VA Study Claims Gastric Bypass Not Associated with Decreased Mortality

Today, Matthew Maciejewski PhD et al published an article in JAMA online titled, “Survival Among High-Risk Patients After Bariatric Surgery”. This was a retrospective cohort study of bariatric surgery programs in Veterans Affairs (VA) medical centers. Mortality was examined for 850 veterans who underwent gastric bypass surgery at 12 VA centers between 2000 and 2006. Using unique statistical modeling, all-cause mortality was compared to 41,244 patients who did not undergo surgery, at an average follow-up of 6.7 years. They found that the observed (crude) mortality rate was lower in the surgery group at 1, 2 and 6 years post-operatively compared with the nonsurgical group, and this difference remained statistically significant after covariate analysis. However, after further statistical manipulation of 1694 propensity-matched patients (847 per group), the mortality difference between the bariatric surgery group and the nonsurgical group was no longer significant.

We asked Dr. Emma Patterson for her opinions on this study. She said, “While this study addressed an important question, it had many limitations, and so the conclusions should be interpreted carefully. The authors do point out some of the limitations including:

1. They focused on predominantly male, sicker patients, so the results may not be generalizable to most bariatric surgery populations that are usually predominantly females.

2. This study did not include patients undergoing laparoscopic gastric banding procedures, so the results are not generalizable to those patients.

3. The sample size was limited in the 1:1 propensity score matching such that large confidence intervals were generated, which cannot rule out that bariatric surgery provides significant harm or benefit to some patients.

4. “The results do not account for unobserved confounding that may persist even after propensity score matching because the analysis was based on a quasi-experimental design from administrative data, not a randomized trial”, the authors of the study stated.

“In addition to these limitations, there are others”, adds Patterson:

5. The 12 VA centers in this study performed a total of 892 bariatric procedures over six years, which averages out to approximately 12 cases per year per center. This is very low volume, and several studies, including one from the NIH-funded LABS consortium, have shown a volume-outcome relationship in gastric bypass surgery.

6. The authors reported 11/850 (1.29%) deaths with 30 days of surgery, which is much higher than most centers currently report. This high perioperative mortality rate alone could explain why they did not find a reduction in mortality after surgery.

7. The follow-up period of six years may be too short, as the Swedish Obese Subjects study (a prospective clinical study), did not show a mortality benefit for 13 years.

Patterson summarizes: “At the end of the day, this study reported that VA medical centers doing gastric bypass surgery at low volumes did not find an association between surgery and reduced mortality at six years. As the authors point out, VA patients may still choose to have gastric bypass surgery, “given the strong evidence for significant reductions in body weight and comorbidities and improved quality of life”.