WILLIAMS, PAUL T.
Benign prostatic hyperplasia (BPH) is a common malady among older men that often produces frequent or uncomfortable urination. These urinary symptoms arise from pressure on the urethra due to greater prostate mass and decreased prostatic smooth muscle tone due to reduced sympathetic nervous activity (14). Bothersome urinary tract symptoms affect over 40% of older men (14). In 2000, BPH accounted for 105,000 hospitalizations, 117,000 emergency room visits, and 4.4 million office visits (30). Reported risk factors for BPH include older age, functioning testes, total and central obesity, hypertension, diabetes, low alcohol intake, nonsmoking status, and possibly African American race (9,11-14,18,21,23,25). Whereas a few studies suggest that physical activity reduces the risk for BPH (10,24) or lower urinary tract symptoms (26,27,33), many report no inverse association (11,18,20,21).
To date, the most compelling evidence relating BPH to physical activity derives from the 30,634 men of the Health Professionals Follow-up Study followed prospectively for 6 yr (24). Their data showed that the odds for BPH were 24% lower for the highest quintile of physical activity than the lowest quintile, and that men who walked 2 to 3 h·wk−1 had 27% lower risk for BPH than those reporting zero walking hours. They also reported that vigorously intense physical activity was less effective in preventing BPH than low- or moderate-intensity activities. Other studies claim BPH risk increases at higher doses of physical activity (12,18). These conflicting results present a compelling need to test the relation between BPH and vigorous exercise. Specifically, there seemed to be no a priori reason to expect that an inverse relationship between BHP risk and physical activity would not extend throughout higher doses of vigorous exercise, as has been observed for other diseases and their risk factors (34,39). Moreover, reductions in disease risk due to exercise are often associated with cardiorespiratory fitness (39), which is more effectively increased by vigorous then by moderate exercise (28) and more effectively increased by more-intense than by less-intense vigorous exercise (31).
The National Runners' Health Study was specifically designed to assess the dose-response relationship between health and running distance (km·wk−1) and 10-km race performance time (an indicator of cardiorespiratory fitness) in men and women who meet or substantially exceed current minimum physical activity guidelines (34-40). In addition to providing independent confirmation of a BPH-physical activity relationship, these analyses extend the activity range to higher doses than previously considered. To our knowledge, this is the only article to date to assess the independent effects of a vigorous physical activity and a fitness-related measure on BPH and to relate BPH to changes in a physical activity.
The design and the methods of the National Runners' Health Study are described elsewhere (34-40). Briefly, a two-page questionnaire was distributed nationally to runners identified through a running magazine subscription list and among participants of foot race events. The questionnaire solicited information on demographics, running history, weight history, smoking habits, prior history of heart attack and cancer, and medications for blood pressure, thyroid, cholesterol, and diabetes. Participants were recruited between 1991 and 1994 (primarily 1993), and follow-up questionnaires were obtained between 1999 and 2002. The study protocol was approved by the University of California Berkeley Committee for the Protection of Human Subjects, and all participants signed committee-approved informed consents.
Participants reported whether a physician had told them they had benign prostate enlargement since their baseline questionnaire and reported the year of diagnosis. Diagnoses dated the same year as the baseline survey or earlier were excluded in order that the end points represented only newly diagnosed incidence of disease. Running distances were reported in usual miles run per week. Although other leisure-time physical activities were not recorded for this cohort, data from runners recruited after 1998 (when the question was added to the survey) show running represents (± SD) 91.5 ± 19.1% of all vigorously intense activity in those men and 73.5 ± 23.7% of their total leisure-time physical activity. Previously, we reported strong correlations between repeated questionnaires for self-reported running distance (r = 0.89) (36), between self-reported and clinically measured height (r = 0.96) and weight (r = 0.96) (36), and for self-reported running distance versus self-reported BMI and body circumference in cross-sectional analyses (34,40). Eighty percent of the 54,956 participants of the National Runners' Health Study provided follow-up information or were known deceased. For this report, participant's best 10-km race during the previous 5 yr (reported as finish time in minutes) as speed in meters per second was included as a performance-based measure associated with cardiorespiratory fitness (3,8,17).
