Background & aims: A protective effect of obesity on death has been reported in the context of various co-morbidities. We studied if the obesity paradox applied to nursing home (NH) older residents according to dementia status.
Methods: Prospective data from 3741 NH residents from France. All-cause mortality was the dependent measure. Subjects were categorized according with body mass index (BMI) as underweight, normal-weight, overweight, and obese. Dementia status was obtained from medical charts. Cox regressions were performed.
Results: There were 344 (9.2%) residents who were underweight, 1367 (43.8%) normal weight, 1069 (28.6%) overweight and 691 (18.5%) obese. 1083 (28.9%) people died during follow-up. In residents with dementia, mortality risk was reduced by almost half in overweight and obese people (HRs of 0.60 [0.48-0.76] and 0.53 [0.38-0.75], respectively; p < 0.001), and increased in underweight (HR = 1.65 [1.29-2.12]; p < 0.001) compared to normal-weight residents; moreover, each 1 kg/m2 increase in BMI decreased the risk of death by 12% and 9% in underweight and normal-weight subjects with dementia. For people without dementia, mortality risk was reduced in overweight and obese people (HRs of 0.80 [0.65-0.99], p = 0.042, and 0.77 [0.60-0.99], p = 0.044, respectively) compared to normal-weight; the 1-unit increase in BMI reduced the risk of death (23% reduction) only in underweight people.
Conclusions: This study showed that the presence of dementia amplifies the obesity paradox in very old and functionally limited NH residents. Therefore, weight loss in NH residents, particularly in people with dementia, should be considered with extreme caution even for obese people.
Keywords: Mortality; Nursing home; Obesity paradox; Older adults; Overweight; Underweight.
Copyright © 2015 Elsevier Ltd and European Society for Clinical Nutrition and Metabolism. All rights reserved.