Adiposity and Physical Activity as Predictors of Mortality
http://www.100md.com
《新英格兰医药杂志》
To the Editor: Hu et al. (Dec. 23 issue)1 show that both increased adiposity and reduced physical activity are strong and independent predictors of death from all causes in the Nurses' Health Study. We observed very similar findings among both men and women in the Cancer Prevention Study II, a prospective cohort study of the American Cancer Society.2 We followed 230,298 women and 90,162 men who were healthy and had never smoked when they were enrolled, in 1982. The subjects' weight, height, and level of physical activity were self-reported at enrollment. Although our measure of physical activity was crude in comparison with that used by Hu et al., our findings are quite similar (Table 1). Among both men and women, mortality rates increased independently across levels of adiposity and physical activity; among subjects who were obese and reported no physical activity, mortality rates were approximately double those among normal-weight subjects who reported heavy physical activity. It is abundantly clear that public health recommendations must focus on the benefits of both a healthy weight and regular physical activity.
Table 1. Relative Risk of Death from All Causes among Women and Men in the Cancer Prevention Study II, 1982 to 2002, According to Joint Categories of Body-Mass Index and Physical Activity.
Eugenia E. Calle, Ph.D.
Lauren R. Teras, M.P.H.
Michael J. Thun, M.D., M.P.H.
American Cancer Society
Atlanta, GA 30329
jcalle@cancer.org
References
Hu FB, Willett WC, Li T, Stampfer MJ, Colditz GA, Manson JE. Adiposity as compared with physical activity in predicting mortality among women. N Engl J Med 2004;351:2694-2703.
Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW. Body-mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med 1999;341:1097-1105.
To the Editor: On the basis of their analysis of data from the Nurses' Health Study, Hu and colleagues conclude that obesity and physical inactivity independently predict mortality. The results, however, are not adjusted for socioeconomic status, despite its known associations with obesity, physical inactivity, and increased mortality among women.1,2,3 The authors maintain that "the relative homogeneity of the cohort with respect to socioeconomic status and educational level reduces confounding."
To examine the socioeconomic homogeneity of nurses, we performed an analysis in which all white women with the same birth years as those in the cohort were selected from the 5 percent microdata sample of the 1980 U.S. Census,4 the census closest to the cohort's baseline year. Women with an occupation code (current or most recent) for "registered nurse" were compared with other white women. Family income was as heterogeneous for nurses as it was for the entire population of women (Table 1). Nurses' incomes were themselves heterogeneous and accounted for the minority of their families' incomes. Therefore, it is not clear whether the magnitude and independence of the effects of obesity and physical inactivity before and after adjustment for socioeconomic status would be similar.
Table 1. Socioeconomic Characteristics of White Women 34 to 59 Years of Age from the 5 Percent Microdata Sample of the 1980 Census.
Diane S. Lauderdale, Ph.D.
University of Chicago
Chicago, IL 60637
lauderdale@uchicago.edu
References
Sorlie PD, Backlund E, Keller JB. US mortality by economic, demographic, and social characteristics: the National Longitudinal Mortality Study. Am J Public Health 1995;85:949-956.
Crespo CJ, Smit E, Andersen RE, Carter-Pokras O, Ainsworth BE. Race/ethnicity, social class and their relation to physical inactivity during leisure time: results from the Third National Health and Nutrition Examination Survey, 1988-1994. Am J Prev Med 2000;18:46-53.
Sobal J, Stunkard AJ. Socioeconomic status and obesity: a review of the literature. Psychol Bull 1989;105:260-275.
Ruggles S, Sobek M, Alexander T, et al. Integrated public use microdata series, version 3.0. Minneapolis: Minnesota Population Center, 2003 (database).
The authors reply: We appreciate the efforts of Calle et al. to replicate our analyses by examining data from the Cancer Prevention Study II. We are gratified to know that their results are consistent with ours in supporting the independent effects of obesity and physical activity in predicting mortality.
We also appreciate Dr. Lauderdale's comments on the potential confounding effects of socioeconomic status on our results. The census data confirm a high degree of homogeneity for educational attainment among registered nurses; 84 percent had at least some college-level education. Also, the median family income among registered nurses was higher than among women in the general population, so that even though the interquartile range and standard deviation of family income were similar to those of the general population, few participants in the Nurses' Health Study would be living in poverty. To address explicitly a potential confounding effect of household income, we further adjusted for the median annual household income (less than $27,600, $27,600 to $36,699, $36,700 to $48,599, and $48,600 or more) from the census track data based on participants' home addresses. Although a higher income was associated with a significantly lower risk of death (relative risk, 0.90; 95 percent confidence interval, 0.84 to 0.97) when the highest category was compared with the lowest, this adjustment had virtually no effect on our results. As compared with women who were lean (i.e., those with a body-mass index of less than 25) and active (i.e., they spent 3.5 or more hours per week exercising), the multivariate relative risk of death was 1.56 (95 percent confidence interval, 1.42 to 1.71) among lean but inactive women, 1.90 (95 percent confidence interval, 1.58 to 2.28) among women who were obese (i.e., those with a body-mass index of 30 or higher) but active, and 2.41 (95 percent confidence interval, 2.13 to 2.72) among inactive, obese women. Further adjustment for the occupations of the participants' parents when they were 16 years old, as well as for the educational attainment of the participants' husbands, did not materially alter our findings. These additional analyses strongly indicate that the independent effects of physical activity and obesity on mortality observed in our study are unlikely to be explained by socioeconomic status.