Intakes of meat, fish, and fruit were based on the questions "During an average week, how many servings of beef, lamb, or pork do you eat," "…servings of fish do you eat," and "…pieces of fruit do you eat." Alcohol intake was estimated from the corresponding questions for 4-oz. (112 mL) glasses of wine, 12-oz (336 mL) bottles of beer, and mixed drinks and liqueurs. Alcohol was computed as 10.8 g per 4-oz glass of wine, 13.2 g per 12-oz bottle of beer, and 15.1 g per mixed drink. Correlations between these responses and values obtained from 4-d diet records in 110 men were r = 0.65 for alcohol, r = 0.46 for red meat, r = 0.38 for fruit, and r = 0.19 for fish intake. These values agree favorably with published correlations between food records and more extensive food frequency questionnaires for red meat (r = 0.50), wine (r = 0.66), beer (r = 0.70), and mixed drinks (r = 0.72), somewhat less favorably for fruit intake (r = 0.50), and less favorably for fish intake (r = 0.51) (19).
Cox proportional hazard analyses (JMP version 5.1, SAS Institute, Cary, NC, USA) were used to estimate the dose-response relationships of incident BPH to distances run and 10-km performance speed. Reported weekly intakes of alcohol, meat, fish, and fruit, along with age and age2, were included as covariates. Current smoking, strict vegetarianism, and diabetes were handled by exclusion rather than adjustment because they represented a very small proportion of the sample. Further adjustments for BMI, running distance (km·wk−1), and 10-km performance (m·s−1) were included where indicated. In addition, relative risks (i.e., risk ratios of the hazard function) are presented that compare distance and performance intervals to all higher activity and performance levels and to the least active and slowest performing runners. The Cox proportional hazard model was also used to test whether changes in distance run per week were related to the incidence of BPH. Those results are presented adjusted for BMI at baseline and end of follow-up.
At baseline, there were 28,612 men who were nonsmoking, nonvegetarian, and nondiabetic who reported their average weekly running distance, height, body weight, and age and who did not report being diagnosed with BPH before their baseline survey. Their baseline characteristics relative to their reported distance run per week are displayed in Table 1. Reported distance run was associated with significantly younger age, slightly less education, lower BMI and narrower waist circumferences, fewer weekly servings of meat, greater fruit intake, and faster 10-km performance times.
Included among the 28,612 men were 1899 (6.64%) who reported incident BPH during (mean ± SD) 7.74 ± 1.84 yr of follow-up. The affected men were significantly older than unaffected men (mean ± SE, 53.6 ± 0.2 vs 44.0 ± 0.1 yr; P < 0.0001). They also ran significantly less per week (34.5 ± 0.5 vs 38.2 ± 0.1 km·wk−1; P < 0.0001) and ran slower 10-km races (3.71 ± 0.01 vs 3.96 ± 0.00 m·s−1; P < 0.0001), which could not be ascribed to their difference in age (age-adjusted significance differences of P = 0.005 and P = 0.002, respectively). The older age of those with BPH accounted for their lower meat consumption (affected vs unaffected, 2.65 ± 0.06 vs 2.85 ± 0.02 servings per week; unadjusted P = 0.001 became P = 0.58 when age adjusted), higher fish (1.73 ± 0.04 vs 1.53 ± 0.01 servings per week; P < 0.0001 became P = 0.57 when age adjusted) and fruit consumption (11.89 ± 0.20 vs 10.98 ± 0.05 servings per week; P < 0.0001 became P = 0.48 when adjusted for age), and broader waist circumferences (85.67 ± 0.14 vs 84.35 ± 0.04 cm; P < 0.0001 became P = 0.55 when age adjusted). Affected and unaffected men also did not differ in their BMI (23.95 ± 0.06 vs 23.86 ± 0.02 kg·m−2; P = 0.16) or weekly alcohol intake (81.32 ± 2.54 vs 81.02 ± 0.70 mL; P = 0.91).
Ten-kilometer performance times were provided on a subset of 24,107 men (84.3%), including 1560 who reported BPH. This fitness subset was younger than men who did not have a 10-km performance time to report (44.48 ± 0.06 vs 45.63 ± 0.17 yr; P < 0.0001), which explains their lower proportion of incident BPH (6.47 ± 0.16 vs 7.52 ± 0.39%; P = 0.009 reduced to P = 0.76 when adjusted for age).
