[Paleopsych] J. Sports Sciences: A natural history of athleticism, health and longevity

Premise Checker checker at panix.com
Sat Aug 20 20:30:17 UTC 2005

A natural history of athleticism, health and longevity
Journal of Sports Sciences, 1998, 16, S31± S45

[This is the crucial part, lifted from the article:

"Estimates of added life gained (to age 80) from a physically active way of 
living, as compared with a sedentary life, was from 1 to > 2 years. For each 
hour that men were physically active, they got to live that hour over again 
plus one or two additional hours (Paffenbarger et al., 1986)."

[Adobe 7 now converts PDF to txt, but, as you can see, it sometimes does a 
horrible job of it, like running across columns. E-mail me for the PDF to get 
the graphs and tables.]


Division of Epidemiology, Department ofHealth Research and Policy, Stanford 
University School of Medicine, Stanford,CA, and Department of Epidemiology, 
Harvard School of Public Health,Boston,MA,USA

Longitudinal observations on the sports play,social habits and health status of 
52,000 men who entered Harvard College or the University of Pennsylvania 
between 1916 and 1950 have afforded means of identifying causes of disease and 
death. These observations were then translated into the effect of sports and 
physical exercise on health and longevity. Student sports play in college 
predicted a decreased risk of developing cardiovascular disease (CVD) at least 
up to age 50 years. Questionnaire surveys showed physical exercise (sports 
play, walking and stair climbing) in middle age to be inversely related to the 
later development of CVD and early death. In a 10-year follow-up between 1962 
and 1972, alumni aged 35± 74 years who expended > 2000 kcal week -1 (8.4 MJ 
week -1) in such activities had a 25% reduced risk of CVD and death compared 
with less active men. But, the 'protective effect' of early athleticism waned 
unless a physically active life was maintained. In contrast, sedentary students 
who took up an active life were at a lower risk of CVD and death than former 
student athletes who gave up or reduced their physical activities in middle 
age. A total of 17,815 Harvard alumni aged 45± 84 years were followed from a 
1977 questionnaire survey through 1992, with 4399 deaths occurring. Death rates 
declined with increased levels of total activity (estimated in kilocalories), 
and declined also with increased intensity of effort measured as from none, to 
light,to moderately vigorous or vigorous sports play.Death rates at any given 
quantity of physical exercise were lower for men playing moderately intense 
sports than for less vigorous men. Over the age range, in the 16-year 
follow-up, Harvard alumni playing moderately vigorous or more intense sports 
gained 1.5 years by age 90 compared with less active men.

Keywords: athleticism, health, lifeway habits, longevity, physical activity.

[snip for bad conversion from PDF to txt. Acrobat 7 is doing a poor job!]

Despite these histories, scientific evidence suYcient to meet twentieth-century 
standards has been hard to come by. The health issues of adequate physical 
activity are still being explored and debated by clinicians, physiologists, 
dietitians, epidemiologists and others. If anything, the issues have become 
more complex and inconsistent. More machines have been devised to do our work 
and carry us about, and many of our jobs have become so 'soft' that we can use 
computers and robots to do them for us while we play bowls or watch television! 
Most heavy labour in particular has been eliminated. Never before has there 
been so little physical activity among so many people as in the Western World 

The adverse effects of these developments on public health became increasingly 
obvious as infectious diseases were brought under control, and cardiovascular 
disease rapidly gained ascendancy. An American physician from St. Louis, 
writing in German (Hammer, 1878; Nuland, 1994) in 1878, had described the 
relation of clinical symptoms to resulting pathological findings of myocardial 
infarction at necropsy of one of his patients. While Rudolf Virchow, Edward 
Jenner and others had previously described ischaemia in connection with 
arteriosclerosis, there seemed to be a rising tide of myocardial infarction 
that was first clearly defined clinically by J.B. Herrick in 1912 (Herrick, 
1912). In 1922, I. Sivertsen and A.W. Dahlstrom classified men in Minnesota 
according to occupational physical activity and observed that death rates 
declined with increasing physical activity in the job (Silversten and 
Dahlstrom, 1922). Yet in 1939, when O.F. Hedley reported that cardiovascular 
disease mortality in Philadelphia was higher for white-collar workers than for 
labourers, the mortality gap was not attributed to their differences in 
physical activity (Hedley, 1939).

The antecedents of modern-day exercise science and sports medicine may have 
begun in the mid-1800s at Oxford and Cambridge with concern for the health and 
longevity of rugby players and oarsmen, true 'gentleman-amateurs' (Park, 1997). 
The prevailing belief regarding vigorous exercise had not changed from the time 
of the ancients, namely that it was harmful (Morgan, 1873). However, studies of 
late nineteenth- and early twentieth-century oarsmen from English universities 
and those of the Eastern United States of America suggested the opposite; that 
is, the life expectancy of these men tended to exceed that of insured or 
general populations (Hartley and Llewellyn, 1939; Yamaji and Shephard, 1977). 
Studies of other sportsmen (e.g. track-and-field athletes; rugby, football and 
baseball players; cricketers; cross-country skiers, etc.) failed to show that 
they died earlier than non-athletes (Rook, 1954). Indeed, other studies even 
showed sportsmen to live longer (Pomeroy and White, 1958). Famed 
demographer-statistician Louis I. Dublin conducted an extensive review of 4976 
university athletes ± participants in baseball, football, crew, track and 
'minor sports' ± from 10 American colleges. In 1928, he concluded that, because 
of selection bias and an inability to adjust for confounding elements, it was 
'not easy to draw final conclusions, [but] the group of college athletes 
studied presented a favourable [longevity] picture' (Dublin, 1928, 1932).

