Heritability of the Weight Loss Response to Gastric Bypass Surgery

Patients and Methods

Results

Preoperatively, patients in this cohort had an average BMI of 50.2 ± 8.6 kg/m², an average age of 44.7 ± 11.3 yr, and were 74.8% female and 86% Caucasian. These characteristics were similar among the different groups studied (Supplemental Table 1, published on The Endocrine Society's Journals Online web site at http://jcem.endojournals.org). After RYGB, patients lost an average of 119.2 ± 41.7 pounds at weight nadir, corresponding to an EBWL of 79.7 ± 21.8%; the population pattern of percent excess weight loss follows the wide and normal distribution observed previously (Fig. 1) (10). To determine whether there is a genetic component to the variation in weight loss after surgery, we compared weight loss after RYGB within pairs of genetically related and genetically unrelated individuals. Unrelated individuals demonstrated far less similar weight loss after surgery than first-degree relatives (Fig. 2), with an average difference in EBWL of 25.4% in unrelated individuals and 9.9% in first-degree relatives (P = 0.001). Because the observed similarity in weight loss could result primarily from shared environmental influences, we compared weight loss within pairs of individuals who were living together but who are genetically unrelated. Pairs of these environmentally related controls had a shared response similar to completely unrelated individuals (mean 26.1% difference in EBWL; P = 0.60), a response that was substantially different from that of first-degree relatives (P = 0.005). The small number of second- and third-degree relative pairs in this cohort precludes statistical analysis of these groups.

The ICC represents the portion of total variation in outcome explained by the pair relationship. Using mixed-effects models adjusted for age, sex, year of surgery, and preoperative BMI, the ICC were 70.4% for first-degree relatives (P = 0.02), 14.3% for environmentally related controls (P = 0.67), and 0.9% for randomly paired individuals (P = 0.48). Because preoperative BMI is strongly associated with postoperative percent EBWL (10), we additionally matched the unrelated controls based on 5-kg/m²-wide BMI groups to mimic the distribution of differences in preoperative BMI between first-degree relatives. After this adjustment, first-degree relative pairs still demonstrated significantly less difference in weight loss compared with the unrelated controls (difference in EBWL for unrelated pairs 22.3 ± 17.9%; P = 0.01 vs. first-degree relatives). Thus, first-degree relatives have weight loss outcomes after surgery that closely and significantly resemble each other, a characteristic not shared by environmentally related controls or randomly paired individuals. Similar results were seen when men and women were examined separately (data not shown).

Discussion

These results suggest that there are strong genetic determinants of weight loss after RYGB. Despite the small number of genetically related individuals in this cohort, there is demonstrably tight clustering of weight loss outcomes within pairs of first-degree relatives. Although some contribution of environmental effects cannot be excluded, this tight clustering is not seen with cohabitating individuals, indicating that there is a substantial genetic impact beyond any potential shared environmental influences. These observations underscore the biological nature of the response to RYGB and suggest that variation in the genetic background of individuals strongly influences weight loss after this operation.

The observed genetic contribution to weight loss after RYGB is consistent with a recent small case study of weight loss after bariatric surgery in four pairs of monozygotic twins. The three twin pairs who underwent RYGB exhibited nearly identical within-pair weight loss outcomes after surgery. Interestingly, the twins who underwent adjustable gastric banding had widely dissimilar weight loss outcomes (13). Although the size of our cohort limits the power to identify the specific genetic signals associated with weight loss by genome-wide association study, we tested the limited number of single-nucleotide polymorphisms known to be associated with the development of obesity. There was no association between any of these 32 genetic loci and weight loss after this procedure (data not shown), a finding that has also been observed in other cohorts (14, 15). A recent report by Gerhard and colleagues (16) suggests that high allelic burden from multiple obesity-associated polymorphisms may be associated with less weight loss after RYGB in surgical patients with relatively low preoperative BMI (<50 kg/m²). It is therefore likely that large-scale genome-wide association studies using one or more large cohorts followed by extensive replication will be required to identify the relevant genetic contributors. This unsupervised approach can yield insight into previously unidentified or unanticipated physiological mechanisms of weight loss after surgery, which appear to be different from the mechanisms that lead to severe obesity in the first place.

The specific biological mechanisms of weight loss after bariatric surgery are poorly understood. Traditional thinking was that weight loss after RYGB resulted from mechanical restriction of food intake and/or calorie malabsorption. However, studies in both rodents and humans demonstrate that macronutrients are not malabsorbed after this operation (17). Moreover, under conditions of caloric restriction, the body responds by decreasing energy expenditure (18), but recent animal and human evidence indicates that this metabolic adaptation is blunted or blocked after RYGB despite a dramatic decrease in food consumption; indeed, both rats and mice exhibit increased energy expenditure after RYGB (6). Studies comparing RYGB to other gastrointestinal weight loss procedures that only decrease the size of the stomach to one comparable to RYGB (e.g. adjustable gastric banding) show that RYGB confers much greater weight loss (7). Thus, the mechanisms of weight loss after RYGB appear far more complex than mere restriction or malabsorption alone. The findings in this study support a physiological mechanism of response to this operation and suggest that biological determinants account for the wide variation in weight loss outcomes among patients.

In conclusion, these data demonstrate that genetic factors explain a significant portion of the variation in weight loss after RYGB. This observation provides additional support for a physiological mechanism of response to surgery. Discovery of the specific genes that predict this response will provide important clues to the molecular mechanisms that account for the high efficacy and long durability of response to RYGB and will advance our understanding of the biological mechanisms of weight loss after this procedure. Moreover, understanding of the profound response to RYGB should illuminate key features of the normal regulation of energy balance and body weight independent of surgery. Identifying these mechanisms will facilitate the development of more effective nonsurgical therapies for obesity that exploit these surgical mechanisms. Determination of the predictive polymorphisms could also lead to the development of tools to distinguish those patients who are likely to benefit most from RYGB. Use of such predictive tools could aid in the optimal selection of patients for RYGB and further improve the risk-to-benefit profile for this highly effective yet invasive treatment.

Supplementary Material