Genetic variants associated with weight loss and metabolic outcomes after bariatric surgery: A systematic review

The extent to which genetic variations contribute to interindividual differences in weight loss and metabolic outcomes after bariatric surgery is unknown. Identifying genetic variants that impact surgery outcomes may contribute to clinical decision making. This review evaluates current evidence addressing the association of genetic variants with weight loss and changes in metabolic parameters after bariatric surgery. A search was conducted using Medline, Embase, Scopus, Web of Science, and Cochrane Library. Fifty‐two eligible studies were identified. Single nucleotide polymorphisms (SNPs) at ADIPOQ (rs226729, rs1501299, rs3774261, and rs17300539) showed a positive association with postoperative change in measures of glucose homeostasis and lipid profiles (n = 4), but not with weight loss after surgery (n = 6). SNPs at FTO (rs11075986, rs16952482, rs8050136, rs9939609, rs9930506, and rs16945088) (n = 10) and MC4R (rs11152213, rs476828, rs2229616, rs9947255, rs17773430, rs5282087, and rs17782313) (n = 9) were inconsistently associated with weight loss and metabolic improvement. Four studies examining the UCP2 SNP rs660339 reported associations with postsurgical weight loss. In summary, there is limited evidence supporting a role for specific genetic variants in surgical outcomes after bariatric surgery. Most studies have adopted a candidate gene approach, limiting the scope for discovery, suggesting that the absence of compelling evidence is not evidence of absence.


| INTRODUCTION
Obesity is a complex, multifactorial, chronic disease that results from the interplay between environmental and biological factors. 1 While during the past decades environmental alterations have driven the rapid increase in obesity rates worldwide, genetic factors influence individual susceptibility to these obesogenic exposures. 27][8] For people with severe obesity, bariatric surgery is the most effective treatment to achieve sustained weight loss and to improve metabolic health. 9,10However, variability in weight change and glycemic control after bariatric surgery is well recognized. 11,124][15] The extent to which specific genetic variants contribute to this interindividual variation is unclear.
Monogenic forms of obesity are often caused by mutations in the melanocortin-4 receptor (MC4R), leptin receptor (LEPR), and proopiomelanocortin (POMC) genes. 16,17Yet, in most instances, obesity is likely to have a polygenic basis, resulting from the cumulative effect of many hundreds or thousands of common variants, such as single nucleotide polymorphisms (SNPs).SNP technologies can be used to identify genes linked to a disease and to understand variation in treatment response. 18Identifying genetic variants that influence outcomes of the surgical treatment of severe obesity may be useful for clinical decision making.Genetics can also help identify novel molecular pathways involved in disease, which has proven extremely powerful for novel target discovery and drug repurposing. 19,20re, we systematically reviewed the published literature on the association of genetic variants with bariatric surgery treatment outcomes.This review was conducted as part of the preparatory phase for an international bariatric surgery consortium (Genetics of Bariatric surgery outcomes [GENBAR]) on the identification of genetic variants associated with clinical outcome phenotypes after bariatric surgery.

| Search strategy
This systematic review was registered in PROSPERO (ID: CRD42021259912) and performed using the PRISMA guidelines. 21 information specialist and two authors (R.M. and V.M.) composed the search strategy (Table S1).The databases of Medline, Embase, Scopus, Web of Science, and Cochrane Library were systematically searched for eligible articles up to July 1, 2022.Titles and abstracts of all retrieved articles were screened independently by two reviewers for eligibility (R.M. and S.M) using Rayyan. 22Full texts of the remaining studies were assessed by the same reviewers.Any inconsistencies were resolved by group discussion (R.M., S.M., and V.M.).Studies were deemed eligible if they conformed to the predetermined inclusion and exclusion criteria.
2. Predictors were genetic variants (SNPs).Exclusion criteria were studies in animals, studies reporting on genetic mutations, and studies in which the type of surgery was unclear and/or postoperative weight/metabolic data were unavailable.The selected language of the articles was English.When articles reported duplicate data, the study reporting the highest number of participants or the longest follow-up was included.Studies from the same group were included when they reported different genetic variants.

