American Journal of Epidemiology

Vol 133. No. 2

Copyright C 1991 by The Johns Hopkins University School of Hygiene and Puttfc Health

PnntedlnUSA

All rights reserved

ORIGINAL CONTRIBUTIONS

Lifetime Occupational Physical Activity and Prostate Cancer Risk

The authors compared the lifetime occupational physical activity of 452 prostate cancer cases identified through the population-based Hawaii Tumor Registry and 899 population controls interviewed from 1981 to 1983. Each job reported was classified into one of five levels of physical activity using published sources. Among men aged 70 years or older, a negative association was found between prostate cancer risk and proportion of life spent in jobs involving only sedentary or light work. Compared with men never employed in such jobs, men who spent more than 54% of their life in these jobs had an odds ratio of 0.5 (95% confidence interval 0.3-0.9). This negative association was dose-dependent, consistent across ethnic groups, and unrelated to socioeconomic status, dietary risk factors, or job-related chemical exposures. The findings for younger men were less clear, but not inconsistent with those for older men. Surprisingly, no association was found with years spent in moderately active or very active jobs in either age group. Although inconclusive, these results suggest that physical activity may be positively associated with the risk of prostate cancer, but this association is likely to be weak and indirect. Am J Epidemiol 1991 ;133:103-11. exercise; occupations; prostatic neoplasms

Recent epidemiologic studies have reported a protective effect of physical activity against colon and breast cancers (1,2). However, the relation of exercise to other cancer sites has received little attention. Prostate cancer is a very common cancer, for which no causal factor has been established. Two

of the main etiologic hypotheses for this cancer, prolonged androgenic stimulation and high intake of fat-derived calories (3), are possibly related to physical activity. We report here on a population-based case-control study of lifetime occupational physical activity and prostate cancer.

Received for publication June 21, 1989, and in final form July 2, 1990. From the Epidemiology Program, Cancer Research Center, University of Hawaii, Honolulu, HI. Reprint requests to Dr Lxic Le Marchand, Epidemiology Program, Cancer Research Center of Hawan, University of Hawai, 1236 Lauhala Street, Suite 407, Honolulu, HI 96813. This work was supported in part by grant nos R26 CA 28943 and NO1 CA 15655 and contract no. NO1 CN 55424 from the National Cancer Institute. The authors thank the following hospitals for their support of this study Castle Medical Center; Kaiser Medical Center; Kuakini Medical Center; Queen's Medical Center; St. Francis Hospital; Straub Clinic and Hospital; and Wahiawa General Hospital. The technical assistance of Lynn Yamanaka and Katharine McDuffie is also gratefully acknowledged.

103

MATERIALS AND METHODS

The methodology used in this study has been previously described in detail (4). Briefly, the cases were all histologically confirmed patients with invasive prostate cancer identified among the residents of the island of Oahu during 1977-1983 through the Hawaii Tumor Registry, a member of the National Cancer Institute's Surveillance, Epidemiology, and Ends Results (SEER) program. An interview was completed for 62 percent (456 men) of the eligible cases. The main reasons for nonparticipation were pa-

Downloaded from https://academic.oup.com/aje/article-abstract/133/2/103/118448 by Insead user on 10 September 2018

Lofc Le Marchand, Laurence N. Kolone), and Carl N. Yoshizawa

104

Le Marchand et a). 4,000 Jobs as Defined in the Dictionary of Occupational Titles (DOT) (8). Occupations held before 1960 were coded according to the DOT physical requirements published in 1956 (8). Physical activity ratings published in 1966 (9) were used to code the occupations held between 1960 and 1969. Those published in 1981 (10) were used to rate jobs held after 1969. The physical activity coding was performed without knowledge of the case-control status of the subjects. In addition, the participants were asked whether they had worked with or around several chemicals that have been associated with prostate cancer or other cancers. We also used a job exposure matrix (11) to impute chemical or physical exposures using the job held the longest by each subject. Because differential effects have been demonstrated in older and younger men for both dietary and hormonal risk factors for prostate cancer (12,13), an a priori decision was made to analyze the data separately for men 70 years or older (57 percent of the sample) and those less than 70 years. The number of years at each activity level was calculated for each subject by summing the number of years in each job with a similar physical requirement rating. Several indices of physical activity were then constructed; they included the proportion of work years in jobs involving sedentary or light work, and the proportion of the subject's life in such jobs. Similar variables were created for jobs with heavy and very heavy work. Unconditional logistic regression models (14) were used to estimate odds ratios (and 95 percent confidence intervals) for several levels (quartiles or tertiles) of these variables, adjusting for age and ethnicity. Additional adjustment variables for some of the models included education, income, saturated fat and beta-carotene intakes, as well as total zinc intake (from dietary and supplement sources). These dietary variables were selected because they were associated with prostate cancer risk in these data, whereas weight and body mass index were not (4). Quartile (or tertile) cutpoints were based on all subjects.

