Heterogeneity of weight loss after gastric bypass, sleeve gastrectomy, and adjustable gastric banding

Abstract

Background

Laparoscopic Roux-en-Y gastric bypass, laparoscopic sleeve gastrectomy, and laparoscopic adjustable gastric banding all lead to substantial weight loss in obese patients. Long-term weight loss can be highly variable beyond 1-year postsurgery. This study examines and compares the frequency distribution of weight loss and lack of treatment effect rates after laparoscopic Roux-en-Y gastric bypass, laparoscopic sleeve gastrectomy, and laparoscopic adjustable gastric banding.

Methods

A total of 1,331 consecutive patients at a single academic institution were reviewed from a prospectively collected database. Preoperative data collected included demographics, body mass index, and percent excess weight loss. Postoperative BMI and %EWL were collected at 12, 24, and 36 months. Percent excess weight loss was analyzed by the percentiles of excess weight lost, and the distribution of percent excess weight loss was evaluated in 10% increments. Lack of a successful treatment effect was defined as <25% excess weight loss.

Results

Of the 1,331 patients, 72.4% (963) underwent laparoscopic Roux-en-Y gastric bypass, 18.3% (243) laparoscopic sleeve gastrectomy, and 9.4%(125) laparoscopic adjustable gastric banding. Mean percent excess weight loss was greatest for laparoscopic Roux-en-Y gastric bypass, followed by laparoscopic sleeve gastrectomy, and then by laparoscopic adjustable gastric banding at every time point: at 2 years mean percent excess weight loss was 77.9± 24.4 for laparoscopic Roux-en-Y gastric bypass, 50.8 ± 25.8 for laparoscopic sleeve gastrectomy, and 40.8± 25.9 for laparoscopic adjustable gastric banding (P < .0001). The rates of a successful treatment effect s for laparoscopic Roux-en-Y gastric bypass, laparoscopic sleeve gastrectomy, and laparoscopic adjustable gastric banding were 0.9%, 5.2%, and 24.3% at 1 year; 0.3%, 11.1%, and 26.0% at 2 years; and 1.0%, 25.3%, and 30.2% at 3 years. At 1 year, the odds ratio of lack of a successful treatment effect of laparoscopic sleeve gastrectomy versus laparoscopic Roux-en-Y gastric bypass was 6.305 (2.125–19.08; P = .0004), the odds ratio for laparoscopic adjustable gastric banding versus laparoscopic Roux-en-Y gastric bypass was 36.552 (15.64–95.71; P < .0001), and the odds ratio for laparoscopic adjustable gastric banding versus laparoscopic sleeve gastrectomy was 5.791 (2.519–14.599; P < .0001). At 2 years, the odds ratio for laparoscopic sleeve gastrectomy versus laparoscopic Roux-en-Y gastric bypass increased to 70.7 (9.4–531.7; P < .0001), the odds ratio for laparoscopic adjustable gastric banding versus laparoscopic Roux-en-Y gastric bypass increased to 128.1 (16.8–974.3; P < .0001), and the odds ratio for laparoscopic adjustable gastric banding versus laparoscopic sleeve gastrectomy decreased to 1.8 (0.9–3.6; P = .09).

Conclusion

This study emphasizes the existing variability in weight loss across bariatric procedures as well as in the lack of a treatment effect for each procedure. Although laparoscopic adjustable gastric banding has the greatest rate of a lack of a successful treatment effect, the rate remained stable over 3 years postoperatively. Laparoscopic sleeve gastrectomy showed a doubling in the rate of a lack of a successful treatment effect every year reaching 25% at year 3. The rates for lack of a successful treatment effect for laparoscopic Roux-en-Y gastric bypass remained stable at about 1% for the first 3 years postoperatively.

Introduction

Bariatric surgery has been shown to be an effective and enduring long-term treatment for morbid obesity.¹ Laparoscopic Roux-en-Y gastric bypass (LRYGB) has remained at the forefront of bariatric procedures for the past 2 decades, with excellent results.² The laparoscopic adjustable gastric banding (LAGB) has seen more variation in overall results.³ Lately, the laparoscopic sleeve gastrectomy (LSG) has become widely adopted, recently surpassing LRYGB as the most commonly performed bariatric procedure in the United States according to the American Society for Metabolic and Bariatric Surgery.⁴ The current literature reports outcomes after bariatric surgery using metrics such as percent excess body weight loss, percent excess body mass index loss, total absolute weight loss, and percent excess weight loss (%EWL). Although each of these metrics of weight loss has its advantages and limitations, all are typically reported using the mean ± standard deviation. Although reporting means and standard deviations as measures of central tendency provides an adequate tool for overall comparisons, these statistics often only partially describe their corresponding distributions. Reports of weight loss as mean ± standard deviation of continuous variables (whether percent or actual weight loss) assume a normal distribution of the data, but if the data are not normally distributed, different statistical tools are needed to analyze the reported data. Therefore, it is of importance to know the variability that exists regarding weight loss after each weight loss procedure and to evaluate if these data follow a normal distribution.

In clinical practice, variation in weight loss between patients undergoing the same bariatric procedure is commonly observed. Such variation is often poorly reflected by procedure-estimated standard deviations alone. Indeed, 1 of the most challenging aspects of LAGB is the wide variation in weight loss that occurs in this population. Bessler et al. demonstrated this variation in a study on frequency distribution of weight loss after LRYGB and LAGB. This study confirmed that the distribution was different between the 2 procedures with initial weight loss being less predictable after LAGB and the 2–3 year weight loss no longer following a normal, single-peak distribution.⁵ The breadth of variation between individuals undergoing the same procedure is potentially important information, with potential clinical implications including in individual selection of the appropriate procedure and patient counseling. This variation in results after LAGB for example has led to multiple studies trying to identify predictors of success after LAGB.6, 7, 8

The comparative rates of lack of a successful treatment effect of each procedure is also an important metric and is seldom reported outside of revisional procedure publications.⁹ Multiple definitions have been published, but no clear consensus exists regarding the definition of lack of a successful treatment effect in bariatric surgery.10, 11 Furthermore, although some comparative effectiveness studies exist comparing the major procedures, few studies have focused directly and specifically on the rates of a lack of a successful treatment effect of the operations.12, 13, 14 This metric could, however, prove to be a useful tool for patient counseling.

The variability of weight loss and rates of a lack of a successful treatment effect after bariatric surgery have not, to the best of our knowledge, been measured nor compared between the 3 current, most frequently performed bariatric procedures. The aim of this study was to evaluate the heterogeneity of weight loss between LRYGB, LSG, and LAGB as well as compare their relative rates of a lack of a successful treatment effect over the first 3 years postoperatively.