| Quality of evidence assessment
Risk of bias in individual studies was assessed using the validated Q-genie tool. 23

| Linkage disequilibrium
Linkage disequilibrium (LD) describes if an allele of one SNP is inherited or correlated with an allele of a neighboring polymorphisms within a population.Assessment of LD is needed to identify genetic markers that tag the actual causal variant to complex traits.The online LD pop tool was used to assess correlations between different SNPs of a particular gene. 24| RESULTS

| Search results
We identified 3697 titles during the initial systematic search (Figure 1).Four additional studies were added after checking reference lists of relevant studies.After removing duplicates (n = 1208), 2493 titles and abstracts remained for screening.A total of 74 full-text articles were assessed for eligibility; 52 articles met the inclusion criteria.Risk of bias was assessed for all studies using the Q-genie assessment tool, 18 studies were rated as poor quality, 33 studies were rated as moderate quality, and one study was rated as good quality (Table S2).

| Demographics
An overview of general characteristics of the included studies is displayed in Table 1.The sample size varied from 21 to 1771

| Fat mass and obesity-associated gene (FTO)
The most frequently studied genetic variants were those representing the signal in FTO and represent the first robust association between a common genetic variant and BMI.This variant, represented in eight (of 10 studies) by the SNP rs9939609, is variably associated with weight loss (Table 2).For example, in four studies, no statistically significant association was found with weight loss. 45,63,67,73Two studies found a positive association with short-term weight loss 3-6 months post-surgery, while longer-term weight loss showed no association with the presence of this variant in patients undergoing a BPD. 29,50ndstein et al found a positive association with %EBMIL (AA 86.1 ± 2.3%, TA/AT 83.0 ± 1.7%, and TT 81.5 ± 2.3%, p = 0.02) 2 years after RYGB. 35In contrast, Rodrigues et al noted that the rs9939609 variant was associated with less weight loss and greater and earlier weight regain. 72Representing the same signal, the rs993056 variant was associated with significantly higher BMI loss (%) 6 months after SG. 28 Sarzynski et al found that the rs16945088 variant, which is weakly correlated with the rs9939609 and rs9930506 variants, was associated with improved weight loss after bariatric surgery in patients undergoing LAGB, but not RYGB. 41

| Melanocortin-4 receptor gene (MC4R)
Of the nine studies evaluating common variants of the MC4R gene, five showed no statistically significant association between genetic variants and postoperative weight loss (Table 2). 30,34,42,43,73Two studies reported SNPs associated with less weight loss at 1-and 3-year follow-up after RYGB and LAGB. 32,71In contrast, the rs17773430 and rs5282087, which are highly correlated, had a positive association with weight loss.Javanrouh et al showed that the rs17773430 variant was related to greater weight loss 12 months after surgery. 62Mirshahi et al concluded that patients with the rs5282087 variant are predisposed to lower weight at nadir, the lowest recorded weight (%EWL 65 ± 0.5% vs. 72 ± 0.1% and 73 ± 0.6% of baseline weight), after RYGB surgery. 66

| Other genes
Various other genes were examined in three or fewer studies (Table 2).van der MEER ET AL.

| Genes associated with weight results
Thirteen different genes were examined and showed a statistically significant association between polymorphisms and weight loss after bariatric surgery.Patients homozygous for the rs490683 C/C genotype of the growth hormone secretagogue receptor gene (GHSR) experienced most weight loss at 30 months, while the C/C genotype at rs9819506 lost slightly less weight. 65Less weight loss was also found for the rs1360780 variant of the FK506-binding protein 5 gene (FKBP5) and the rs1126535 variant of the CD40 ligand gene (CD40L). 31,44A positive relation with surgical weight loss (up to 9.9% more TWL and up to 5.1 kg/m 2 higher decrease in BMI after 1 year) was reported for genetic variants of ghrelin (GHRL), proopiomelanocortin (POMC), interleukin-6 (IL-6), estrogen receptor-1 (ESR1), lysophospholipase-like 1 (LYPLAL1), mitochondrial translational initiation factor 3 (MTIF3), ST8 sialyltransferase 2 (ST8SIA2), glucagon-like peptide 1 receptor (GLP1R), fatty acid amid hydrolase (FAAH) gene, and Acyl-CoA synthetase long-chain 5 (ACSL5). 25,29,33,38,41,44,49,56,59,7310.2| Genes with no statistically significant association with weight loss For numerous genes, only one or more SNPs were examined, and no significant association was found between the genetic variant and weight loss after bariatric surgery.33,36,

.3 | Metabolic outcomes
There were 25 studies reporting on metabolic outcomes after bariatric surgery (Table 2).Follow-up ranged from 6 months to 5 years.