Table 2 presents the relative risk from survival analyses for incident BPH versus baseline running distance (km·wk−1) and 10-km performance (m·s−1) both separately and when included simultaneously and with and without adjustment for BMI. The table shows that both longer running distance and faster 10-km race performance predicted lower BPH risk. The risk reduction attributable to distance was the same for the entire cohort as for the fitness subset (i.e., those reporting 10-km race times). Adjustment for BMI slightly increased the reductions in incidence associated with distance and performance. Restricting the analyses to men 45 yr old and older yielded essentially identical results (not displayed). Figure 1 displays the monotonic risk reductions associated with baseline 10-km performance. Men who ran ≥4.5 m·s−1 had 32% lower risk than the slowest performing men (P = 0.0006). The dose-response relationship extended across the entire performance range, with men in the fastest performance category having significantly lower risk than men in the penultimate fastest performance category (P = 0.03).
There were 25,936 men who provided their running distance on end of follow-up questionnaire. These data were analyzed in conjunction with their reported baseline level in Table 3. Two models were used differing only in the way the baseline and end of follow-up running distances were parameterized as independent variables. Model 1 included the baseline and the follow-up distances separately as independent variables. In model 1, the significance for the baseline distance represents its effect independent of the end of follow-up distance, and, correspondingly, the significance for the follow-up distance represents its effect independent of baseline distance. These analyses showed that baseline distance was a more significant predictor of BPH than follow-up distance. In model 2, the baseline and the follow-up distances were reparameterized as the average distance run and the change in distance run between baseline and follow-up. The analyses show that the average distance predicted the risk for BPH whereas the change in distance did not. Figure 2 displays the linear decline in risk with average running distance. There was significantly lower risk for men who ran >16 than <16 km·wk−1 (P = 0.05), >32 than 16-32 km·wk−1 (P = 0.02), and >48 than 32-48 km·wk−1 (P = 0.04).
Our results add to the accumulating evidence that important health benefits accrue at greater exercise doses and greater exercise intensities than currently recommended. Specifically, guidelines from the American College of Sports Medicine and the American Heart Association (16) and the Centers for Disease Control and Prevention (29) emphasize the health benefits of walking 2 miles (3.2 km) briskly on most days of the week (the energy equivalent of running 8-12 km·wk−1) (1). This corresponds to the referent activity levels of Figure 2 (<16 km·wk−1) (1). These guidelines (29), particularly the updated versions (16), acknowledge that additional benefits may accrue for longer, more intense activity. Consistent with this acknowledgement, Figure 2 shows that BPH risk continued to decline with longer running distances through at least 64 km·wk−1. Compared with ≤16 km·wk−1 (guideline levels), running 16-32, 32-48, 48-64, and >64 km·wk−1 reduced BPH risk by 6%, 12%, 24%, and 33%, respectively. In other articles, we have demonstrated that running appears to improve men's health incrementally up to 80 km·wk−1 by lowering BMI (34,40); through 64 km·wk−1 by raising plasma HDL cholesterol (34), lowering plasma triglycerides and LDL cholesterol concentrations (34), and reducing incident diabetes and hypercholesterolemia (39); and through 48 km·wk−1 by reducing hypertension (34,39).
The Health Professionals Follow-up Study suggests that physical activity reduces the odds for BPH even if the activity is only of low to moderate intensity (climbing stairs, walking, rowing, or calisthenics) (24). Our analyses confirm the inverse relationship between physical activity and BPH. This is important because few articles report this benefit compared with other health outcomes such as cardiovascular disease (29). Small sample size may explain the nonsignificance of the trend between physical activity and BPH in the 9-yr follow-up of the Massachusetts Male Aging study of 1019 men (21) and the lack of a significant total physical activity difference between Chinese men with and without BPH (20).
In addition to providing confirmation, our article extends the findings of the Health Professionals Follow-up Study to higher exercise doses and intensities. The highest quintile of total physical activity they reported was ≥33.8 MET·h per week of activity (24), the energy equivalent of 33 km·wk−1 (1), whereas Figure 2 suggests declining risk through ≥64 km·wk−1. Although the Health Professionals Follow-up Study also demonstrates that running specifically reduces the odds for BPH (P = 0.02) (24), their highest category of running was ≥2 h·wk−1. Our Figure 2 shows significant incremental decreases in risk for running over 16 km·wk−1 (P = 0.05), 32 km·wk−1 (P = 0.02), and 48 km·wk−1 (P = 0.04). Whereas the Health Professionals Follow-up Study reported that the decline in BPH risk was less for high intensity (running, jogging, biking, swimming, tennis, racquet ball, or squash) than low- to moderate-intensity physical activity (flights of stairs climbed, walking, rowing, or calisthenics), our data provide ample evidence for dose-dependent, clinically important risk reductions with vigorous exercise.
