Associations between general and abdominal adiposity and mortality in individuals with diabetes mellitus

Diewertje Sluik*, Heiner Boeing, Jukka Montonen, Tobias Pischon, Rudolf Kaaks, Birgit Teucher, Anne Tjønneland, Jytte Halkjaer, Tina L. Berentzen, Kim Overvad, Larraitz Arriola, Eva Ardanaz, Benedetta Bendinelli, Sara Grioni, Rosario Tumino, Carlotta Sacerdote, Amalia Mattiello, Annemieke M.W. Spijkerman, Daphne L. Van Der A, Joline W. BeulensYvonne T. Van Der Schouw, Peter M. Nilsson, Bo Hedblad, Olov Rolandsson, Paul W. Franks, Ute Nöthlings

*Corresponding author for this work

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Individuals with diabetes mellitus are advised to achieve a healthy weight to prevent complications. However, fat mass distribution has hardly been investigated as a risk factor for diabetes complications. The authors studied associations between body mass index, waist circumference, waist/hip ratio, and waist/height ratio and mortality among individuals with diabetes mellitus. Within the European Prospective Investigation into Cancer and Nutrition, a subcohort was defined as 5,435 individuals with a confirmed self-report of diabetes mellitus at baseline in 1992-2000. Participants were aged 57.3 (standard deviation, 6.3) years, 54% were men, the median diabetes duration was 4.6 (interquartile range, 2.0-9.8) years, and 22% of the participants used insulin. Body mass index, as indicator of general obesity, was not associated with higher mortality, whereas all measurements of abdominal obesity showed a positive association. Associations generally were slightly weaker in women. The strongest association was observed for waist/height ratio: In the fifth quintile, the hazard rate ratio was 1.88 (95% confidence interval: 1.33, 2.65) for men and 2.46 (95% confidence interval: 1.46, 4.14) for women. Measurements of abdominal, but not general, adiposity were associated with higher mortality in diabetic individuals. The waist/height ratio showed the strongest association. Respective indicators might be investigated in risk prediction models.

Original languageEnglish
Pages (from-to)22-34
Number of pages13
JournalAmerican Journal of Epidemiology
Volume174
Issue number1
DOIs
Publication statusPublished - 1 Jul 2011

Cite this