November 30, 2017
2 min read

Link between obesity, mortality weakens as patients age

You've successfully added to your alerts. You will receive an email when new content is published.

Click Here to Manage Email Alerts

We were unable to process your request. Please try again later. If you continue to have this issue please contact


Associations between obesity and mortality decreased as patients aged, according to findings recently published in Obesity.

Researchers wrote that the findings overturned previous studies that indicated that age strengthened the association between obesity and mortality.

“BMI as a measure of obesity is associated with mortality,” Zhiqiang Wang, PhD, of the University of Queensland, Brisbane, Australia, and colleagues wrote. “Most studies have found that the association of obesity and mortality weakens with age, but several recent studies have argued that the association strengthens with increasing age after controlling for age at the survey and for birth cohort membership.”

In studies reporting that age strengthens the relationship between obesity and mortality, Wang and colleagues wrote, authors used multiple regression analysis. However, calculating HRs with this method “can be complicated because the effect of obesity on mortality depends on the values of several coefficients and variables, including age at the survey,” they wrote.

In the current analysis, researchers used the National Health Interview Survey to gather data on 240,184 white men aged 40 to 84 years and 51,697 participants with grade 2 or 3 obesity aged 25 to 84 years who were matched with participants of normal weight. They calculated age-specific HRs of mortality for grade 2 or grade 3 obesity compared with a BMI between 18.5 kg/m2 and 25 kg/m2 using a flexible parametric survival model for participants and Cox proportional hazard models for the matched pairs.

HRs seemed to increase with age when the model included age at the time of the survey as a single value, the researchers reported; however, when Wang and colleagues included adults with various ages during the survey, HRs were higher among younger adults compared with older adults with the same duration of follow-up.

In the matched cohort, HRs were higher for younger patients (age 40 to 49 years; HR = 2.14; 95% CI, 1.9-2.4) than older adults (age 90 years or older; HR = 1.22; 95% CI, 0.91-1.63).

In an accompanying editorial, Jay S. Kaufman, PhD, of the department of epidemiology, biostatistics and occupational health at McGill University, wrote that though the study by Wang and colleagues demonstrates how using an alternative model specification can overturn a pattern of results, the data do not resolve etiologic questions surrounding the association between obesity and age, including a causal interpretation.

“The prospects for this whole research question look rather discouraging when approached by using cohort data in which people of various ages report a baseline BMI at the time of interview and are then followed up for mortality of subsequent decades,” he wrote.

“As with most observational epidemiology, thinking in terms of a ‘target trial’ would go a long way to illuminating most of the deficiencies of these existing studies, although it is not obvious that these can be surmounted with any ethical or feasible design. Rather, we could probably make more progress by focusing on the study of interventions like weight-loss programs or bariatric surgery. It’s not hazard ratios, but rather epidemiologic designs that really need to get stronger over time.” – by Andy Polhamus

Disclosures: Kaufman reports no relevant financial disclosures. The authors report no relevant financial disclosures.