Body mass index and mortality in women: Follow-up of the Canadian National Breast Screening Study cohort

M. G. Jain, A. B. Miller, T. E. Rohan, J. T. Rehm, S. J. Bondy, M. J. Ashley, J. E. Cohen, R. G. Ferrence

Research output: Contribution to journalArticlepeer-review

15 Scopus citations


OBJECTIVE: The present study was conducted to examine the relationship between obesity and all-cause mortality in women. STUDY DESIGN AND SETTINGS: The subjects were women enrolled from 1980 to 1985 in a Canadian randomized trial, the National Breast Screening Study (NBSS) to evaluate the efficacy of mammographie screening. Mortality was ascertained by record linkage to the Canadian Mortality Data Base. Hazard ratios (HR) for the association between body mass index (BMI) and all-cause mortality were obtained from Cox proportional hazard regression models. RESULTS: During an average follow-up period of 16.5 years, 2566 deaths were identified among the 49165 women, age 40-59 y at enrolment. The risk of all-cause death increased linearly above a BMI of 22 kg/m2 and the trend was statistically significant. The HR (and 95% confidence intervals) in the various categories of BMI (kg/m2) were: BMI < 18.5: 1.12 (0.99-1.25); BMI 18.5-21.9: 1.00 (reference); BMI 22-24.9: 1.15 (1.11-1.18); BMI 25.0-27.9: 1.28 (1.24-1.32); BMI 28.0-29.9: 1.34 (1.29-1.39); BMI 30.0-34.9: 1.30 (1.25-1.35); and BMI ≥ 35.0: 1.40 (1.33-1.47). CONCLUSION: This study confirms the association of high BMI with increased all-cause mortality in women.

Original languageEnglish (US)
Pages (from-to)792-797
Number of pages6
JournalInternational Journal of Obesity
Issue number7
StatePublished - Jul 2005


  • Body mass index
  • Cohort
  • Mortality
  • Trial
  • Women

ASJC Scopus subject areas

  • Medicine (miscellaneous)
  • Endocrinology, Diabetes and Metabolism
  • Nutrition and Dietetics


Dive into the research topics of 'Body mass index and mortality in women: Follow-up of the Canadian National Breast Screening Study cohort'. Together they form a unique fingerprint.

Cite this