Downloaded from https://academic.oup.com/aje/article-abstract/133/2/103/118448 by Insead user on 10 September 2018

tient refusals (13 percent), physician refusal (15 percent), and untraceability, language barrier, or death with absence of suitable surrogate for interview (9 percent). In addition, four cases were excluded because their answers were judged unreliable by the interviewer. Sixty-seven percent of the cases were diagnosed at a localized stage, whereas 13 and 19 percent were diagnosed at regional and distant stages, respectively. The staging for seven cases could not be confirmed from the Hawaii Tumor Registry. Two population controls were frequencymatched to each case by 5-year age group. For cases over age 65, controls were selected from a 10 percent random sample of all male Oahu residents registered with the Health Care Financing Administration. For cases under age 65, we used a random digit dialing procedure similar to that described by Waksberg (5). An interview was completed for 76 percent (914 men) of the selected controls. Nonparticipation was due to refusal (11 percent), and inability to locate, language barrier, or death with absence of surrogate (13 percent). Fifteen completed control interviews were rejected because they were judged unreliable by the interviewer. Most interviews were conducted at home. When the subject was too ill or had died, an interview was sought from a suitable substitute (usually the spouse). Proxy interviews were obtained from 30 percent of the cases and 11 percent of the controls. In addition to dietary, demographic, and social information, the questionnaire included a complete lifetime occupational history which documented the job title, the type of industry, and period of time worked for each occupation held. These occupations were coded by job title and industry according to the US Department of Commerce Alphabetical Index of Occupations for the 1970 Census (6). Following the method used by Vena et al. (7), each of these jobs was then assigned one of five levels of physical activity (sedentary work, light work, medium work, heavy work, very heavy work), as rated by the Department of Labor in their Estimates of Worker Trait Requirements for

Occupational Physical Activity and Prostate Cancer

RESULTS

The numbers of cases and controls, as well as some background characteristics for the subjects, are presented in table 1 by age group. There were no statistically significant differences between cases and controls in educational achievement, lifetime duration of employment, smoking and drinking habits, and body size. The age- and ethnicity-adjusted risk estimates for years spent in jobs with sedentary or light work and for proportions of the subjects' working life and total life in such jobs are presented by age group in table 2. Although some of the odds ratios were significantly reduced, no clear association was observed among men less than 70 years of age. Among older men, a pattern compatible with a weak negative association was apparent. Compared with men never employed in jobs with sedentary or light work, men who spent less than 54 percent of their life in such jobs had a 30 percent reduction in risk. Those who spent more than 54 percent of their life in these jobs had a 40 percent reduction in risk. These risk reductions, as well as the corresponding trend test, were of borderline statistical significance. Com-

parable results were found for the two other activity classifications in the table. Table 3 presents the results for proportion of life in sedentary or light work for men aged 70 years and older after further adjustment for two measures of socioeconomic status (income and education). Unfortunately, 49 cases and 66 controls had to be excluded from this analysis because of missing information on these covariates. The odds ratio for the fourth quartile was slightly reduced by this adjustment and reached statistical significance. The p value for the corresponding trend test was 0.04. Similar effects were observed when this analysis was also repeated for the other classification measures in table 2. Further adjustment for the dietary variables found associated with risk in these data (saturated fat, beta-carotene, and total zinc intakes) did not materially change the odds ratios. Table 4 presents the age- and ethnicityadjusted risk estimates for proportions of life spent in moderately active jobs, and in jobs with heavy or very heavy work, by age group. No associations were found with these variables in either of the age-categories. Very similar risk patterns were obtained for number of years and proportion of working life in these jobs. Further adjustment for socioeconomic status and the other covariates did not modify these results. To examine the internal consistency of the findings for the sedentary and light work variables, we computed ethnic-specific risk estimates among older men using common tertile cutpoints (table 5). The negative association with proportion of life spent in