| Adiponectin gene (ADIPOQ)
Four studies found a positive association between SNPs at the ADI-POQ locus and changes in measures of glucose homeostasis after bariatric surgery. 51,52,69,75The rs266729 variant was associated with a weight-independent decrease of fasting insulin levels (Δ À9.1 ± 3.1 vs. À4.1 ± 2.3 mUI/L) and HOMA IR (Δ À1.3 ± 0.2 vs. À0.8 ± 0.3 units) up to 3 years after BPD. 50Similar results (nonadjusted for weight loss) were seen for the rs1501299 variant and rs3774261 variant, which are likely to represent the same signal. 51,66l patients showed an improvement in lipid profile, and for the rs17300539 and rs266729 variants, the decrease in total cholesterol, LDL cholesterol, and triglycerides was greater in the C/C genotype. 51,75For the latter variant, the improvement was independent of weight loss.In addition, the A-C haplotype, a unique combination of the rs266729 and rs17300539 variant, showed a greater reduction in LDL concentrations after RYGB (À43.0%A-C carriers vs. À18.1% G-G carriers, p = 0.019).

| Other genes
For polymorphisms of 17 genes, the association between SNPs with postoperative metabolic improvement was only examined in one study.Neuropeptide Y (NPY) (rs16147), brain-derived neurotrophic factor (BDNF) (rs10767664), apolipoprotein A1 (APOA1) (rs670), and cannabinoid receptor 2 (CB2R) (rs3123554) variants were associated with increased improvement in glucose homeostasis including fasting glucose, fasting insulin, and HOMA IR after BPD. 53,54,57,68Human catalase gene promotor haplotype (CAT) (rs769214) was associated with lower CRP and diastolic blood pressure levels 12 months after RYGB, LAGB, and SG. 74Higher decreases in plasma triglycerides were seen in T-carriers of the ACSL5 rs2419261 variant. 595,76 This review describes the relationship between genetic variants and clinical outcomes after bariatric surgery.We found a positive association between polymorphisms of the ADIPOQ gene and postoperative changes in measures of glucose homeostasis and lipid profiles, but not with weight loss after surgery.Furthermore, SNPs of well-known obesity-related genes (FTO and MC4R) demonstrated contradictory results regarding their effect on bariatric surgery-induced weight loss and metabolic improvement.In T-allele carriers of the UCP2 variant rs660339, greater postsurgical weight loss was found.There were only a few overlapping SNPs associated with both weight loss and improvement of metabolic health suggesting that some SNPs affect metabolic outcomes independent of weight loss.These results show that the effect of specific genetic variants on weight loss and metabolic outcomes after bariatric surgery remains largely unknown.
In the present study, we identified a positive association between the ADIPOQ gene and metabolic improvement, while weight loss was similar for all genotypes.None of the included studies (n = 6) found a direct association between weight loss and the ADIPOQ gene.The ADIPOQ gene is expressed in adipose tissue exclusively.The encoded protein, adiponectin, suppresses hepatic glucose production, enhances insulin sensitization, and improves fatty acid oxidation in skeletal muscle. 779][80][81][82] SNPs at ADIPOQ (rs226729, rs1501299, rs3774261, and rs17300539) were associated with improvement of insulin sensitivity (lower fasting insulin and HOMA-IR) after bariatric surgery.Furthermore, the C/C genotype (rs266729 and rs17300539) was associated with greater improvement in lipid profiles after surgery.However, some degree of the improvement of insulin sensitivity after bariatric surgery in subjects with these variants is probably caused by the amount of weight loss itself. 83 this systematic review, we identified association signals for SNPs in 41 different genes; however, a large part of the studies focused on genetic variants of the FTO (n = 10) and MC4R (n = 9) genes.In 2007, FTO was discovered as the first GWAS-identified obesity-susceptibility gene. 84FTO seems to play a potential role in the regulation of energy metabolism; however, the molecular mechanisms are not well known. 85The hypothalamic MC4R pathway is involved in controlling food intake and energy expenditure.Loss-of-function mutations of MC4R are a well-known cause for monogenetic obesity. 86Furthermore, previous research revealed that MC4R polymorphisms are associated with common obesity as well. 87,88Overall, we found that SNPs of FTO and MC4R genes showed inconsistent results in their relationship with weight loss and metabolic improvement after bariatric surgery.This might be explained by the fact that in humans, weight loss is affected by multiple polymorphisms. 89The interindividual variability in surgery outcomes may not be explained by one genetic variant alone, and for that reason, the contribution of FTO and MC4R in polygenic risk scores could be more appropriate for predicting outcomes after bariatric surgery.This is in line with the systematic review of Gupta et al that reported on different genetic risk scores that predicted weight loss or metabolic outcomes 12-96 months after bariatric surgery. 90These promising findings underscore the need of large and representative prospective studies investigating the validation and clinical utility of polygenic risk scores in clinical decision making.Moreover, genetic studies for obesity-related traits identified a collection of different variants for the FTO and MC4R gene. 91,92ry limited studies have been performed on the myriad of variants of these two genes.