The disease experience and length of life of men who were athletes in their 
youth had attracted interest over the decades despite the fact that, whatever 
the findings, interpretation was often uncertain, even confusing. The fact that 
men commonly adopted or avoided athletics suited to their size, strength, 
coordination, health and desires, whether self-selected or encouraged by 
others, made it diYcult to draw conclusions from epidemiological studies on the 
role of exercise, or its lack of a role, in the development of, say, 
cardiovascular disease or survival.

Some examples of these uncertainties follow: Michigan State University athletes 
born between 1855 and 1919 experienced no appreciable difference in longevity 
as compared with classmates (Montoye et al., 1956). Apart from violent deaths 
(especially military action and accidents), there were no differences in the 
causes of death. Extension of studies of these same Michigan State athletes 
versus classmates suggested that moderate levels of aerobic activity might 
reduce the risk of hypertension and cardiovascular disease, thus leading to 
longer life (Quinn et al., 1990). Tokyo University athletes were reported to 
live longer than their classmates, including medical graduates (Ishiko, 
1967a,b). Danish athletic champions born between 1880 and 1910 had a 39% lower 
risk of death than the general male population when aged 25± 50 years, but 
similar risks when aged 50± 64 and 65± 80 years; the rates of causes of death 
were also similar (Schnohr, 1971). Harvard and Yale University oarsmen 
matriculating between 1882 and 1902 lived 61± 4 years longer than randomly 
selected classmates (Prout, 1972). Finnish champion skiers born between 1845 
and 1910 lived 3± 4 years longer than the general male population (Karvonen et 
al., 1974). Studies of athletes who attended Harvard College between 1880 and 
1916 were followed for longevity and causes of death (Polednak, 1972). 'Major' 
university athletes were the shortest lived, by 1± 3 years, in relation to 
'minor' athletes and recreational athletes in each of three birth decades. 
Major athletes died from coronary heart disease (CHD) significantly earlier 
than others.

In most comparisons of university athletes with non- athletes, or of sportsmen 
versus less physically active individuals, the subjects self-selected 
themselves into differing levels of physical activity, yet almost no attention 
has been paid to the reason why they made the choice. Also, investigators have 
often been unable to account for potential confounding elements, both personal 
characteristics and other ways of living. A comparison of subsequent disease 
and death rates of athletes with non-athletic students or with insured lives, 
or with general populations, is insuYcient or often misleading. A further 
weakness of these early studies is that they concerned host and environmental 
conditions that existed as much as a century ago. Nevertheless, these 
shortcomings do not lessen the importance of examining critically the available 
data, the methods used, and their interpretation so as to uncover limitations 
that must be placed on the conclusions.

More recent observations of the effect of sports play on length of life were 
conducted on New Zealand international rugby payers as compared with New 
Zealand life tables for males. The life expectancy of All Black players was the 
same as for the general population, but non-Maori All Blacks lived almost 10 
years longer than Maori All Blacks (Beaglehole and Stewart, 1983).

Survival curves and standardized mortality ratios for US professional baseball 
players were very similar to those of the male general population (Waterbor et 
al., 1988). Dutch male long-distance skaters experienced a lower mortality rate 
than the age-comparable general population, but the data suggested that 
competitive racers experienced higher death rates than noncompetitive skaters 
(van Saase et al., 1990).

Finnish male world-class athletes were compared with a sample of age-and 
area-matched Finnish Defence forces conscripts. As contrasted with the 
reference population, endurance sportsmen lived an extra 5.7 years; team sports 
players, jumpers and short-distance runners, 4.0 years; and power sportsmen, an 
extra 1.6 years. But no differences were observed among the groups for maximum 
life span. A 1985 questionnaire survey did indicate that former athletes had a 
physically more active lifestyle and were healthier than the reference 
population (Sarna et al., 1993).

Recently, a national cross-sectional postal survey of male recreational runners 
in the USA assessed the weekly distance run as related to the prevalence of 
numerous personal characteristics and other lifestyle characteristics. Compared 
with runners who ran less than 16 km week -1, men who ran exceptionally long 
distances (> 80 km week -1) showed nearly a 50% reduction in clinically 
recognized hypertension, an 85% reduction in clinically low levels of 
high-density lipoprotein cholesterol, and a 2.5-fold increased prevalence of 
clinically defined high levels of high-density lipoprotein cholesterol. 
Estimated age-adjusted 10-year CHD risk was 30% lower in runners who averaged > 
64 km week -1 than in those who averaged < 16 km week -1 (Williams, 1997). 
Gradient responses to distance run were evident for each end-point measured. 
The implication drawn by the author of this study, and a second done on women 
runners with similar results, is that substantial health benefits occur at 
exercise levels that exceed current minimal guidelines and do not exceed a 
point of diminishing return to the health benefits of running any distance less 
than 80 km week -1 (Williams, 1996).

Data from these studies of recreational runners indicate a clear benefit of 
vigorous physical activity, and add to data showing a continuum of health 
benefits with increasing amounts and intensity of activity. Thus, while the 
message of moderate physical activity may be more palatable for those 
recalcitrant to exercise, vigorous-intensity activity should not receive less 
emphasis among those for whom such activity is not contraindicated. In today' s 
time-oriented world, a half- hour of vigorous exercise can expend as much 
energy as light to moderate exercise carried out for twice or three times as 

Modern-day exercise

The modern story of exercise and CHD began in 1949 when Jeremy N. Morris and 
his colleagues in London first began to understand how both vocational and 
leisure-time physical activity relate to cardiovascular fitness and risk of CHD 
(Morris et al., 1953). Initially, they found that highly active conductors on 
London' s double-decker omnibuses were at lower risk for CHD than bus drivers 
who worked sitting at the wheel; what disease the conductors did develop was 
less severe, and they were less likely to die following a heart attack. Morris 
et al. (1953) also found that postmen delivering mail on foot had lower rates 
of CHD than sedentary supervisors and telephonists, and a gradient appeared for 
positions of intermediate physical activity.