Various mechanisms have been proposed to explain how physical activity reduces BPH risk. Androgens are involved in the development and the maintenance of BPH, and men who engage in chronic endurance training have, as a group, significantly lower resting levels of testosterone (15,32). Whereas this explanation appears apropos to the high-mileage runners reported by us, Platz et al. (24) discounted this explanation for walking and other low- to moderate-intensity physical activities of the Health Professionals Follow-up Study. They attributed their results to reduced sympathetic nervous activity rather than reduced prostate volume because activity was more strongly related to symptomatic BPH than prostatic enlargement.
Our analyses may offer particular clues to the biological process(es) involved: 1) it must have little if any relationship to the leanness of the runners; 2) it must be at least as strongly related to the physical activity as cardiorespiratory fitness; and 3) it must affect BPH chronically rather than acutely. The third clue is derived from the analyses of Table 3, which show that BPH was more strongly related to the baseline than the end of follow-up activity, and more strongly related to average than change in activity during follow-up. These clues distinguish BPH from hypertension, diabetes, and lipoproteins in their relationship to physical activity. First, in other analyses (34,35,39), we have shown that adjustment for the runners' BMI substantially reduced the decrease in blood pressure, hypertension, diabetes, and triglycerides and the increases in HDL cholesterol associated with running distance (adjustment did not attenuate the exercise-BPH relationship). Second, the risks for hypertension and diabetes appear to be affected more acutely by changes in exercise (37,38). Third, the risks for hypertension and diabetes show no relationship to running distance when adjusted for 10-km race performance but show a strong relationship to 10-km performance when adjusted for distance (39) (BPH is more strongly related to running distance activity than 10-km performance). Thus, the processes that elevate HDL cholesterol, reduce triglycerides, improve glucose intolerance, and reduce diabetes risk with exercise may be poor candidates for explaining the BPH-exercise relationship. We are of the opinion that if suppression of elevated sympathetic nervous system activity causes the BPH risk reduction due to exercise, then it would be more strongly related to follow-up than baseline running distance, as observed with hypertension.
The important and influential work by Blair et al. (4,5) shows a variety of health benefits associated with cardiorespiratory fitness, including reduced cardiovascular morbidity and mortality. In formulating public health guidelines for physical activity, there was great utility in interpreting cardiorespiratory fitness as a more objective measure of physical activity than self-reported activity itself (5,29). However, Table 2 shows that BPH risk is separate, statistically independent associations with cardiorespiratory fitness (as measured by 10-km performance) and vigorous physical activity. Elsewhere, we have demonstrated that declines in the prevalence of antidiabetic, antihypertensive, and LDL-cholesterol-lowering medications with increasing fitness (m·s−1) were also independent of reported physical activity (39). Furthermore, we have shown that blood pressure levels were more strongly related to cardiorespiratory fitness than physical activity (35). These observations suggest the need for a clearer interpretation of fitness than simply as a surrogate for physical activity. For example, cardiorespiratory fitness is known to reflect a substantial hereditary component (6,7), and genetic factors associated with higher fitness levels might also reduce BPH risk. They also suggest that cardiorespiratory fitness per se warrants consideration as an important clinically indicated risk factor independent of activity.
These analyses are based on self-reported distance run per week, which, although comparable to other self-reported physical activity measures with respect to reproducibility over duplicate questionnaires, is still subject to error and potential bias that could affect the associations described in this article. The prospective cohort design used in our study is less subject to the biases associated with the case-control studies, which rely heavily on the adequacy of matching cases with controls. However, we do not wish to overstate our findings as proof of cause and effect. Although the prospective design ensures that physical activity levels preceded the clinical diagnosis of BPH, the relationship is probably better viewed as cross-sectional between the activity and probability of BPH. There is also the inherent limitation of self-reported physician diagnosis, which, although used by other epidemiological studies of BPH (2,22), is necessarily less accurate than clinically validated conditions. Our survey question on incident BPH has not been verified in relation to medical records or independent physician diagnoses and lacks specificity with respect to symptomatic and asymptomatic disease. In addition, the question does not distinguish prostate enlargement from obstructive and irritative lower urinary tract symptoms. However, the Health Professionals Follow-up Study reported the same inverse association for BPH defined by surgery or by symptoms (24). This study also found that self-reported surgery for prostate enlargement to be fully verified by medical records, and they included self-reported enlarged prostate detected by digital rectal examination as an end point. Nevertheless, our results are likely to exclude cases where these symptoms were not attributed to BPH. The ability to distinguish benign and malign disease may depend on the experience of the physician. However, it is important to emphasize that misdiagnosis and underreporting of BPH cannot account for the significant association reported here between BPH and running distance and 10-km performance unless the rates of misdiagnosis and underreporting were directly related to the reported distance run or 10-km performance time, which we believe to be unlikely.