TABLE 1. Mean characteristics* of prostate cancer cases and controls, by age, Hawaii, 1977-1983 Age 0.54

No. of

No. of controls

OR (95% Cl)

86

171 88 88 95

1.0 0.7(0.4-1.2) 0.7(0.4-1.2) 0.5 (0.3-0.9)

42 44 42

p = 0.04t * Odds ratios (ORs) adjusted for age, ethnicity, Income, and education by multiple logistic regression analysis Cl, confidence Interval t Trend test.

sedentary and light work was apparent among four of the five ethnic groups included in the study. Similar results were obtained with the two other sedentary and light work variables. Further adjustment for income and education did not alter these findings.

Because sedentary and light activity jobs are less likely to involve exposures to chemicals than more physically strenuous ones, we compared the frequency with which selected job-related chemical exposures were reported by cases and controls aged 70 years and older (table 6). No statistically signifi-

Downloaded from https://academic.oup.com/aje/article-abstract/133/2/103/118448 by Insead user on 10 September 2018

No. of years In job with sedentary or Bght work None 1-22 23-37 >37

No. of cases

Occupational Physical Activity and Prostate Cancer

107

TABLE 4. Odds ratios* for prostate cancer by age and proportion of life in moderate and heavy activity jobs, Hawaii, 1977-1983 Age 0.43

Age £70 years

OR (95% CO

1.0t

0.9(0.6-1.5) 0.7 (0.4-1.2) 1.0(0.6-1.7)

No. of cases

No. of controls

OR (95% Cl)

90 61 57 55

176 104 128 100

1.0 1.2(0.8-1.8) 1.0(0.6-1.5) 1.1 (0.7-1.7)

p = 0.92J Proportion of Irfe with heavy or very heavy work 0.00 0.01-0.22 X3.22

108 44 37

200 91 100

1.0 1.0(0.6-1.5) 0.8 (0.5-1.3)

p = 0.79

128 64

245 128 135

71

1.0 1.1 (0.7-1.6) 1.3(0.9-1.9)

p = 0.27

p = 0.25

• Odds ratios (ORs) adjusted for age and ethnicity by muttipte logistic regression analysis. Cl, confidence Interval, t Reference category. t Trend test

TABLE 5. Odds ratios* for prostate cancer by job activity level and ethnicity among men aged 70 years or older, Hawaii, 1977-1983 Proportion of Ife in jobs wtth sedentary or light worit

Japanese (103/193)t

Caucasian (71/107)

Filipino (40/122)

Chinese (28/59)

Hawaiian (21/27)

OR (95% Cl)

OR (95% Cl)

OR (95% Cl)

OR (95% Cl)

OR (95% CO

1.0 0.7(0.3-1.8) 0.9 (0.2-4.8)

0.9 (0.2-3.6) 0.7 (0.2-3.2)

0.61

0.68

1.0

1.0

0.00* 0.01-0.45 >0.45

0.6(0.4-1.2)

1.0 0.6(0.2-1.3)

0.6(0.4-1.2)

0.5(0.2-1.1)

0.5(0.1-1.7) 1.4 (0.5-4.2)

p for trend

0.13

0.48

0.30

• Odds ratios (ORs) adjusted for age by multiple logistic regression analysis. Cl, confidence interval, t No. of cases/no, of controls. t Reference category.

TABLE 6. Odds ratios* for prostate cancer by self-reported chemical occupational exposures among men aged 70 years and older, Hawaii, 1977-1983 No. exposed Cases

Controls

Odds ratio

95% confidence interval

37 31 3 8 5 24

66 65 8 10 11 42 2

1.3 0.9 0.7 1.5 0.8 1.0 0.8

0.8-2.0 0.5-1.4 0.2-2.6 0.6-4.0 0.3-2.3 0.6-1.8 0.1-8.5

Type of chemical

Pesticides Paints Nickel Zinc Cadmium Asbestos Rubber (manufacturing process)

1

' Adjusted for age and ethnicity by multiple logistic regression.

Downloaded from https://academic.oup.com/aje/article-abstract/133/2/103/118448 by Insead user on 10 September 2018

No. of cases

108

Le Marchandet al.

