New Year’s Resolution: Lose Weight, Live Longer

Published - Written by Rena Xu

Obesity has long been recognized as a public health enemy.  The suspected health consequences of being overweight or obese run the gamut from heart disease to cancer to diabetes.  But how does being overweight or obese actually affect one’s risk of mortality?

This week, NEJM publishes the findings of a study by Berrington et al. that investigates the relation between body-mass index (BMI) and all-cause mortality in white (non-Hispanic) adults.  BMI is the weight in kilograms divided by the square of the height in meters; larger BMIs indicate obesity.  Whereas previous analyses of BMI and all-cause mortality may have included smokers and patients with preexisting diseases such as cancer and heart disease, the Berrington study systematically excluded these subgroups. Their rationale for doing so was that smoking and disease-related weight loss were possible confounding factors that might artifactually increase the calculated mortality associated with lower BMIs.


The authors performed a pooled analysis of 19 prospective studies for a total of 1.46 million white adults aged 19 to 84, 58% of whom were women. Within this population, 47% reported at baseline that they had never smoked, while 13% were self-reported current smokers.  Analyses were performed for all participants combined as well as for separate subgroups – for instance, for never-smokers, past smokers, and current smokers. The data were further adjusted for study, age, level of physical activity, alcohol consumption, education, and marital status.

Overall, the study data suggested that a BMI between 22.5 and 25 was associated with the lowest all-cause mortality, while being overweight or obese were both associated with increased all-cause mortality.  Moreover, as smokers and patients with cancer or heart disease at baseline were systematically excluded from the analysis, the shape of the hazard-ratio-to-BMI curve changed.  Specifically, the hazard ratios for high BMIs (25 or higher) steadily increased, while the hazard ratios for low BMIs (less than 22.5) steadily decreased.

Among healthy female never-smokers, the range associated with lowest all-cause mortality expanded to include BMIs between 20 and 25.  Meanwhile, hazard ratios were found to steadily increase for BMIs above this reference range.  The calculated hazard ratios were 1.13 for women who were overweight (BMI between 25 and 30), 1.44 for those in obesity class I (BMI between 30 and 35), 1.88 for those in obesity class II (BMI between 35 and 40), and 2.51 for those in obesity class III (BMI between 40 and 50).  Similar hazard ratios were calculated for male participants.

Increased hazard ratios were also seen for low BMIs (less than 20).  However, these were attenuated with longer-term follow-up, suggesting that these values might be artifacts of underlying disease processes as opposed to a true association.

These findings are largely consistent with those of previous studies, many of which also suggested an optimal BMI range of 20 to 25.  And while this study has its limitations – the height and weight data used to calculate BMI were self-reported, for instance, and changes in preexisting conditions were not captured over time – it is potentially valuable in reaffirming the impact of being overweight or obese on mortality.  Recognizing this association, some people may even find the additional motivation they need to shed those extra pounds this winter – and to keep them off.

The common saying is that you cannot be too thin; do you think these data really support that idea? Will the findings of this study affect how you treat patients in the overweight or obese categories (BMIs above 25)?