Associations of body mass index and waist circumference with all cause mortality in the oldest old with cognitive impairment: a prospective cohort study.

Importance: In clinical practice, reducing body mass index (BMI), and waist circumference (WC) is a crucial treatment target for minimizing health risks. However, the association patterns between BMI, WC, and all cause mortality in cognitively impaired older adults remain unknown. Objective: To investigate the association patterns between body mass index (BMI), waist circumference (WC), and all cause mortality among cognitively impaired oldest old. Design setting and participants: The cognitively impaired oldest old from the Chinese Longitudinal Healthy Longevity Survey (CLHLS) in the 2011-2014 wave was included. A restricted cubic spline based on Cox proportional hazards model was used to examine the association patterns. Exposures: The global cognitive function of participants was assessed by the Minimum-Mental State Examination. Main outcomes and measures: All cause mortality was the outcome. Results: A total of 2,124 participants (1,522 females [71.7%]) were included in this study and 1,071 (50.4%) deaths were documented. We found J shaped association between levels of BMI and all cause mortality in cognitively impaired oldest old, with low BMI levels associated with increased mortality risk whereas high BMI levels associated with reduced mortality risk. Compared to BMI in quartile 4, the multivariable adjusted hazards ratios of all cause mortality were 1.53 (95% confidence interval, 1.28 to 1.83), 1.36 (1.13 to 1.64), and 1.31 (1.09 to 1.57), respectively, for BMI in quartiles 1, 2 and 3. We also observed a linear association between levels of WC and all cause mortality in the cognitively impaired oldest old, with low levels associated with high mortality risk and high levels associated with low mortality risk. In contrast with lower WC (quartile 1), the hazard ratios for all cause mortality were 0.82 (0.70 to 0.97) for quartile 2, 0.77 (0.65 to 0.91) for quartile 3, and 0.70 (0.59 to 0.83) for quartile 4, respectively. Joint association analyses revealed that participants in the highest quartile of BMI and the highest quartile of WC had the lowest mortality risk. Conclusions and relevance: Among the cognitively impaired oldest old, we found a J shaped association between BMI and all-cause mortality, and a linear association between WC and all-cause mortality, with increased levels of each associated with reduced mortality risk. Contrary to clinical practices that aim to reduce BMI and WC to minimize health risks, this study emphasizes the importance of maintaining higher BMI and WC levels in cognitively impaired oldest old.