The Changing Body Mass–Mortality Association in the United States: Evidence of Sex-Specific Cohort Trends from Three National Health and Nutrition Examination Surveys

Yan Yu

Research output: Contribution to journalArticlepeer-review

6 Citations (Scopus)
1 Downloads (Pure)

Abstract

The association between body mass index (BMI) categories and mortality remains uncertain. Using three National Health and Nutrition Examination Surveys covering the 1971–2006 period for cohorts born between 1896 and 1968, this study estimates separately for men and women models for year-of-birth (cohort) and year-of-observation (period) trends in how age-specific mortality rates differ across BMI categories. Among women, relative to the normal weight (BMI 18.5–24.9 kg/m2), there are increasing trends in mortality rates for the overweight (BMI 25–29.9) or obese (BMI ≥ 30). Among men, mortality rates relative to the normal weight decrease for the overweight, do not change for the moderately obese (BMI 30–34.9), and increase for the severely obese (BMI ≥ 35). Period and cohort trends are similar, but the cohort trends are more consistent. In the latest cohorts, compared with the normal weight, mortality rates are 50 percent lower for overweight men, not different for moderately obese men, and 100–200 percent higher for severely obese men and for overweight or obese women. For U.S. cohorts born after the 1920s, a lower overweight than normal weight mortality is confined to men. I speculate on possible reasons why the mortality association with overweight and obesity varies by sex and cohort
Original languageEnglish
Pages (from-to)143-163
Number of pages21
JournalBiodemography and Social Biology
Volume62
Issue number2
DOIs
Publication statusPublished - 2016
Externally publishedYes

Fingerprint

Dive into the research topics of 'The Changing Body Mass–Mortality Association in the United States: Evidence of Sex-Specific Cohort Trends from Three National Health and Nutrition Examination Surveys'. Together they form a unique fingerprint.

Cite this