We also caution that our analyses of incident BPH in relation to concurrent changes in running distance are not prospective and recall bias may spuriously associate follow-up distance to disease status. Moreover, our analysis of exercise change between baseline and follow-up does not take into account substitution of other activities for running.
Chronologic age is the strongest risk factor for BPH (14,21). Ten-kilometer performance times decline sharply with age (35), and therefore, the validity of these analyses depend strongly on the adequacy of our age adjustment. Age and age2 were included as covariates in all analyses, which appeared sufficient to eliminate the highly significant differences in meat, fish, and fruit consumption between affected and unaffected men and thus presumably would also have eliminated an association between fitness and BPH due solely to age. When the analyses of Tables 2 and 3 were repeated on men 45 yr old and older, the results were essentially indistinguishable from the entire sample of men. (Analyses not displayed.)
Urological complications of BPH could lead to reduced mobility and recreational activities, yet our findings are generally consistent with those for low to moderately intense activity, which presumably would be less affected (24). Dal Maso et al. (10) argue against this explanation because the relationships are consistent in younger and older men. We do not believe that the declining incidence of BPH with increasing running distance is likely due to greater stoicism in more athletic men, for this would presumably be reflected in the relationship of running distance to other maladies including degenerative arthritis (P = 0.23), kidney stones (P = 0.35), or gastric ulcers (P = 0.83), which was not observed (unpublished observations).
Although BPH has not generally been considered a preventable condition, our findings suggest that exercise may lower BPH risk. They also provide further support for encouraging physical activity at doses and intensities that substantially exceed minimum guideline levels and for recognizing cardiorespiratory fitness as a clinically important risk factor independent of physical activity.
This study was supported in part by grant AG-032004 from the National Institutes of Health and was conducted at the Ernest Orlando Lawrence Berkeley Laboratory (Department of Energy grant DE-AC03-76SF00098 to the University of California). Dr. Williams had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis. The results of the present study do not constitute endorsement by ACSM.
1. Ainsworth BE, Haskell WL, Leon AS, et al. Compendium of physical activities: classification of energy costs of human physical activities. Med Sci Sports Exerc
2. Arrighi H, Guess H, Metter E, Fozard J. Symptoms and signs of prostatism as risk factors for prostatectomy. Prostate
3. Balke B, Ware RW. An experimental study of physical fitness of Air Force personnel. US Armed Forces Med J
4. Blair SN, Jackson AS. Physical fitness and activity as separate heart disease risk factors: a meta-analysis. Med Sci Sports Exerc
5. Blair SN, Kohl HW 3rd, Paffenbarger RS Jr, Clark DG, Cooper KH, Gibbons LW. Physical fitness and all-cause mortality. A prospective study of healthy men and women. JAMA
6. Bouchard C, Lesage R, Lortie G, et al. Aerobic performance in brothers, dizygotic and monozygotic twins. Med Sci Sports Exerc
7. Bouchard C, Daw EW, Rice T, et al. Familial resemblance for V˙O2
max in the sedentary state: the HERITAGE family study. Med Sci Sports Exerc
8. Cooper KH. A means of assessing maximal oxygen intake: correlation between field and treadmill testing. JAMA
9. Crispo A, Talamini R, Gallus S, et al. Alcohol and the risk of prostate cancer and benign prostatic hyperplasia. Urology
10. Dal Maso L, Zucchetto A, Tavani A, et al. Lifetime occupational and recreational physical activity and risk of benign prostatic hyperplasia. Int J Cancer
11. Fritschi L, Tabrizi J, Leavy J, Ambrosini G, Timperio A. Risk factors for surgically treated benign prostatic hyperplasia in Western Australia. Public Health
12. Gann P, Hennekens C, Longcope C, Verhoek-Oftedahl W, Grodstein F, Stampfer M. A prospective study of plasma hormone levels, nonhormonal factors, and development of benign prostatic hyperplasia. Prostate
