Weight Loss and Mortality

A number of studies relating "generic weight loss" (cause of weight loss unknown), "weight cycling" (cycles of weight loss followed by weight regain), and mortality have been published (308-311).  In most (308, 309, 311), but not all (310), of these studies, generic weight loss and weight cycling are associated with increases in mortality. None of these studies, however, differentiated between intentional and unintentional weight loss (311).  With the exception of the studies below, very little is currently known about factors related to intentional and unintentional weight loss in the general population or about the relationship between weight loss intention and mortality (312).

Two studies of factors related to weight loss intention have been carried out in the general population: French and colleagues assessed correlates of intentional and unintentional weight loss of > 20 lb in the Iowa Women's Health Study, a cohort study of approximately 29,000 women with a mean age of about 65 years (313-315); and Meltzer and Everhart analyzed data on 1-year self-reported weight change from approximately 9,000 participants in the nationally representative U.S. National Health Interview Survey, aged 45 years and above (316).

The results of these two studies suggest the following:

• In cross-sectional studies of weight loss recall, heavier persons are more likely to report intentional weight loss than unintentional weight loss, while the reverse is true for leaner persons;
   
• Intentional and unintentional weight losses occur with similar frequency in the U.S. population and contribute similarly to long-term weight fluctuation;
   
• The frequency of intentional weight loss is lower at older ages, while the frequency of unintentional weight loss is higher at older ages;
   
• Unintentional weight loss occurs more often in persons who report that their health status is poor, who use medications for chronic health conditions, and who smoke (312).

To date, only three studies have examined the relationship between intentional weight loss and mortality. Singh and colleagues (317) published results from a 1-year randomized controlled trial of a "cardioprotective diet" in East Indian patients hospitalized with recent myocardial infarction (mean age 50 years, mean BMI about 24 kg/m²). Although this study was not designed to specifically test the efficacy of intentional weight loss on lower mortality, the authors found that those who lost at least 0.5 kg (1.1 lb) had a 50 percent lower incidence of cardiac events and a 54 percent lower risk of overall mortality compared with counterparts who lost < 0.5 kg (1.1 lb) (317).

Williamson and colleagues (318) published a 12-year prospective observational study of weight loss and mortality that directly assessed weight loss intention. They analyzed data from 43,457 overweight (BMI > 27), never-smoking, white women ages 40 to 64 years. Mortality ratios were compared for women who intentionally lost weight with those for women who had no change in weight. In women with obesity-related comorbidities, intentional weight loss of any amount was associated with a statistically significant 20 percent reduction in all-cause mortality, primarily due to a significant 40 to 50 percent reduction in mortality from obesity-related cancers; diabetes-related mortality was also significantly reduced by 30 to 40 percent in those who intentionally lost weight. In women with no comorbidities, intentional weight loss was generally unrelated to mortality.  However, after subdividing intentional weight loss by time interval, it was found that a loss of at least 20 lb. that occurred within the previous year was associated with small to modest increases in mortality. The authors concluded that the association between intentional weight loss and longevity in middle-age overweight women depends on health status. In addition, preliminary evidence suggests that intentional weight loss in middle-age overweight men may be associated with a similar reduction of diabetes-related mortality as was observed in the overweight women (319).

The ongoing Swedish Obesity Study is a controlled trial of surgically induced weight loss and subsequent morbidity and mortality over a 10-year follow-up period (1,006 participants aged 37 to 57 years; initial BMI of 34 in men and 38 in women [(320, 321)]. Although the study is not randomized (participants self-select for surgery), the controls (who receive a behavioral weight loss program) are computer-matched to surgical participants on a large number of potential confounders including weight (320). In a preliminary abstract (322), the study reported the 2-year incidence rates shown in the table below.

These results for disease and risk factor incidences suggest that 10-year mortality will ultimately be lower in the surgical intervention group. Definitive mortality results have not been reported to date.