Since the early research by Morris and co-workers, further studies of 
occupational or leisure physical activity and CHD rates have concerned farmers 
and non-farmers in North Dakota, Iowa and Puerto Rico; American letter carriers 
and mail clerks; American and Italian railroad trackmen and clerks; Israeli 
kibbutzim workers in various occupations; San Francisco long- shoring cargo 
handlers and warehousemen; those insured by the Health Insurance Plan of New 
York; college students and alumni in various activities; residents of Iowa, 
California and Framingham, Massachusetts; American Cancer Society volunteers; 
Japanese-American men in Hawaii; Seventh Day Adventists in California; Multiple 
Risk Factor Intervention Trial men; and those from various regions of Denmark, 
Finland, Holland, Australia, New Zealand, Norway, Sweden and Germany. While 
these studies showed lower CHD risk with higher levels of exercise, some did 
not address potential confounding by cigarette smoking, diet, heredity, stress 
or other characteristics. Many of these studies have been summarized by Powell 
et al. (1987), Berlin and Colditz (1990), Blair and Connelly (1996), 
Paffenbarger and Lee (1996), Lee and Paffenbarger (1996) and, in particular, 
The Surgeon General' s Report on Physical Activity and Health (1996).

The College Alumni Health Study

In an early study of physical activity, athleticism and other personal 
characteristics related to the occurrence of specific chronic diseases in later 
life, pre-documented data provided material from records made years in advance 
of the clinical recognition of disease (Paffenbarger et al., 1966). The study 
was based on 52,500 men who were born between 1896 and 1934, entered Harvard 
College or the University of Pennsylvania between 1916 and 1950, and were 
followed continuously for cause-specific mortality from time of college 
entrance to present time. Multiple host and environmental characteristics of 
college students in the 1920s, 1930s and 1940s were considered to predict 
future coronary mortality. Individuals who died from coronary disease were less 
likely to have played sports in college than their surviving classmates. 
Habitual cigarette smoking and higher blood pressure levels were the most 
detrimental characteristics, while participation in college athletics was the 
most protective against early coronary mortality.

In compressing investigative time by using data recorded years earlier, later 
observations reaYrmed that student athleticism led to a substantial sparing 
effect from CHD mortality through about age 55 years, but not beyond. A graded 
dose± response existed; namely, the more athletically involved, the lower the 
risk of coronary mortality. In combination with cigarette abstention, lean body 
composition, lower blood pressure level and a favourable family history of 
longevity, a physically active lifestyle, especially vigorous exercise, 
predicted a low risk of early coronary mortality. Since habits of exercise, 
smoking and weight control are matters of personal choice, and their 
independent and combined in¯ uences are substantial, it becomes apparent that 
intervention strategies are likely to promote the public' s heath (Paffenbarger 
and Wing, 1969).

Further observations contributed to our current -understanding of the 
influences of physical energy expenditure on other lifestyle habits, on 
personal characteristics and on cardiovascular health, largely by excluding 
selection bias and eliminating confounding elements in showing that physical 
activity protected against heart attack. This phase of the study concerned 
nearly 17,000 Harvard College men assessed in the 1960s for walking, stair 
climbing, recreational activities and other characteristics, in addition to 
self-reported physician-diagnosed disease. Classified by these and previously 
documented characteristics from college days, men aged 35± 74 years were 
followed through 1972 for the occurrence of non-fatal and fatal CHD. 
Substantively, college athletes retained their low risk of both non-fatal and 
fatal CHD only if they remained physically active as adults (62% lower risk 
than for athletes who became sedentary). Also, sedentary students reduced their 
risk of coronary disease (by 61%) only if they adopted an active life since 
their college days. Both vigorous leisure-time physical activities and other 
less demanding activities in the 1960s, as measured on a linear scale of 
increasing energy expenditure, were independently associated with a subsequent 
lower risk of heart attack. At any given level of energy expenditure, the risk 
of heart -attack was significantly lower with vigorous activities than with 
more casual activities. Finally, the study -showed that alumni who expended 
less than 2000 kcal week -1 (8.4 MJ week -1) in leisure-time activities, who 
smoked cigarettes and were hypertensive were nearly eight times more likely to 
have a heart attack in mid- life than alumni with none of these characteristics 
(Paffenbarger et al., 1978).

The next phase of the study extended observations on men who entered college in 
the years 1916± 50 and were resurveyed in the 1960s for their personal 
athleticism in relation to both morbidity from CHD and to specific causes of 
death. Follow-up through 1978 reaYrmed that only contemporary exercise, not 
college sports play, protected against cardiovascular disease. Walking, 
climbing and recreational activities totalling 2000 or more kcal week -1 (8.4 
MJ week -1) reduced the risk of both fatal CHD and stroke in a gradient 
fashion. The benefits of exercise from a physically active lifestyle were 
independent of normotension, but not smoking, stable body weight or an absence 
of family history of cardiovascular disease. The community benefit from 
habitual physical exercise, measured as population attributable risk, was 
substantially stronger than corresponding figures for each of these other 
characteristics. Although inversely related to both cardiovascular and 
respiratory mortality, exercise level was less related to cancer or other 
causes of death (Paffenbarger et al., 1984).

Next, we made quantitative estimates of influences on length of life related to 
physical activity (again expressed in kilocalories of weekly energy 
expenditure) and other lifeway characteristics. In the follow-up of men from 
1962 to 1978, the data showed that exercise benefited all types: old and young; 
large and small; hypertensive and normotensive; men who have never smoked, ex- 
smokers, and light, heavy or very heavy smokers. Death rates declined steadily 
with increased levels of energy expenditure, at least up to 3500 kcal week -1 
(14.6 MJ week -1). Rates were one-quarter to one-third lower among men who 
expended 2000 or more kilocalories (8.4 MJ) than among less active individuals. 
Estimates of added life gained (to age 80) from a physically active way of 
living, as compared with a sedentary life, was from 1 to > 2 years. For each 
hour that men were physically active, they got to live that hour over again 
plus one or two additional hours (Paffenbarger et al., 1986).

