Weight regain after bariatric surgery: a systematic literature review and comparison across studies using a large reference sample

Published estimates of weight regain (WR) after bariatric surgery vary greatly. Understanding the sources of variability in the literature and clarifying the magnitude of WR after Roux-en-Y gastric bypass (RYGB) and sleeve gastrectomy (SG) are critical for informing expectations and planning interventions. A literature search through January 2019 yielded 15 English-language studies that reported WR in at least 30 participants, not selected based on weight loss or WR, at least 3 years after primary RYGB (n = 11) or SG (n = 5). Median follow-up was 5.0 (range, 3.2-10.0) years. Median sample size was 62 (range, 33-464). Samples represented a median of 54.3% (range, 10.7%-100%) of eligible participants. Nadir weight was determined by serial research assessments (n = 1), medical records (n = 7), participant recall (n = 4), or an undisclosed method (n = 4). Three continuous and 8 binary WR measures (the latter, based on various thresholds for clinically meaningful WR) were reported. To enable comparison across studies, the percentage difference in WR in each study versus a reference sample (n = 1433 RYGB), matched on time since surgery and WR measure, was calculated. Median WR in the reference sample increased from 8.2 (25th-75th percentile: 0-19.5) to 23.8 (25th-75th percentile: 9.0-33.9) percent of maximum weight lost, 3 to 6 years post RYGB surgery. Studies of RYGB versus SG, with larger versus smaller samples, with higher versus lower participation rates, that determined nadir weight via participant recall versus medical records, and reported continuous versus binary WR measures tended to have WR values closer to the reference sample and each other. Variation in WR estimates was explained by heterogeneity in WR measures, timing of assessment, surgical procedure, and study design characteristics. The best estimate of WR after RYGB likely comes from the large reference sample. WR after SG versus RYGB appears higher. However, additional high-quality studies with uniform reporting of WR by surgical procedure are needed.