TABLE 7. Odds ratios* for prostate cancer by mafor occupational categories using the longest job held by men aged 70 years or older, Hawaii, 1977-1983 Occupational category

32 32 14 29 17 60 9 13 15 27

No. of controls

Odds ratio

95% confidence interval

62 53 35 45 32 115 24 24 40 53

0.9 1.1 0.7 1.5 1.0 1.0 0.8 1.1 0.8 1.2

0.5-1.4 0.7-1.7 0.4-1.4 0.9-2.5 0.5-1.8 0.7-1.4 0.4-1.8 0.6-2.2 0.4-1.5 0.7-2.0

* Adjusted for age and ethnicity by multiple logistic regression.

cant differences were found. A similar comparison was made with the exposures imputed from a job-exposure matrix (11), based on the job held the longest. Again, no association with prostate cancer risk was found (results not shown). We also compared the distribution of cases and controls by longest held occupation (table 7), and found no statistically significant excess risk for any category. Thus, the protection afforded by employment in sedentary or light activity jobs does not appear to be due to a differential in exposure to potential occupational carcinogens. DISCUSSION

In this population-based case-control study of prostate cancer, we did not find any association with years spent in moderately active or very active jobs. However, we found a negative association with years spent in sedentary or light activity occupations among men aged 70 years or older. Although the results for sedentary and light activity jobs in younger men were less clear, they were not inconsistent with a similar protective association, since the odds ratios were all lower than or equal to 1.0. This negative association with years spent in sedentary jobs in older men was "dosedependent," unrelated to socioeconomic status or dietary risk factors, and consistent across ethnic groups. It was also unlikely to be due to a relative lack of job-related chemical exposures, since no particular occupa-

tion or exposure (either reported by the subjects or imputed from a job-exposure matrix) was associated with prostate cancer risk. This is consistent with the findings of past studies which suggest that, overall, occupational exposures are not of major importance in the etiology of prostate cancer (15, 16). The protective association observed in our data with years spent in sedentary or light activity jobs is difficult to interpret in the absence of a direct association with years spent in active jobs. Vena et al. (7), who used the same methodology in a colon cancer study, did not report on the risk associated with years spent in moderately or very active jobs. However, the positive association that they observed with years spent in sedentary and light activity occupations agrees with the results of studies which included more direct measures of occupational and/or leisure time physical activity (1, 7, 17-19). Thus, it is possible that employment in a sedentary occupation is a better marker of physical activity than years spent in active jobs. Because we were unable to take into account factors such as seniority and incapacitating injuries or illnesses, which may lead to less strenuous assignments without a change in job title, more misclassification is likely to have occurred in the moderately and very active job categories than in the sedentary job categories. Such a nondifferential misclassification may have resulted in our inability to detect a positive association

Downloaded from https://academic.oup.com/aje/article-abstract/133/2/103/118448 by Insead user on 10 September 2018

Professional/tech nica) Managers, administrators Sales Service Clerical Craftsmen Operatives Transport operatives Laborers Farmers/farm managers

^ ^

Occupational Physical Activity and Prostate Cancer

decreased for men in the lower tertile of arm muscle area, as estimated from triceps skinfold thickness and arm circumference. Men in the middle and upper tertiles had similar risks, and no relation was found with arm fat area. These results would be consistent with those of Paffenbarger et al. (20) and ours, if men with small muscle mass were more likely to have been inactive in college and to have taken on a sedentary occupation at an earlier age than more muscular men. Increased androgen levels could constitute a possible mechanism for a positive association between physical activity and prostate cancer. Testosterone is suspected to play an important role in prostate carcinogenesis (27). Androgens are also known to be related to muscle mass and, possibly, physical activity. The pubertal surge of testosterone is known to have an hypertrophic effect on the upper-body muscle mass, and male hypogonadism is associated with muscular hypotrophy (28). Androgens (anabolic steroids) may increase muscle mass and are thought to generate an increase in aggressive behavior and drive for training in the athletes who use them (29). However, there is no evidence that an increase in moderate exercise would result in a change in androgen levels (30), and we are not aware of any hormonal studies comparing men with sedentary occupations with men employed in more vigorous jobs. In summary, although the evidence is still sketchy, an association of physical activity with prostate cancer would probably be indirect and could reflect the effect of circulating androgen levels. The potential limitations of this study need to be addressed. No attempt was made to validate the occupational histories. However, subjects were interviewed at home and were asked to refer to available records to complement the recalled information. There is no reason to suspect that prostate cancer cases would differentially recall sedentary occupations. Some bias may also have been introduced into our data by the use of surrogate respondents. To investigate this possibility, we repeated our analysis for men 70 years and older excluding the 53 case and 76 control surrogate interviews. The risk