Favourable changes in personal characteristics and habits, including the 
adoption of a physically active way of living between college years and middle 
age, had been shown to reduce the risk of cardiovascular disease. Now, men aged 
35-84 years from the same large cohort of former college students were studied 
for the adoption or maintenance of physical activity and other optional lifeway 
habits between 1962 and 1977 for the influence of such changes on 
cardiovascular and total mortality over the subsequent 8-9 years. Men who 
increased their activity levels to 1500 kcal week -1 (6.3 MJ week -1) or who 
adopted moderately vigorous sports play habits between the 1960s and 1970s, 
lowered their risk of premature death by 25% in the ensuing 12 years. Quitting 
cigarette smoking, maintaining normal blood pressure and avoiding obesity were 
separately associated with lower mortality rates from CHD and all-cause 
mortality. These volitional changes represent a kind of intervention, one of 
the epidemiological canons generally assumed to be necessary to establish 
cause- and-effect relations between host or environmental characteristics and 
disease outcome. The estimates of years of life gained to age 85 years by men 
who had reported favourable changes in levels of energy expenditure and these 
other lifeway habits were substantial when contrasted to survival rates for men 
whose style of living or personal characteristics remained detrimental. 
Estimates from later observations predicted that men who took up moderately 
vigorous sports play, quit cigarette smoking and remained normotensive enjoyed 
4-10 more years of active living (an average of 7 years) than those who did 
not. Of special significance, these findings add further support to the 
hypothesis that sedentary living, cigarette smoking, hypertension and obesity 
independently cause specific chronic diseases and reduce longevity 
substantially (Paffenbarger et al., 1993, 1994).

Contemporary studies

Most recent observations have examined the experiences of 17,815 Harvard 
College alumni aged 45-84 years who, in 1977, had responded to questionnaires 
on their health habits and health status. Again classified by walking, stair 
climbing, recreational activities, cigarette smoking habit, blood pressure 
status, body mass index, alcohol consumption, history of parental mortality and 
chronic disease occurrence, the alumni were then followed for 16 years, through 
1992 or to age 90 years, for mortality. The experiences represent 254,636 man 
years of observation, during which 4399 men died. Mortality rates were computed 
for each lifeway pattern, standardized for age and all other pertinent patterns 
to account for potential confounding influences. To examine the influence of 
ageing, some analyses are presented in age-specific groups in 1977, with 
follow-up for the 16-year period through 1992. It is important to recognize 
that comparisons were made among separate age groups, each followed for 16 
years, and not as trackings of one group as it passed through successive ages.

Age trends of lifeways

To examine the influence of lifeway patterns on mortality at various stages of 
the ageing process, we categorized the Harvard population into four 10-year age 
groups: 45-54, 55-64, 65-74 and 75-84 years in 1977. Baseline characteristics 
are given in Table 1, together with measures of median attained ages during 
follow-up. The age of survivors through 1992 ranged 61 years to censure at 90 

The frequencies of lifeway patterns in these separate age groups are also given 
in Table 1. The proportions of men who played moderately vigorous sports 
declined steadily and substantially with advancing age, as did the proportions 
who climbed stairs equalling > 55 floors, and who expended > 2000 kcal week -1 
(8.4 MJ week -1). Although the proportions of men playing light sports 
increased with age, total sports play declined from 80.2% in the youngest men 
to 73.0%, 59.2% and 44.0% in successively older age groups.

No consistent age trend was noted for patterns of walking, blood pressure 
status, or reported quantities of alcohol consumption. With increasing age, 
lower proportions of men reported smoking cigarettes, fewer were overweight 
(body mass index > 26 units) or re- ported a history of early parental 
mortality, and more were free of self-reported physician-diagnosed chronic 
disease. These last four cited patterns presumably related in part to previous 
(before 1977) mortality rates; that is, some men who had exhibited one or more 
of these lifeways may have died and not been available to report when these 
observations began. Nevertheless, interaction of all of these lifeway patterns 
could have influenced morbidity and mortality rates in all age groups during 
the follow-up interval of 16 years.

Table 1 Baseline characteristics of Harvard alumni and frequencies (percent of 
man-years) of lifeway patterns in 1977 [omitted]

Lifeways and mortality

The physical activity index is distributed into arbitrary levels at < 1000, 
1000± 1999 and > 2000 kcal week -1, dividing the population into roughly equal 
man-years of follow-up (Table 2). Using < 1000 kcal week -1 (4.2 M J week -1) 
as the comparison group, men in the middle group had a 19% lower risk of death 
during the interval 1977± 92, and men in the most active group a 25% lower 
risk. The gradients of lower risk for both walking and stair climbing were less 
steep but both activities implied reduced mortality rates with greater 
quantities of these activities. With absence of sports play representing the 
standard of reference, men playing only light sports (activities requiring < 
4.5 METs intensity; 10% of men) were at a 16% lower risk of death, while men 
participating in moderately vigorous sports play (> 4.5 METs; 62% of men) were 
at a 27% lower risk.

We distributed the population into arbitrary or customary levels of lifeways 
for other comparisons. The dangers of cigarette smoking are evident, with light 
smokers having a one-third lower risk and non-smokers a nearly 50% lower risk 
of death than heavy smokers. Normotensive men experienced only 75% the 
mortality during follow-up as hypertensive men. Both body mass index and 
alcohol consumption patterns displayed U-shaped distributions of mortality when 
examined in approximately thirds, with mid-range levels experiencing the lowest 
rates, suggesting that both lower and higher levels of each pattern carried 
higher risk.