Uncoupled proteins (UCPs) are anion carriers present in the inner mitochondrial membrane. 93UCP2 is widely expressed in tissues throughout the body, including skeletal muscle, white adipose tissue, and pancreatic islets.Recent studies have shown that UCP2 gene polymorphisms are involved in regulation of energy expenditure and have a role in the pathogenesis of metabolic disorders, such as obesity and diabetes. 94,95Even though there were only a few studies regarding the rs660339 variant of the UCP2 gene, we found a positive association with weight loss after surgery.Interestingly, this SNP causes a conservative amino acid change (position 55 of exon 4), and, to date, there is no evidence that this alteration leads to a functional change in the protein.The rs660339 variant is tightly linked with the À886G/A polymorphism (rs659366), which has an evident effect on UCP2 gene tissue expression. 96Therefore, the rs660339 variant might not be responsible for the observed association with weight loss, but instead could be reflecting the effects of other functional polymorphisms.In addition, results from a recent published systematic review suggest that UCP2 expression in adipose tissue and skeletal muscle seems to be affected by metabolic changes after bariatric surgery. 97The upregulation of UCP2 expression after RYGB might lead to greater energy expenditure and weight loss.Future research exploring the predictive value of UCP polymorphisms and the effect of (changes in) gene expression on surgery outcomes may reveal novel targets for obesity therapy.
Our review shows that there is very limited evidence supporting a role of specific genetic variants and surgical outcomes after bariatric surgery.The absence of compelling evidence might be explained by the lack of well-designed and adequately powered studies.The candidate gene approach has been widely used to identify genes linked to a disease and to understand variation in treatment response. 98,99It focuses on assessing the effect of SNPs within pre-specified genes that were chosen based on current knowledge of its biology and pathophysiology.Candidate gene studies are often limited by small sample sizes and the lack of statistical power. 100Additional criticism on this approach is that most genetic associations are not replicated in subsequent cohorts.Given poor reproducibility rates, the credibility of most of the genetic associations has been questioned. 101,102In recent years, genetic technology has undergone a major paradigm shift.
Genome-wide association studies (GWAS) have increased the speed of discovery and found novel genetic associations that were not anticipated by the candidate gene approach. 103,104GWAS is a genotyping technology that screens hundreds of thousands of SNPs across the entire genome for associations between genetic variants and complex clinical conditions. 105Highly reliable results have emerged from GWAS, and as a result, GWAS is preferred over candidate gene approach. 106th only two discovery studies (GWAS) included, most of the studies in our review are candidate gene studies.Meta-analysis of the published literature could be an important tool to increase validity van der MEER ET AL.
by increasing sample sizes and assessing the cumulative evidence for association between genetic variants and clinical outcomes after bariatric surgery.However, the lack of a standardized definition for biological endpoints after bariatric surgery impedes comparability.Future genetic studies will benefit from consistent definitions of clinical outcome phenotypes.Moreover, international collaboration is needed to provide sufficient power to perform a GWAS and to replicate findings in independent studies across sexes and ethnicities.Therefore, the GENBAR consortium aims to bring together multiple bariatric surgery cohorts that include genetic information (GWAS) and clinical outcome phenotypes after bariatric surgery.

| Limitations
There were several limitations for this review that should be considered, such as differences in the design of the studies including sample size, sample collection, length of follow-up, gene selection, and genotyping.In addition, the overall quality of the studies was low with substantial inter-study heterogeneity in terms of quality control of samples, patient selection, and adjustment for confounding variables.Most of the studies were small, and the sample size estimation or power analysis was lacking.Furthermore, studies lost points due to poorly described nontechnical classification of the genetic variant and inadequate disclosure of potential sources of bias.For example, there is an increased risk of publication bias since statistical tests for multiple genetic variants are assessed with reporting of mainly positive results.Therefore, we rated the majority of the studies as moderate quality and a large part as poor quality, in contrast with the previous published review on genetic determinants of weight loss. 90Lastly, there was a multitude of different types of surgery included in this review.Associations between surgery outcomes and genetic variants examined in cohorts with less commonly performed procedures, such as LAGB, OAGB, VGB, and BPD, may not translate to SG and RYGB.

