Dementia and BMI: An Unexpected Finding Reveals That Higher Weight Is Associated With Lower Mortality

An unexpected discovery is raising new questions about body weight and dementia. In a long-term study conducted in the United States, individuals who were overweight or obese at the time of diagnosis had a lower risk of death than those with a weight considered normal. The findings suggest that body weight may offer clues regarding prognosis, but they do not demonstrate that obesity protects against death or that gaining weight is beneficial.

Dementia is a condition that progressively affects brain functions such as memory, reasoning, language, and the ability to perform daily activities. While age, coexisting medical conditions, and other factors influence the progression of the disease, research has also explored whether body weight might be linked to survival following a diagnosis.

Some studies have observed that individuals who are overweight or obese appear to live longer after a dementia diagnosis than those whose weight is considered normal.

This result may seem surprising, given that excess weight is associated with various health problems. To better understand this relationship, researchers investigated whether body weight at the time dementia was identified was linked to the risk of death, and whether this association could be explained by other factors.

Scientists used data from a large sample of the United States population, tracked over an extended period. The objective was to compare the survival of individuals diagnosed with dementia across different body weight categories. To do this, they utilized the body mass index (BMI), a metric calculated by dividing a person’s weight (in kilograms) by the square of their height (in meters).

This index allows for the classification of individuals into categories such as normal weight, overweight, and various degrees of obesity. The researchers then monitored the participants to determine how long they lived following their diagnosis and compared the risk of death across these categories.

The analysis accounted for the statistical association between weight and mortality, while also seeking to identify potential biases that could make this relationship appear stronger, or different, than it actually is.

One of the factors investigated was weight loss linked to health issues occurring prior to a dementia diagnosis. Certain medical conditions can lead to involuntary weight loss before they are formally identified; furthermore, dementia itself may be associated with changes in eating habits, difficulties in preparing meals, or weight loss as the condition progresses. Consequently, an individual classified as having a "normal" weight at the time of diagnosis might have previously lost a significant amount of weight due to an underlying health problem.

In this scenario, a "normal" weight would not necessarily be the cause of lower survival rates; rather, it could serve as an indicator of an already compromised health condition. To investigate this possibility, the researchers assessed whether weight loss and other factors might account for the observed association.

They also examined two more complex statistical issues: selection bias, which can arise when the inclusion of individuals in a study depends on characteristics related to both the exposure and the outcome, and survival bias, which occurs when an analysis includes only those who survived long enough to be observed. These phenomena can create the appearance of an association even when no direct causal relationship exists.

The results showed that individuals with class I obesity at the time of diagnosis had a 27% lower risk of death compared to those classified as having a "normal" weight. The risk was also lower among participants who were overweight or had class II or III obesity.

To reach this conclusion, the researchers employed a statistical measure known as the hazard ratio, which compares the occurrence of an event, in this case, death during the follow-up period, between two groups.

For class I obesity, the estimate was 0.73, indicating a relative risk approximately 27% lower during the period analyzed. This does not mean that every individual with obesity will live longer, nor that their life expectancy will increase by 27%.

This is an observed difference between groups, not an individual prediction. Furthermore, additional analyses did not suggest that the biases investigated were the primary drivers of the observed association.

Despite these results, researchers do not yet know exactly why individuals with higher body weight exhibited lower mortality following a dementia diagnosis. One possibility is that body weight serves as an indirect indicator of energy reserves and nutritional status, particularly in a condition that can make eating difficult and lead to weight loss. However, this hypothesis was not confirmed by the study.

Other biological factors, differences among participants, and characteristics of the disease itself may also contribute to these findings.

Moreover, the lack of evidence that specific biases account for this association does not rule out all potential sources of distortion. The main conclusion is that body weight at the time of diagnosis may provide additional information regarding prognosis, but it must be interpreted in the context of age, nutritional status, coexisting medical conditions, and clinical progression.

The study does not demonstrate that being overweight or obese protects against death, nor does it justify recommending weight gain as a strategy to extend the lives of people with dementia.

READ MORE:

Body mass index and mortality in incident dementia

Virginia W. Chang, Eleanor Hayes-Larson, Hui Zheng, and Yuan S. Zhang

Journal of Alzheimer’s Disease. Oct 1, 2026

DOI: 10.1177/13872877261491694

Abstract: 

Limited evidence suggests that higher body weight may be associated with improved survival among individuals with dementia relative to those in the normal weight range. The mechanisms underlying this association are uncertain and may include both physiological and methodological factors.

To examine the association between weight status at incident dementia and mortality and to assess potential sources of bias. We examined the association between weight status at incident dementia and mortality in a large, U.S. population-based sample with extended follow-up. We also evaluated potential sources of bias affecting this association, including confounding due to weight loss, collider stratification bias, and survivor bias. Compared with the normal weight category, class I obesity at incident dementia was associated with a 27% lower risk of mortality (HR 0.73, 95% CI: 0.62–0.86). Overweight and class II/III obesity were also associated with lower mortality. Analyses evaluating potential sources of bias did not support these mechanisms as major contributors to the observed associations. In a large, U.S. population-based sample with extended follow-up, overweight and obesity at incident dementia were associated with lower mortality relative to normal weight. While the underlying mechanisms are not fully understood, these findings suggest that weight status at diagnosis may serve as a useful prognostic indicator alongside other factors.