Early parental mortality was only a minimal indicator of increased alumni 
mortality in this interval; but the presence of a potentially lethal chronic 
disease (representing CHD, stroke, diabetes mellitus, chronic obstructive lung 
disease or cancer; 18% of men) was a strong predictor of death in these men, 
doubling the mortality over that for men not reporting any of these ailments.

Not shown is the proportion of deaths (i.e. 13.2%) that might have been avoided 
between 1977 and 1992 if all alumni had expended > 2000 kcal week -1 (8.4 M J 
week -1). Corresponding figures for playing moderately vigorous sports would 
have been 13.9%; for not smoking, 9.3%; for remaining normotensive, 7.9%; and 
for being free of chronic disease, 19.7%. Such attributable risk estimates 
comprise a public health perspective to the Harvard alumni population.

Age trends of mortality

Table 3 shows the age-specific relative risks of death according to the 
distributions of lifeway patterns presented in Table 2. In general, these 
gradients for the separate age groups ran parallel to those for age- 
standardized relative risks, there being some exceptions that may have useful 
implications for preventive inter-were unrelated to delayed mortality; only 
moderately vention as men get older. Perhaps most notable with vigorous sports 
play was inversely (and significantly) respect to physical activity, walking, 
stair climbing and related to death in the 16-year follow-up period for total 
energy expenditure in men aged 45± 54 years those youngest men. In addition, an 
inverse relation with death for body mass index was evident only for the At the 
oldest end of the age range, cigarette smoking youngest men. Other age groups 
did not show this was not significantly related to death, again perhaps 
relation of a higher death risk for the more overweight because smoking had 
taken a considerable toll before men. this population reached age 75 years. 
Also, among men aged 75± 84 years in 1977, whose median attained age would have 
reached 88 years by 1992, and in whom the frequencies of favourable lifeways 
would have declined substantially relative to those in younger men, signifi 
cant relations with mortality during follow-up were not evident from stair 
climbing, body mass index, alcohol consumption or early parental mortality.

Table 2 Rates and relative risks of deatha among Harvard alumnib from 1977 to 
1992, according to lifeway patterns in 1977 [omitted]

Table 3 Age-specific relative risks of deatha among Harvard alumni from 1977 to 
1992, according to lifeway patterns in 1977 [omitted]

Youth versus adult sports

It might be asked if student athleticism per se is related to premature 
mortality in a population that had been classified by contemporary sports play 
status when aged 45± 84 years of age. For this purpose, 11,286 Harvard alumni 
(a sample of the larger population) were followed over 163,935 man-years, 
during which 2377 deaths occurred. Mortality rates were computed for the next 
16 years. Figure 1 shows standardized rates and relative risks of death by 
student and alumni physical activity patterns. Considering college activity as 
university (intercollegiate), recreational (intramural) and non-athletic 
(minimal or none), a cross-tabulation is shown with alumni (contemporary) 
sports play at three levels: none, light only (< 4.5 METs) and moderately 
vigorous or vigorous (> 4.5 METs). Thus risks are given for men who were less 
physically active as students but more active as alumni, or vice versa, or who 
showed little or no change in activity level.

While contemporary athleticism coded as none, light sports only, or moderately 
vigorous to vigorous sports play, carried a significant decline in death rates 
that favoured increased activity, student sports play did not. There was no 
significant trend in death rates among men classified by college sports play, 
except for former university athletes who had given up sports play since 1977. 
They experienced a significantly higher death rate through 1992 than former 
recreational athletes or non-athletes who played no sports as alumni.

Combined lifeways

To assess any confounding influences among major precursors of mortality, and 
with a special interest in the protective effect afforded by moderately 
vigorous and vigorous sports play, Fig. 2 presents rates and relative risks of 
paired combinations of such sports play with four other precursors, among the 
17,815 Harvard alumni, 4399 of whom died during follow-up. As shown in the 
stereograms, each of the lifeway characteristics is associated with mortality 
in the 16-year follow-up period. Moderately vigorous sports players who do not 
smoke are at a 54% lower risk of death than their opposites. Normotensive 
sports players are at a 42% lower risk, lean players a 26% lower risk, and 
alumni without chronic disease a 60% lower risk than men who were more active 
and healthy.

Changes in physical activity

Next, we tested the hypothesis that favourable changes in physical activity 
might reduce the risk of premature mortality by computing death rates and 
plotting survival curves (1977± 92) for the 14,782 alumni who had reported on 
their physical activities in 1962 or 1966 and again in 1977 (Fig. 3). A total 
of 3528 men died in 212,410 man-years in the 16-year follow-up. Men who 
increased or decreased their energy expenditure between the 1960s and 1977 by 
less than 250 kcal week -1 (1.05 MJ week -1) were placed in an 'unchanged' 
category and used as a reference. Compared with this standard, gradient 
reductions in mortality became evident with increasing levels of physical 
activity, and gradient increases in mortality with decreasing levels of 
activity. The patterns were consistent within each of the four age groups 
studied, the three oldest groups being presented. Note in particular that even 
elderly men would seem to benefit from becoming more physically active.

Figure 1 Rates and relative risks of death per 10,000 man-years among Harvard 
alumni aged 45± 84 years, between 1977 and 1992, according to combinations of 
athleticism in college and sports play in 1977. Rates and relative risks were 
standardized for age, cigarette smoking habit, blood pressure status, body mass 
index, alcohol consumption, early parental mortality and chronic disease 
(except for given characteristic) (11,286 men; 2377 deaths; 163,935 man-years).

