Survival After Bariatric Surgery Among High-Risk Patients

Abstract

To the Editor: Dr Maciejewski and colleagues reported that mortality was not reduced in severely obese patients in Veterans Affairs medical centers following bariatric surgery compared with a nonsurgical group of obese individuals. Their conclusion is in contrast to that of a meta-analysis of 8 previous reports comparing mortality risk following bariatric surgery with a variety of comparator groups. We are unaware of other evidence to suggest that substantial weight loss is not associated with a reduction in long-term mortality risk in such a population. The authors suggested several possible explanations for their different conclusion, emphasizing the predominantly male and somewhat older (mean age, 49.5 years) population studied. We believe that several methodological issues are a more likely explanation for the negative conclusion reached by Maciejewski et al. The mortality hazard ratio (HR) for the surgical patients after adjustment for baseline covariates was 0.80 (95% CI, 0.63-0.995). Following propensity score matching, the HR was 0.83 (95% CI, 0.61-1.14), minimally changed but no longer statistically significant. The statistical power was greatly reduced in this analysis, possibly accounting for the loss of significance. There should be longer follow-up before making a definitive statement regarding mortality benefit. The mean follow-up in the study was 6.7 years. Two larger studies from Utah and Sweden had mean follow-up times of 7.1 years and 10.9 years, with HRs reaching 0.60 and 0.76, respectively. The survival benefit became progressively greater as the years of follow-up continued. Another possible factor was the relatively high perioperative mortality. The authors suggested the reason for the higher perioperative mortality rates was their use of a cohort of primarily male and somewhat older patients than previously studied. Another plausible explanation is the relatively low volume of bariatric surgeries performed as evidenced by a mean number of 10.6 per center per year. Lower hospital volume of obesity surgery has been associated with higher mortality.