Downloaded from https://academic.oup.com/aje/article-abstract/133/2/103/118448 by Insead user on 10 September 2018

with years spent in active jobs. Thus, our data may still be supportive of a positive association between physical activity and prostate cancer risk. The epidemiologic literature on physical activity and prostate cancer is inconsistent. Paffenbarger et al. (20) found that students who engaged in at least 5 hours of sports per week while in college were more likely to develop prostate cancer than were less active students (relative risk = 1.66, p = 0.03). However, these authors did not find an association with a more contemporaneous measure of physical activity among a subset of the same cohort, nor did they find an association between a measure of energy expenditure on the job and prostate cancer mortality in a cohort of longshoremen (20). In a cohort of Japanese men in Hawaii, Severson et al. (21) found no association between prostate cancer incidence and several measures of physical activity at baseline. The lack of association observed in these studies with estimates of recent physical activity would be expected if only exercise during early adulthood would affect prostate cancer risk. Such an age-specific effect is suggested by the results in college alumni of Paffenbarger et al. (20), described above. Unfortunately, we were unable to examine the independent effect of sedentary or light work during early adulthood in our data because once a subject had taken such a job, he tended to remain in it for the rest of his working life. No age effect was detected for years spent in moderately or very active jobs. Further insight into the possible relation of physical activity to prostate cancer can be gained from the results of studies of body mass and this cancer. Although data from case-control studies have been inconsistent (4, 12, 22), prospective studies have invariably shown that men with an increased body mass are at greater risk for prostate cancer (23-25). Body mass may reflect body fat and/or lean tissue (26). Severson et al. (25) have recently suggested that this association might be accounted for more by muscle mass than by fat tissue. They found in their cohort of Japanese Americans in Hawaii that risk of prostate cancer was significantly

109

110

Le Marchand et al.

4. 5. 6.

7. 8.

9.

10. 11.

12. 13.

14.

15.

16.

17. 18.

REFERENCES 1. Slattery ML, Schumacher MC, Smith KR, et al. Physical activity, diet, and risk of colon cancer in Utah. Am J Epidemiol 1988; 128:989-99. 2. Frisch RE, Wyshak G, Albright NL. Lower prevalence of breast cancer and cancers of the reproductive system among former college athletes compared to non-athletes. Br J Cancer 1985;52:88591. 3. Mandel JS, Schuman LM. Epidemiology of cancer of the prostate. In: LUienfeld AM, ed. Reviews in

19.

20.

21. 22.

cancer epidemiology. Vol 1. New York; Elsevier/ North Holland, 1980:1-83. Kolonel LN, Yoshizawa CN, Hankin JH. Diet and prostatic cancer a case-control study in Hawaii. Am J Epidemiol 1988,127:999-1012. Waksberg J. Sampling methods for random digit dialing. J Am Stat Assoc 1976;73:4O-6. US Department of Commerce, Bureau of the Census. 1970 census of the population, alphabetical index of occupations and industries. Washington, DC: US GPO, 1970. Vena JE, Graham S, Zielezny M, et al. Lifetime occupational exercise and colon cancer. Am J Epidemiol 1985;l22:357-65. US Department of Labor, Employment Service, Bureau of Employment Security. Estimates of worker trait requirements for 4000 jobs as defined in the Dictionary of Occupational Titles (an alphabetical index). Washington, DC: US GPO, 1956. US Department of Labor. Selected characteristics of occupations (physical demands, working conditions, training time). A supplement to the Dictionary of Occupational Titles, 3rd ed. Washington, DC: US GPO, 1966. US Department of Labor. Selected characteristics of occupations defined in the Dictionary of Occupational Titles. Washington, DC: US GPO, 1981. Hoar SK, Morrison AS, Cole P, et al. An occupation and exposure linkage system for the study of occupational carcinogenesis. J Occup Med 1980,22:722-6. Graham S, Haughey B, Marshall J, et al. Diet in the epidemiology of carcinoma of the prostate gland. JNCI 1983;70:687-92. Zumoff B, Levin J, Strain GW, et al. Abnormal levels of plasma hormones in men with prostate cancer evidence toward a "two-disease theory." Prostate 1982;3:579-88. Breslow NE, Day NE. Statistical methods in cancer research. Vol I. The analysis of case-control studies. IARC scientific publication no. 32. Lyon: IARC, 1980. Ernster V, Selvin S, Brown S, et al. Occupation and prostatic cancer. A review and retrospective analysis based on death certificates in two Califomian counties. J Occup Med 1979;21:175-83. Pearce NE, Sheppard RA, Fraser J. Case-control study of occupation and cancer of the prostate in New Zealand. J Epidemiol Community Health 1987;41:130-2. Garabrant DH, Peters JM, Mack TM, et al. Job activity and colon cancer risk. Am J Epidemiol 1984; 119:1005-14. Gerhardsson M, Floderus B, Norell SE. Physical activity and colon cancer risk. Int J Epidemiol 1988; 17:743-6. Wu AH, Paganini-HiU RK, Henderson BE. Alcohol, physical activity and other risk factors for colorectal cancer a prospective study. Br J Cancer 1987;55:687-94. Paffenbarger RS, Hyde RT, Wing AL. Physical activity and incidence of cancer in diverse populations: a preliminary report Am J Clin Nutr I987;45:3I2-17. Severson RK, Nomura AMY, Grove JS, et al. A prospective analysis of physical activity and cancer. Am J Epidemiol 1989; 130:522-9. Talamini R, La Vecchia C, DeCarli A, et al. Nutri-