Figure 2 Rates and relative risks of death per 10,000 man-years among Harvard 
alumni aged 45± 84 years, between 1977 and 1992, according to paired 
combinations of moderately vigorous sports play and (A) cigarette smoking 
habit, (B) hypertension, (C) body mass index and (D) chronic disease status in 
1977. Rates and relative risks were standardized for age, cigarette smoking 
habit,blood pressure status,body massindex,alcoholconsumption, early 
parentalmortality and chronic disease (except for given characteristic) (17,815 
men; 4399 deaths; 254,636 man-years).

Quantity versus intensity of physical activity

To explore the relative importance of the quantity as opposed to the intensity 
of physical activity in deferring mortality, the college alumni were first 
separated into two groups: those whose activities might require> 4.5 METs of 
sustained effort (76% of the total) and those whose activities might be 
expected to require less. In these analyses, standardized mortality rates for 
1977± 88 were computed for 14,787 alumni aged 45± 84 years in 1977. A total of 
2206 men had died in 123,285 man- years of observation. The quantity of 
activity amassed from walking and climbing stairs was arbitrarily assigned to 
the less vigorous group. Death rates were then plotted against the quantity of 
energy expended in kcal week -1 (Fig. 4). The data indicate that, with 
increasing quantities of energy expenditure, both light and moderately vigorous 
effort were related to lower rates of mortality. But at any given quantity of 
energy expenditure, rates were lower for moderately vigorous activities than 
for activities requiring less intensity of effort.

Figure 3 Survivorship curves and relative risks of death, among Harvard alumni 
aged 45± 84 years, between 1977 and 1992, according to change or lack of change 
in physical activity index (kcal week -1) between 1962 or 1966 and 1977. Curves 
and relative risks were standardized for age, cigarette smoking habit, blood 
pressure status, body mass index, alcohol consumption, early parental mortality 
and chronic disease (14,787 men; 3528 deaths; 212,410 man-years).

The two vertical marks on each curve of Fig. 4 (tertile ticks) separate men 
into thirds according to man-years of experience. From these graphic displays, 
the mean death rate was about 170 per 10,000 for the less vigorous, compared 
with about 120 per 10,000 for the more vigorous, or 20% lower. Of further 
importance in this time-oriented era, at some given quantity of moderately 
vigorous activity, say 1500 kcal of energy expended per week (6.3 MJ week -1), 
where the death rate was about 120 per 10,000, the same low death rate was 
achieved among the less vigorously active men only at about 3000 kcal week -1 
(12.6 MJ week -1). This difference might be equated to 3± 4 h of jogging versus 
6± 8 h of regular walking.

Figure 4 Death rates among Harvard alumni aged 45± 84 years, between 1977 and 
1988, according to quantity and intensity of physical activity. Rates were 
standardized for age, cigarette smoking habit, blood pressure status, body mass 
index, alcohol consumption, early parental mortality and chronic disease 
(14,787 men; 2206 deaths; 123,285 man-years).

Added life

The mortality and survival experiences of the Harvard alumni during the 16-year 
follow-up period were used to develop estimates of years of added life gained 
(up to age 90) from favourable as opposed to detrimental lifeway patterns 
(Table 4). The results are tabulated for 10-year age groups and for selected 
habits or characteristics at the start of the follow-up. Estimates for 
individual lifeway patterns are standardized for age (5year groups) and each of 
the other patterns listed. The physical activity index shows the estimated 
years of survival gained by men expending > 2000 kcal week -1 (8.4 MJ week -1) 
in the combination of self-reported walking, stair climbing and playing sports, 
as compared with those expending less energy in those activities. The estimated 
added years of life from a more active life was 1.19 years. Walking > 15 km, 
> 55 floors and playing moderately vigorous sports each week versus
walking and climbing less and not playing such sports each week were estimated 
to yield 0.43, 0.70 and 1.44 added years of life, respectively. Even the oldest 
men, all of whom were followed until death or age 90, experienced at least an 
extra 6 months of life by being active as compared with being sedentary. 
Important, too, was the finding that these physically active alumni, as 
contrasted with their less active counterparts, reported that they 'felt 
younger than their years' , and were more likely to be 'feeling fine and 
enjoying life' (Paffenbarger et al., 1993). The additional life gained as a 
result of not smoking cigarettes versus smoking was 2.95 years, and 1.56 years 
for remaining normotensive as opposed to having become hypertensive. Men who 
had avoided chronic disease as of the 1977 questionnaire assessment gained 3.96 
years more than men who manifested at least one of these ailments by that time. 
Considering the range of starting ages (i.e. 45± 84 years) and ageing through 
90 years, these are truncated analyses and the difference in survival between 
groups being compared might be expected to become somewhat narrower as the life 
experience of this population approaches its completion.

Table 4 Added years of life gained (to age 90) from favourable lifeway patterns 
among Harvard alumni as estimated from standardized death rates, 1977± 92a 

These actuarial estimates predict what the logical extension or outcome might 
be sometime in the future. The mortality and survival experience of these 
alumni during a 16-year follow-up (i.e. 1962± 78) provided estimates of added 
years to age 80 from an active life- way (Paffenbarger et al., 1986). The 
observed mortality differentials for the follow-up period (i.e. 1977± 92) have 
shown that these earlier predictions were remarkably accurate to date.


Extended longevity is the result of a number of interacting favourable 
influences, some susceptible to optional adjustment and others not. The 
findings reported here on college alumni represent only a part of the overall 
picture, but they suggest a protective effect of moderately vigorous exercise, 
abstention from cigarette smoking and maintenance of normotension against all- 
cause mortality in all age groups studied, and therefore an indication of 
additional years of life expectancy.

To the extent that these alumni increased their exercise, reduced their smoking 
and received treatment for hypertension since 1977, when these lifeway patterns 
were assessed, such changes would minimize the importance of the patterns 
reported here. Thus, the true strength of the protective effect of these 
lifeways would be even larger than observed. Nevertheless, the estimates of 
years gained reflect the self-reported experience of college men, already 
recognized as long- lived, whose level of aZuence and social behaviour might 
differ from other populations, and this might invite caution against 
unwarranted extrapolation to other populations.

