Researchers from Columbia University's Mailman School of Public Health and NYU School of Global Public Health report that body weight at the time someone is diagnosed with dementia may offer clues about how long that person is likely to live. Their findings were published in the Journal of Alzheimer's Disease. Compared with people in the normal weight range, those who were overweight or obese at diagnosis had a significantly lower risk of death, even after accounting for smoking, existing medical conditions, and demographic characteristics. Overweight people (BMI 25 to 29.9) had a 16 percent lower risk of death, people with class 1 obesity (BMI 30 to 34.9) had a 27 percent lower risk, and people with class 2 and 3 obesity (BMI 35 or higher) had a 22 percent lower risk.
The result runs against expectations, since obesity is well established as a risk factor for heart disease and diabetes, and midlife overweight is linked to a higher chance of later dementia. Earlier research had hinted that the relationship might flip once dementia develops. The authors, led by Yuan Zhang and first author Virginia Chang, were careful to test whether the pattern was an artifact of bias. They drew on the Health and Retirement Study, which follows Americans aged 51 and older, and analyzed records from 1998 to 2022 for 4,523 participants who developed dementia during the study. Because the data spanned many years, the team could track weight changes before diagnosis, which matters because earlier work by the same researchers found that weight loss can begin as much as 10 years before a diagnosis. That pattern could make people who are already declining look thin and more at risk, flattering heavier patients by comparison. The team also examined survivor bias and other distortions, and their additional analyses suggested these biases were unlikely to explain the link.
The researchers offered possible biological explanations, including that stored fat or muscle may help patients withstand the weight loss, muscle wasting, and feeding difficulties that dementia can bring. Chang described how changes in the brain can reduce appetite and initiative, raise energy needs, and make it hard to feed oneself. However, the study did not establish that these factors directly account for the survival differences, and the authors say they cannot yet say for certain why the pattern exists. They are explicit that the findings are not a recommendation for people with dementia to gain weight, since the study was not designed to test that. What they do suggest is that weight at diagnosis may be a useful prognostic signal and that nutrition deserves attention in people living with dementia.