Elsevier

Annals of Epidemiology

Volume 14, Issue 10, November 2004, Pages 731-739
Annals of Epidemiology

Original Report
Body mass index and mortality among US male physicians

https://doi.org/10.1016/j.annepidem.2003.10.008Get rights and content

Abstract

Purpose

To assess the relationship between body mass index and mortality in a population homogeneous in educational attainment and socioeconomic status.

Methods

We analyzed the association between body mass index (BMI) and both all-cause and cause-specific mortality among 85,078 men aged 40 to 84 years from the Physicians' Health Study enrollment cohort.

Results

During 5 years of follow-up, we documented 2856 deaths (including 1212 due to cardiovascular diseases and 891 due to cancer). In age-adjusted analyses, we observed a U-shaped relation between BMI and all-cause mortality; among men who never smoked a linear relation was observed with no increase in mortality among leaner men (P for trend, <0.001). Among never smokers, in multivariate analyses adjusted for age, alcohol intake, and physical activity, the relative risks of all-cause mortality increased in a stepwise fashion with increasing BMI. Excluding the first 2 years of follow-up further strengthened the association (multivariate relative risks, from BMI<20 to ⩾30 kg/m2, were 0.93, 1.00, 1.00, 1.16, 1.45, and 1.71 [P for trend, <0.001]). In all age strata (40–54, 55–69, and 70–84 years), never smokers with BMIs of 30 or greater had approximately a 70% increased risk of death compared with the referent group (BMI 22.5–24.9). Higher levels of BMI were also strongly related to increased risk of cardiovascular mortality, regardless of physical activity level (P for trend, <0.01).

Conclusions

All-cause and cardiovascular mortality was directly related to BMI among middle-aged and elderly men. Advancing age did not attenuate the increased risk of death associated with obesity. Lean men (BMI<20) did not have excess mortality, regardless of age.

Introduction

Obesity is a burgeoning health problem in both developed and developing countries. In the United States, national surveys have documented an increase in obesity (defined as body mass index ⩾ 30 kg/m2) despite the dissemination and adoption of a wide variety of fat-free and low-fat foods. The 2000 Behavioral Risk Factor Surveillance System estimated an increase in prevalence of obesity among adults age 18 or older from 12.0% in 1991 to 19.8% in 2000 (1). Another large survey, the National Health and Nutrition Examination Survey (NHANES) reported an increase in prevalence of obesity among adult men and women aged 20 to 74 years (2) from 14.5% between 1971 and 1974 (NHANES II) to 30.9% in most recent data from NHANES 1999–2000. Recent surveys from countries in Asia and South America show that these populations are also becoming more obese (3). Such trends are alarming, given the clear association between obesity and morbidity (4) and mortality 5., 6..

While there is minimal controversy concerning the mortality hazards for individuals at the upper end of the weight distribution curve, there is still some dispute concerning hazards for those at the lower end of the curve. Previous studies have not demonstrated a clear relationship between body weight and mortality—some studies have shown no association 7., 8., 9., 10., 11. and others have shown a direct linear association 12., 13., 14., 15., 16., 18., 19., 20.. Still others have found J-shaped 7., 8., 9., 17., 21., 22., 23. or U-shaped curves 24., 25., 26., 27., 28., 29., 30., 31., 32. which suggest that both lean and obese individuals are at increased risk of all-cause mortality.

The relationship between body weight and mortality can be obscured by several potential biases and confounders. These include failure or inability to control for cigarette smoking (smokers tend to be lean and also have elevated mortality rates), failure or inability to control for subclinical disease (underlying disease may cause weight loss and also increase mortality), and over-adjustment for intermediate risk factors such as hypertension, diabetes, and high cholesterol (biological effects of obesity that mediate its influence on mortality, rather than confounding variables) (33). Moreover, in most studies demonstrating a direct association between body weight and mortality, the relationship was observed predominantly among younger and middle-aged participants, but not among the elderly 11., 17., 18., 34..

The Physicians' Health Study enrollment cohort afforded a unique opportunity to assess the relationship between body mass index and all-cause mortality, as well as mortality from specific causes, among more than 85,000 US male physicians aged 40 to 84 years. In this cohort, confounding by educational attainment, socioeconomic status, and access to health care was minimized and we were able to examine factors such as age, cigarette smoking, and physical activity that may influence the association between body mass index and mortality.