The opportunity to examine long-range parallels between findings based on 
experience of student days and the alumni follow-up gave a different dimension 
to this investigation. If it is postulated that university sports play 
reflects, at least in part, a selective attribute of personal health (e.g. 
cardiovascular fitness), the present findings show that such selection alone is 
insuYcient to explain delayed death in later adult (alumni) years. Alumni who 
had not been especially athletic as students, but rated high physical activity 
status at middle and advanced age, were at lower risk of mortality than former 
athletes whose later exercise level did not include moderately vigorous sports 

The medical treatment of such chronic diseases as cardiovascular ailments and 
diabetes mellitus, and the management of such intermediate variables as 
abnormal lipoprotein profiles and impaired glucose tolerance, may defer 
mortality and extend longevity. Yet, the effect of adequate physical activity 
is partly independent of these influences and it counteracts many detrimental 
variables through metabolic and other processes. Even when other means of 
health promotion and disease prevention are used to increase longevity, the 
relevance of adequate physical activity is likely to remain.

Physical activity is a positive and pervasive element of health maintenance and 
disease prevention, and lifetime extension for everyone, including the elderly 
and aged persons. Habitual physical activity can influence fitness, which in 
turn may modify the level of physical activity of the individual. Other 
lifeways and personal characteristics, physical and social environments, and 
genetic endowment also affect these complex relations. In the late James A. 
Michener' s 1976 book entitled Sports in America, he wrote: 'I defend sports as 
a means of obtaining exercise pleasurably. I am more impressed with sports as a 
developer of health than as a developer of character, and I want to see them 
prosper for health reasons, because I know of no other human activity which so 
well serves that purpose' . Perhaps Michener' s ideas have not changed much 
from those of Hippocrates and Galen many centuries earlier.


This study was supported by US Public Health Service Research grants HL34174 
from the National Heart, Lung and Blood Institute, and CA44854 from the 
National Cancer Institute, National Institutes of Health. This is Report Number 
LXVIII in a series on chronic disease in former college students.


Beaglehole, R. and Stewart, A. (1983). The longevity of international rugby 
players. New Zealand Medical Journal, 96, 513± 515.

Berlin, J.A. and Colditz, G.A. (1990). A meta-analysis of physical activity in 
the prevention of coronary heart disease. American Journal of Epidemiology, 
132, 612± 628.

Berryman, J.W. (1989). The tradition of the 'six things nonnatural': Exercise 
and medicine from Hippocrates through ante-bellum America. In Exercise and 
Sports Science Reviews, Vol. 17 (edited by J.O. Holloszy), pp. 515± 559. 
Baltimore, MD: Williams & Wilkins.

Berryman, J.W. (1995). Out of Many, One: A History of the American College of 
Sports Medicine. Champaign, IL: Human Kinetics.

Blair, S.N. and Connelly, J.C. (1996). How much physical activity should we do? 
The case for moderate amounts and intensities of physical activity. Research 
Quarterly for Exercise and Sport, 67, 193± 205.

Dublin, L.I. (1928). Longevity of college athletes. Harper' s Monthly, 157, 
229± 238. Dublin, L.I. (1932). College honor men longer lived. Statistical 
Bulletin, Metropolitan Life Insurance Co., 13, 5.

Guy, W.A. (1843). Contributions to a knowledge of the influence of employments 
upon health. Journal of the Royal Statistical Society, 6, 197± 211.

Hammer, A. (1978). Ein Fall von thrombotischen Verclusse einer der 
Kranzarterien des Herzens. Weiner Medizinsche Wochenschrift, 2 February.

Hartley, P.H.S. and Llewellyn, G.F. (1939). The longevity of oarsmen: A study 
of those who rowed in the Oxford and

Paffenbarger and Lee

Cambridge boat race from 1829± 1928. British Medical Journal, 1, 657± 662.

Hedley, O.F. (1939). Analysis of 5,116 deaths reported as due to acute coronary 
occlusion in Philadelphia, 1933± 1936. US Weekly Public Health Reports, 54, 

Herrick, J.B. (1912). Clinical features of sudden obstruction of the coronary 
arteries. Journal of the American Medical Association, 59, 2015± 2020.

Ishiko, T. (1967a). Aerobic capacity and external criteria of performance. 
Canadian Medical Association Journal, 96, 746± 751.

Ishiko, T. (1967b). Participation in athletics: Commentary. Canadian Medical 
Association Journal, 96, 821.

Karvonen, M.J., Klemola, H., Virkaj„rvi, J. and Kekkonen, A. (1974). Longevity 
of endurance skiers. Medicine and Science in Sports, 6, 49± 51.

Lee, I.-M. and Paffenbarger, R.S., Jr. (1996). Do physical activity and 
physical fitness avert premature mortality? In Exercise and Sports Science 
Reviews, Vol. 24 (edited by J.O. Holloszy), pp. 135± 171. Baltimore, MD: 
Williams & Wilkins.

Lyons, A.S.and Petrucelli,R.J.(1978). Medicine: An Illustrated History.New 
York:Abradale Press/Harry N. Abrams,Inc.

MacAuley, D. (1994). A history of physical activity, health and medicine. 
Journal of the Royal Society of Medicine, 87, 32± 35.

Michener, J.A. (1976). Sports in America. New York: Random House.

Montoye, H.J., Huss, W.D.V., Olson, H., Hudec, A. and Mahoney, E. (1956). Study 
of the longevity and morbidity of college athletes. Journal of the American 
Medical Association, 17, 1132± 1134.

