TY - JOUR
T1 - Changes in the Relationships Between Body Mass Index and Health Outcomes Across Middle Age and Older Adulthood
AU - Peeters, G. M.E.E.
AU - Herber-Gast, Gerrie Cor M.
AU - Dobson, Annette J.
AU - Brown, Wendy J.
PY - 2015/7/1
Y1 - 2015/7/1
N2 - Objective: To examine patterns of the incidence of diabetes, hypertension, and mortality by single units of body mass index (BMI) and 5-year age groups using longitudinal data from middle-aged and older women. Patients and Methods: Middle-aged (born between 1946 and 1951; N=13,715) and older (born between 1921 and 1926; N=12,432) participants in the Australian Longitudinal Study on Women's Health completed surveys in 1996 and at approximately 3-year intervals thereafter until 2011. Proportions of women with diabetes, hypertension, and mortality over 3-year intervals were estimated for each unit of BMI and 5-year age group (45 to <50, 50 to <55, 55 to <60, 70 to <75, 75 to <80, and 80 to <85 years) using generalized additive modeling with adjustment for time-varying covariates. Results: Three-year incidence of diabetes (1.2%-3.6%), hypertension (5.2%-17.8%), and death (0.4%-9.5%) increased with age. For both diabetes and hypertension, the associations with BMI were curvilinear in middle-aged women and became almost linear in older women. With increasing age, the slope became steeper, and the increase started at lower BMI values. For hypertension, there was a marked increase in intercept from 75 years onward. In contrast, mortality risks were highest for low BMI (≤20) in all age groups. A clear U-shaped curve was observed only in the oldest age group. Conclusion: The shapes of the relationships between BMI and incidence of diabetes, hypertension, and mortality change with age, suggesting that weight management interventions should be tailored for different age groups.
AB - Objective: To examine patterns of the incidence of diabetes, hypertension, and mortality by single units of body mass index (BMI) and 5-year age groups using longitudinal data from middle-aged and older women. Patients and Methods: Middle-aged (born between 1946 and 1951; N=13,715) and older (born between 1921 and 1926; N=12,432) participants in the Australian Longitudinal Study on Women's Health completed surveys in 1996 and at approximately 3-year intervals thereafter until 2011. Proportions of women with diabetes, hypertension, and mortality over 3-year intervals were estimated for each unit of BMI and 5-year age group (45 to <50, 50 to <55, 55 to <60, 70 to <75, 75 to <80, and 80 to <85 years) using generalized additive modeling with adjustment for time-varying covariates. Results: Three-year incidence of diabetes (1.2%-3.6%), hypertension (5.2%-17.8%), and death (0.4%-9.5%) increased with age. For both diabetes and hypertension, the associations with BMI were curvilinear in middle-aged women and became almost linear in older women. With increasing age, the slope became steeper, and the increase started at lower BMI values. For hypertension, there was a marked increase in intercept from 75 years onward. In contrast, mortality risks were highest for low BMI (≤20) in all age groups. A clear U-shaped curve was observed only in the oldest age group. Conclusion: The shapes of the relationships between BMI and incidence of diabetes, hypertension, and mortality change with age, suggesting that weight management interventions should be tailored for different age groups.
UR - http://www.scopus.com/inward/record.url?scp=84934299930&partnerID=8YFLogxK
U2 - 10.1016/j.mayocp.2015.04.006
DO - 10.1016/j.mayocp.2015.04.006
M3 - Article
C2 - 26141330
AN - SCOPUS:84934299930
SN - 0025-6196
VL - 90
SP - 903
EP - 910
JO - Mayo Clinic Proceedings
JF - Mayo Clinic Proceedings
IS - 7
ER -