Section snippets

Study population

The Physicians' Health Study was a randomized, double-blind, placebo-controlled trial testing the role aspirin and β-carotene in primary prevention of cardiovascular diseases and cancer. A detailed description of the study design has been published elsewhere 35., 36.. Briefly, in 1982 and 1983, letters of invitation, informed consent forms, and baseline questionnaires were mailed to 261,248 eligible US male physicians aged 40 to 84 years. By December 31, 1983, 112,528 men had responded to the

Results

During 5 years of follow-up, we identified 2856 deaths. Of these, 1212 were due to cardiovascular diseases, 891 were due to cancer, and 753 were due to other causes. In Table 1, we present the baseline characteristics of the 85,078 participants according to BMI categories at baseline. As expected, a greater proportion of participants in higher BMI categories reported a positive history of hypertension, high cholesterol, and diabetes, while fewer subjects reported engaging in vigorous physical

Discussion

In this cohort of apparently healthy male physicians aged 40 to 84 years at baseline who were followed for 5 years, we observed a linear trend between body mass index and all-cause mortality after accounting for pre-existing disease. This linear trend was observed particularly among those who never smoked, an alternate method of accounting for effects of smoking. Smoking status was a powerful confounder of the association between body mass index and mortality and therefore, simply controlling

Acknowledgements

The authors acknowledge the contributions of the entire staff of the Physicians' Health Study, including Charlene Belanger, Mary Breen, Vadim Bubes, Jean MacFayden, Geneva McNair, David Potter, Leslie Power, Harriet Samuelson, Miriam Schvartz, Mickie Sheehey, Joanne Smith, Phyllis Johnson Wojciechowski, P.J. Skerrett, and Philomena Quinn for their expert help. We are also indebted to all dedicated and committed participants of this study.

References (43)

  • W.C. Willett et al.

    Guidelines for healthy weight

    N Engl J Med

    (1999)
  • M. Schroll

    A longitudinal epidemiological survey of relative weight at age 25, 50, and 60 in the Glostrup population of men and women born in 1914

    Dan Med Bull

    (1981)
  • J.P. Vandenbroucke et al.

    Weight, smoking, and mortality

    JAMA

    (1984)
  • J. Tuomilehto et al.

    Body weight and risk of myocardial infarction and death in the adult population of eastern Finland

    Br Med J (Res Clin Ed)

    (1987)
  • J. Stevens et al.

    Body mass index and body girths as predictors of mortality in black and white women

    Arch Intern Med

    (1992)
  • P. Diehr et al.

    Body mass index and mortality in nonsmoking older adults: The Cardiovascular Health Study

    Am J Public Health

    (1998)
  • Society of Actuaries

    Build and Blood Pressure Study 1959

    (1959)
  • R.J. Garrison et al.

    Cigarette smoking as a confounder of the relationship between relative weight and long-term mortality. The Framingham Heart Study

    JAMA

    (1983)
  • K. Lindsted et al.

    Body mass index and patterns of mortality among Seventh-day Adventist men

    Int J Obes

    (1991)
  • I.M. Lee et al.

    Body weight and mortality. A 27-year follow-up of middle-aged men

    JAMA

    (1993)
  • J.E. Manson et al.

    Body weight and mortality among women

    N Engl J Med

    (1995)
  • Cited by (131)

    • Interrelations between body mass index, frailty, and clinical adverse events in older community-dwelling women: The EPIDOS cohort study

      2018, Clinical Nutrition
      Citation Excerpt :

      In Europe, 48% of people aged 65 and older were overweight. Obesity is associated with higher mortality in middle-aged individuals [1–4], but whether the same is true in older people is unclear [1,5–9]. Whether obesity is a risk factor for clinical adverse events such as death, falls, fracture or hospital admission remains controversial in older persons and may depend on the underlying medical condition.

    • From Genome-Wide Association Study to Phenome-Wide Association Study: New Paradigms in Obesity Research

      2016, Progress in Molecular Biology and Translational Science
      Citation Excerpt :

      Obesity is an abnormal accumulation of body fat (>20% over an individual's body weight), which causes many health and medical problems worldwide.1–5

    • Age-dependent association between obesity and mortality

      2015, Obesity Research and Clinical Practice
    View all citing articles on Scopus

    Supported by a grant (HL-42441) from the National Institutes of Health, Bethesda, MD, USA. Dr. Lotufo is the recipient of a fellowship (97-02424-8) from FAPESP (Fundação de Amparo a Pesquisa do Estado de São Paulo), São Paulo, Brazil.

    View full text