Morgan, J. (1873). University Oars, Being a Critical Enquiry into the After 
Health of Men Who Rowed in the Oxford and Cambridge Boat-Race from the Year 
1829± 1869, Based on the Personal Experiences of the Rowers Themselves. London: 

Morris, J.N., Heady, J.A., RaZe, P.A.B., Roberts, C.G. and Parks, J.N. (1953). 
Coronary heart disease and physical activity of work. Lancet, 2, 1053± 1057, 
1111± 1120.

Nuland, S.B. (1994). How We Die: Reflection on Life's Final Chapter. New York: 
Alfred A. Knopf.

Paffenbarger, R.S., Jr. and Lee, I.-M. (1996). Exercise and fitness. In 
Prevention of Myocardial Infarction (edited by

J.E. Manson, P.M. Ridker, J.M. Gaziano and C.H. Hennekens), pp. 172± 202. New 
York: Oxford University Press. Paffenbarger, R.S., Jr. and Wing, A.L. (1969). 
Chronic disease in former college students: X. The effects of single and 
multiple characteristics on risk of fatal coronary heart disease. American 
Journal of Epidemiology, 90, 527± 535.

Paffenbarger, R.S., Jr., Wolf, P.A., Notkin, J. and Thorne,

M.C. (1966). Chronic disease in former college students: 1. Early precursors of 
fatal coronary heart disease. American Journal of Epidemiology, 83, 314± 328. 
Paffenbarger, R.S., Jr., Wing, A.K. and Hyde, R.T. (1978). Chronic disease in 
former college students:

XVI. Physical activity as an index of heart attack risk in college alumni. 
American Journal of Epidemiology, 108, 161± 175.

C.H. (1984). Chronic disease in former college students: XXV. A natural history 
of athleticism and cardiovascular health. Journal of the American Medical 
Association, 252, 491± 495.

Paffenbarger, R.S., Jr., Hyde, R.T., Wing, A.L. and Hsieh, C.-C. (1986). 
Physical activity, all-cause mortality and longevity of college alumni. New 
England Journal of Medicine, 314, 605± 613.

Paffenbarger, R.S., Jr., Hyde, R.T., Wing, A.L., Lee, I.-M., Jung, D.L. and 
Kampert, J.B. (1993). Chronic disease in former college students: XXXVII. The 
association of changes in physical-activity level and other lifestyle 
characteristics with mortality among men. New England Journal of Medicine, 328, 
538± 545.

Paffenbarger, R.S., Jr., Kampert, J.B., Lee, I.-M., Hyde, R.T., Leung, R.W. and 
Wing, A.L. (1994). Chronic disease in former college students: LII. Changes in 
physical activity and other lifeway patterns influencing longevity. Medicine 
and Science in Sports and Exercise, 26, 857± 865.

Park, R.J. (1997). High protein diets, 'damaged hearts' , and rowing men: 
Antecedents of modern sports medicine and exercise science, 1867± 1928. In 
Exercise and Sports Sciences Reviews, Vol. 25 (edited by J.O. Holloszy), pp. 
137± 169. Baltimore, MD: Williams & Wilkins.

Polednak, A.P.(1972).Longevity and cardiovascular mortality among former 
college athletes. Circulation, XLVI, 649± 654.

Pomeroy, W.C. and White, P.D. (1958). Coronary heart disease in former football 
players. Journal of the American Medical Association, 167, 711± 714.

Powell, K.E., Thompson, P.D., Caspersen, C.J. and Kendrick,

J.S. (1987). Physical activity and the incidence of coronary heart disease. 
Annual Review of Public Health, 8, 253± 287. Prout, C. (1972). Life expectancy 
of College oarsmen. Journal of the American Medical Association, 220, 1709± 
1711. Quinn, T.J., Sprague, H.A., Van Huss, W.D. and Olson, H.W. (1990). 
Caloric expenditure, life status and disease in former male athletes and 
non-athletes. Medicine and Science in Sports and Exercise, 22, 742± 750.

Ramazzini, B. (1700/1964). Disease of Workers (in Latin) (translated by W.C. 
Wright). New York: Hafner.

Rook, A. (1954). An investigation into the longevity of Cambridge sportsmen. 
British Medical Journal, 1, 657± 662.

Sarna, S., Sahi, T., Koskenvuo, M. and Kaprio, J. (1993). Increased life 
expectancy of world class male athletes. Medicine and Science in Sports and 
Exercise, 25, 237± 244.

Schnohr, P. (1971). Longevity and causes of death in male athletic champions. 
Lancet, 18 December, pp. 1364± 1366.

Silversten, I. and Dahlstrom, A.W. (1922). The relation of muscular activity to 
carcinoma: A preliminary report. Journal of Cancer Research, 6, 365± 378.

Smith, E. (1864). Report on the sanitary conditions of tailors in London. In 
Report of the Medical OYcer, pp. 416± 430. London: The Privy Council.

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

van Saase, J.L.C.M., Noteboom, W.M.P. and Vandenbouch‚ ,

J.P. (1990). Longevity of men capable of vigorous physical activity exercise: A 
32 year follow-up of 2,259 participants in the Dutch eleven cities ice skating 
tour. British Medical Journal, 301, 1409± 1411. Waterbor, J., Cole, P., 
Delzell, E. and Andjelkovich, D. (1988). The mortality experience of 
major-league baseball players. New England Journal of Medicine, 318, 1278± 

Williams, P.T. (1996). High-density lipoprotein cholesterol and other risk 
factors for coronary heart disease in female runners. New England Journal of 
Medicine, 334, 1298± 1303.

Williams, P.T. (1997). Relationship of distance run per week to coronary heart 
disease risk factors in 8,283 male runners: The National Runners' Health Study. 
Archives of Internal Medicine, 157, 191± 198.

Yamaji, K. and Shephard, R.J. (1977) Longevity and causes of death of athletes. 
Journal of Human Ergology, 6, 15± 27.

More information about the paleopsych mailing list