Loss of height predicts total and cardiovascular mortality: a cohort study of northern European women

OBJECTIVE: To examine height changes in middle-aged northern European women in relation to overall and cardiovascular mortality. DESIGN: Population-based cohort studies with longitudinally measured heights and register-based mortality. SETTING: Sweden and Denmark. PARTICIPANTS: Population-based samples of 2406 Swedish and Danish women born on selected years in 1908-1952, recruited to baseline examinations at ages 30-60, and re-examined 10-13 years later. MAIN OUTCOME MEASURE: Total and cardiovascular disease (CVD) specific mortality during 17-19 years of follow-up after last height measure. RESULTS: For each 1 cm height loss during 10-13 years, the HR (95% CI) for total mortality was 1.14 (1.05 to 1.23) in Swedish women and 1.21 (1.09 to 1.35) in Danish women, independent of key covariates. Low height and high leisure time physical activity at baseline were protective of height loss, independent of age. Considering total mortality, the HR for major height loss, defined as height loss greater than 2 cm, were 1.74 (1.32 to 2.29) in Swedish women and 1.80 (1.27 to 2.54) in Danish women. Pooled analyses indicated that height loss was monotonically associated with an increased mortality, confirming a significant effect above 2 cm height loss. For cause-specific mortality, major height loss was associated with a HR of 2.31 (1.09 to 4.87) for stroke mortality, 2.14 (1.47 to 3.12) for total CVD mortality and 1.71 (1.28 to 2.29) for mortality due to causes other than CVD. CONCLUSION: Height loss is a marker for excess mortality in northern European women. Specifically the hazard of CVD mortality is increased in women with height loss during middle age, and the results suggest that the strongest cause-specific endpoint may be stroke mortality. The present findings suggest attention to height loss in early and mid-adulthood to identify women at high risk of CVD, and that regular physical activity may prevent early onset height loss.