisco longshoremen in cohorts classified annually by work activity (WA) was followed for fatal heart attack (FHA) over a 22-year period. In 57,632 person-years of follow-up, 410 men died from heart attack. A multi-factor logistic analysis was used to study work energy output adjusted for age, race, systolic blood pressure, smoking, body mass index, glucose intolerance and EKG status as predictors of FHA. After adjustment for these factors, men with a high WA of 7 kcal/min above basal metabolism at the beginning of a yearly follow-up period had a FHA rate about one-half the rate for men in the lowest WA category of 1 kcal/min above basal metabolism (p = 0.0003). In addition, the predictability of average WA during the four years preceding a yearly follow-up period was isolated to see if the lower risk associated with high WA derives from selective factors. After adjustment for factors described above and also for rate of change in WA used to represent selective factors, subjects with WA of 7 kcal/min still had about one-half the FHA rate observed for men at the lowest WA level (p = 0.0006). The findings from these and other phases of the analysis are consistent with the hypothesis that a substantial protective effect against FHA results from vigorous physical exertion.

In a 22-year followup of 3686 San Francisco longshoremen, a cohort analysis assessed job activity and six personal characteristics in relation to 395 fatal heart attacks. Four cohorts aged 35-44, 45-54, 55-64, and 65-74 in 1951 were studied annually for job shifts affecting energy output and for sudden or delayed death from heart attack by age 75. All subjects underwent multiphasic screening for heavy cigarette smoking, higher blood pressure, history of prior heart disease, obesity, abnormal glucose metabolism, and higher blood cholesterol. The first three of these characteristics added risk of fatal heart attack. The amount of risk varied in the four cohorts. Higher energy output on the job reduced risk of fatal heart attack, especially sudden death, in the two younger cohorts, where less active workers were at threefold increased risk. Lack of this effect in the two older cohorts could imply real differences in their work habits, such as being less energetic in heavy jobs or more energetic in light jobs than the younger cohorts. Or, before the study began, early deaths may have winnowed susceptibles from the two older cohorts. Combined low-energy output, heavy smoking, and higher blood pressure increased risk by as much as 20-fold. By elimination of these adverse influences, this population might have had an 88% reduction in its rate of fatal heart attack during the 22 years.