Frank B. Hu, M.D., Ph.D.
Walter C. Willett, M.D., Dr.P.H.
Harvard School of Public Health
Boston, MA 02115
frank.hu@channing.harvard.edu
JoAnn E. Manson, M.D., Dr.P.H.
Brigham and Women's Hospital
Boston, MA 02115
Table 1. Relative Risk of Death from All Causes among Women and Men in the Cancer Prevention Study II, 1982 to 2002, According to Joint Categories of Body-Mass Index and Physical Activity.
Eugenia E. Calle, Ph.D.
Lauren R. Teras, M.P.H.
Michael J. Thun, M.D., M.P.H.
American Cancer Society
Atlanta, GA 30329
jcalle@cancer.org
References
Hu FB, Willett WC, Li T, Stampfer MJ, Colditz GA, Manson JE. Adiposity as compared with physical activity in predicting mortality among women. N Engl J Med 2004;351:2694-2703.
Calle EE, Thun MJ, Petrelli JM, Rodriguez C, Heath CW. Body-mass index and mortality in a prospective cohort of U.S. adults. N Engl J Med 1999;341:1097-1105.
To the Editor: On the basis of their analysis of data from the Nurses' Health Study, Hu and colleagues conclude that obesity and physical inactivity independently predict mortality. The results, however, are not adjusted for socioeconomic status, despite its known associations with obesity, physical inactivity, and increased mortality among women.1,2,3 The authors maintain that "the relative homogeneity of the cohort with respect to socioeconomic status and educational level reduces confounding."
To examine the socioeconomic homogeneity of nurses, we performed an analysis in which all white women with the same birth years as those in the cohort were selected from the 5 percent microdata sample of the 1980 U.S. Census,4 the census closest to the cohort's baseline year. Women with an occupation code (current or most recent) for "registered nurse" were compared with other white women. Family income was as heterogeneous for nurses as it was for the entire population of women (Table 1). Nurses' incomes were themselves heterogeneous and accounted for the minority of their families' incomes. Therefore, it is not clear whether the magnitude and independence of the effects of obesity and physical inactivity before and after adjustment for socioeconomic status would be similar.
Table 1. Socioeconomic Characteristics of White Women 34 to 59 Years of Age from the 5 Percent Microdata Sample of the 1980 Census.
Diane S. Lauderdale, Ph.D.
University of Chicago
Chicago, IL 60637
lauderdale@uchicago.edu
References
Sorlie PD, Backlund E, Keller JB. US mortality by economic, demographic, and social characteristics: the National Longitudinal Mortality Study. Am J Public Health 1995;85:949-956.
Crespo CJ, Smit E, Andersen RE, Carter-Pokras O, Ainsworth BE. Race/ethnicity, social class and their relation to physical inactivity during leisure time: results from the Third National Health and Nutrition Examination Survey, 1988-1994. Am J Prev Med 2000;18:46-53.
Sobal J, Stunkard AJ. Socioeconomic status and obesity: a review of the literature. Psychol Bull 1989;105:260-275.
Ruggles S, Sobek M, Alexander T, et al. Integrated public use microdata series, version 3.0. Minneapolis: Minnesota Population Center, 2003 (database).
The authors reply: We appreciate the efforts of Calle et al. to replicate our analyses by examining data from the Cancer Prevention Study II. We are gratified to know that their results are consistent with ours in supporting the independent effects of obesity and physical activity in predicting mortality.
We also appreciate Dr. Lauderdale's comments on the potential confounding effects of socioeconomic status on our results. The census data confirm a high degree of homogeneity for educational attainment among registered nurses; 84 percent had at least some college-level education. Also, the median family income among registered nurses was higher than among women in the general population, so that even though the interquartile range and standard deviation of family income were similar to those of the general population, few participants in the Nurses' Health Study would be living in poverty. To address explicitly a potential confounding effect of household income, we further adjusted for the median annual household income (less than $27,600, $27,600 to $36,699, $36,700 to $48,599, and $48,600 or more) from the census track data based on participants' home addresses. Although a higher income was associated with a significantly lower risk of death (relative risk, 0.90; 95 percent confidence interval, 0.84 to 0.97) when the highest category was compared with the lowest, this adjustment had virtually no effect on our results. As compared with women who were lean (i.e., those with a body-mass index of less than 25) and active (i.e., they spent 3.5 or more hours per week exercising), the multivariate relative risk of death was 1.56 (95 percent confidence interval, 1.42 to 1.71) among lean but inactive women, 1.90 (95 percent confidence interval, 1.58 to 2.28) among women who were obese (i.e., those with a body-mass index of 30 or higher) but active, and 2.41 (95 percent confidence interval, 2.13 to 2.72) among inactive, obese women. Further adjustment for the occupations of the participants' parents when they were 16 years old, as well as for the educational attainment of the participants' husbands, did not materially alter our findings. These additional analyses strongly indicate that the independent effects of physical activity and obesity on mortality observed in our study are unlikely to be explained by socioeconomic status.
Frank B. Hu, M.D., Ph.D.
Walter C. Willett, M.D., Dr.P.H.
Harvard School of Public Health
Boston, MA 02115
frank.hu@channing.harvard.edu
JoAnn E. Manson, M.D., Dr.P.H.
Brigham and Women's Hospital
Boston, MA 02115