Downloaded from https://academic.oup.com/aje/article-abstract/133/2/103/118448 by Insead user on 10 September 2018

estimates that we obtained for percent of life in sedentary or light jobs were very similar to those based on the entire sample. Men employed in sedentary and light activity jobs may require a lower caloric intake than those employed in more active jobs. However, in our data, energy intake was not associated with prostate cancer risk, and further adjustment for total calories did not materially modify our risk estimates for proportion of life spent in sedentary and light jobs. Although we do not know of any past report of an association between energy intake and prostate cancer, the possibility that caloric intake earlier in life could affect the risk for this cancer cannot be excluded. Finally, because of the clinical insignificance of some slow-growing prostatic carcinomas, the implications of our results would be limited if they only applied to less aggressive tumors. This was not the case, because when we repeated our analysis for sedentary and light activity occupations in older men by stage category (localized vs. regional + distant), we found similar negative associations in each category. Although no firm conclusions can be drawn from our data, this study suggests that physical activity might be positively associated with prostate cancer risk, but this association is likely to be weak and indirect. Future studies should obtain detailed measurements of lifetime occupational, as well as leisure-time, physical activity, and should assess total diet and lean body mass. Important insights might also be gained from investigations of the relation between androgen levels and physical activity.

Occupational Physical Activity and Prostate Cancer

23. 24.

26.

27. Ross RK, Shimizu H, Paganini-Hill A, et al. Casecontrol studies of prostate cancer in Blacks and Whites in Southern California. JNCI 1987;78:86974. 28. Hall R, Anderson J, Smart GA, et al. Clinical endocrinology. Philadelphia: JB Lippincott Co, 1974. 29. Cowart V. Steroids in sports: after four decades, time to return these genies to bottle? JAMA 1987;257:421-7. 30. Adlercreutz H, Harkonen M, Kuoppasalmi K, et al. Effect of training on plasma anabolic and catabolic steroid hormones and their response during physical exercise. Int J Sports Med 1986;7:27-8.

Downloaded from https://academic.oup.com/aje/article-abstract/133/2/103/118448 by Insead user on 10 September 2018

25.

tion, social factors and prostatic cancer in a Northem Italian population. Br J Cancer I986;53:81721. Snowdon DA, Phillips RL, Choi W. Diet, obesity, and risk of fatal prostate cancer. Am J Epidemiol 1984; 120:244-50. Garfinkel L. Overweight and mortality. Cancer 1986;58:1826-9. Severson RK, Grove JS, Nomura AMY, et al. Body mass and prostatic cancer a prospective study. Br Med J 1988-^97:713-15. Gam SM, Leonard WR, Hawthorne VM. Three limitations of the body mass index. Am J Clin Nutr 1986;44:996-7.

111

Lifetime occupational physical activity and prostate cancer risk.

The authors compared the lifetime occupational physical activity of 452 prostate cancer cases identified through the population-based Hawaii Tumor Reg...
584KB Sizes 0 Downloads 0 Views