13. Giovannucci E, Rimm EB, Chute CG, et al. Obesity and benign prostatic hyperplasia. Am J Epidemiol
14. Guess H. Benign prostatic hyperplasia: antecedents and natural history. Epidemiol Rev
15. Hackney AC. The male reproductive system and endurance exercise. Med Sci Sports Exerc
16. Haskell WL, Lee IM, Pate RR, et al. Physical activity and public health: updated recommendation for adults from the American College of Sports Medicine and the American Heart Association. Med Sci Sports Exerc
17. Hellerstein HK. Limitations of marathon running in the rehabilitation of coronary patients: anatomic and physiologic determinants. Ann NY Acad Sci
18. Hong J, Kwon S, Yoon H, et al. Risk factors for benign prostatic hyperplasia in South Korean men. Urol Int
19. Hu FB, Rimm E, Smith-Warner SA, et al. Reproducibility and validity of dietary patterns assessed with a food-frequency questionnaire. Am J Clin Nutr
20. Lacey JV Jr, Deng J, Dosemeci M, et al. Prostate cancer, benign prostatic hyperplasia and physical activity in Shanghai, China. Int J Epidemiol
21. Meigs JB, Mohr B, Barry MJ, Collins MM, McKinlay JB. Risk factors for clinical benign prostatic hyperplasia in a community-based population of healthy aging men. J Clin Epidemiol
22. Meigs JB, Barry MJ, Giovannucci E, Rimm EB, Stampfer MJ, Kawachi I. High rates of prostate-specific antigen testing in men with evidence of benign prostatic hyperplasia. Am J Med
23. Parsons JK. Modifiable risk factors for benign prostatic hyperplasia and lower urinary tract symptoms: new approaches to old problems. J Urol
24. Platz E, Kawachi I, Rimm E, et al. Physical activity and benign prostatic hyperplasia. Arch Internal Med
25. Platz E, Rimm E, Kawachi I, et al. Alcohol consumption, cigarette smoking, and risk of benign prostatic hyperplasia. Am J Epidemiol
26. Prezioso D, Catuogno C, Galassi P, D'Andrea G, Castello G, Pirritano D. Life-style in patients with LUTS suggestive of BPH. Eur Urol
. 2001;40 Suppl 1:9-12.
27. Rohrmann S, Crespo CJ, Weber JR, Smit E, Giovannucci E, Platz EA. Association of cigarette smoking, alcohol consumption and physical activity with lower urinary tract symptoms in older American men: findings from the third National Health and Nutrition Examination Survey. BJU Int
28. Swain DP. Moderate or vigorous intensity exercise: which is better for improving aerobic fitness? Prev Cardiol
29. US Department of Health and Human Services. Physical Activity and Health: A Report of the Surgeon General
. Atlanta, GA: US Department of Health and Human Services, Centers for Disease Control and Prevention, National Center for Chronic Disease Prevention and Health Promotion, 1996, pp. 1-276. ISBN-13: 978-0763706364.
30. Wei JT, Calhoun E, Jacobsen SJ. Urologic Diseases in America project: benign prostatic hyperplasia. J Urol
31. Wenger HA, Bell GJ. The interactions of intensity, frequency and duration of exercise training in altering cardiorespiratory fitness. Sports Med
32. Wheeler GD, Wall SR, Belcastro AN, Cumming DC. Reduced serum testosterone and prolactin levels in male distance runners. JAMA
33. Wong SY, Woo J, Hong A, Leung JC, Kwok T, Leung PC. Risk factors for lower urinary tract symptoms in southern Chinese men. Urology
34. Williams PT. Relationship of distance run per week to coronary heart disease risk factors in 8283 male runners: the National Runners' Health Study. Arch Intern Med
35. Williams PT. Relationships of heart disease risk factors to exercise quantity and intensity. Arch Intern Med
36. Williams PT. Vigorous exercise and the population distribution of body weight. Int J Obesity
37. Williams PT. Changes in vigorous physical activity and incident diabetes in male runners. Diabetes Care
38. Williams PT. A cohort study of incident hypertension in relation to changes in vigorous physical activity in men and women. J Hypertens
39. Williams PT, Franklin B. Vigorous exercise and diabetic, hypertensive, and hypercholesterolemia medication use. Med Sci Sports Exerc
40. Williams PT, Pate RR. Cross-sectional relationships of exercise and age to adiposity in 60,617 male runners. Med Sci Sports Exerc