Body Fat Distribution and 5-Year Risk of Death in Older Women

Aaron R. Folsom, Susan A. Kaye, Thomas A. Sellers, Ching Ping Hong, James R. Cerhan, John D. Potter, Ronald J. Prineas

Research output: Contribution to journalArticlepeer-review

540 Scopus citations


Objective.—To test the hypothesis that both body mass index (expressed as the ratio of weight in kilograms per height in meters squared) and the ratio of waist circumference to hip circumference are positively associated with mortality risk in older women. Design.—Prospective cohort study with a 5-year follow-up period. Setting.—General community. Participants.—Random sample of 41 837 Iowa women aged 55 to 69 years. Main Outcome Measure.—Total mortality (1504 deaths). Main Results.—Body mass index, an index of relative weight, was associated with mortality in a J-shaped fashion: rates were elevated in the leanest as well as in the most obese women. In contrast, waist/hip circumference ratio was strongly and positively associated with mortality in a dose-response manner. Adjusted for age, body mass index, smoking, education level, marital status, estrogen use, and alcohol use, a 0.15-unit increase in waist/hip circumference ratio (eg, a 15-cm [6-in] increase in waist measurement in a woman with 100-cm [40-in] hips) was associated with a 60% greater relative risk of death. The observed associations were not explained to any great degree by bias from weight loss prior to baseline or higher early deaths among lean participants. Conclusions.—Waist/hip circumference ratio is a better marker than body mass index of risk of death in older women. Waist/hip circumference ratio should be measured as part of routine surveillance and risk monitoring in medical practice.

Original languageEnglish (US)
Pages (from-to)483-487
Number of pages5
JournalJAMA: The Journal of the American Medical Association
Issue number4
StatePublished - Jan 27 1993


Dive into the research topics of 'Body Fat Distribution and 5-Year Risk of Death in Older Women'. Together they form a unique fingerprint.

Cite this