| CONCLUSION
This systematic review provides an overview of genetic variants associated with weight loss and metabolic outcomes after bariatric surgery.We showed that polymorphisms of the ADIPOQ gene are associated with increased improvement in measures of glucose homeostasis and circulating lipid profiles.Moreover, this review suggests that the UCP gene is associated with greater weight loss after surgery.The use of the candidate gene approach in most studies and inconsistency of the results make it difficult to identify genetic variants that can explain interindividual variation after bariatric surgery.
Future research in this field is needed and will benefit from international collaboration to use GWAS, which is superior to candidate gene approach, to identify genetic variants that can help predict surgery outcomes and elucidate pathophysiological mechanisms of obesity for precision medicine and drug target identification.
participants.Most of the studies based their inclusion criteria for surgery on the IFSO guidelines, including BMI > 40 kg/m 2 or BMI 35 kg/ m 2 in the presence of obesity-related medical problems.The mean preoperative BMI ranged from 41.0 to 51.0 kg/m 2 , the mean age ranged from 31.0 to 48.9 years, and the proportion of percentage of female participants was between 61.8% and 100%.

FABP2:
Regulation of energy homeostasis, melanocyte stimulation, and immune modulation rs1042571 Vélazquez-Fernández et al 73 " year Various additional variants Potoczna et al 32 = years IL-6: Interleukin 6 Role in inflammation and maturation of B receptor-1 Regulation of transcription of estrogen-inducible genes that play a role in growth, metabolism, sexual development, -like 1 Enzymatic function is unclear but linked to fat distribution, waist-to-hip ratio, and nonalcoholic fatty liver disease rs4846567 Bandstein et al 25 " years MTIF3: Mitochondrial translational initiation factor 3 Role in mitochondrial protein synthesis rs4771122 Rasmussen-Torvik et al 41 " 1-9 years ST8SIA2: ST8 sialyltransferase 2 Contribution to a variety of neuronal events by producing polysialic acid rs17702901 Hatoum et al 38 " 1-7 years GLP1R: Glucagon-like peptide 1 receptor Stimulates glucose-induced insulin secretion rs6923761 de Luis et al 56 " = months FAAH: Fatty acid amide hydrolase Hydrolyzes endogenous bioactive fatty acid amides, including anandamide and oleamide (both playing a role in food intake) C358A de Luis et al 49 " = year ACSL5: Acyl-CoA synthetase long-chain 5 Regulation of fatty acid metabolism and transportation of fatty acids into cells rs2419621 de Luis et al 59 " " year CD40L: CD40 ligand Regulation of B cell function by engaging CD40 on the B cell surface rs1126535 Vitolo et al 44 # year BDNF: Brain-derived neurotrophic factor Assists the survival of existing neurons and encourages the growth and differentiation of new Fatty acid binding protein 2 Uptake, intracellular metabolism, and transport of long-chain fatty acids rs1799883 de Luis et al 58 Kops et al 63 = = = year years NR3C1: Nuclear receptor subfamiliy 3 group c member 1 Regulation of the hypothalamic-pituitary-adrenal axis by modulating availability of the stress A1 Central role in lipoprotein metabolism.Component of high-density lipoprotein (HDL) -gamma-inducible lysosomal thiol reductase Can reduce protein disulfide bonds and has an role in MHC class IIbinding protein 3 Central role in lipoprotein metabolism.Main apolipoprotein of chylomicrons and lowdensity lipoproteins adaptor protein Mediates activation of various kinases involved in multiple signaling pathways, including the receptors of insulin, insulinlike growth factor and brain-derived neurotrophic factor rs7498665 Novais et al 67 = 1 year TAS1R2: Taste 1 receptor member 1 Contributes to sweet taste receptor activity rs35874116 Novais et al 67 = Decomposes hydrogen peroxide generated during biological oxidation to water and oxygen rs769214 Alili et al 74 = 1 year CETP: Cholesteryl ester transfer protein Involved in the transfer of cholesteryl ester from high-density lipoprotein (HDL) to other Suppression of insulin ability to stimulate glucose uptake into adipose cells rs1862513 Gasparotto et al 75 = 39 months Note: #: poorer postoperative outcomes, ": better postoperative outcomes, =: no association.
This tool is developed for genetic research and consists of 11 questions with a 7-point Likert scale ranging from 1 (poor) to 7 (excellent).Items included are study rationale, outcome, comparability, exposure, bias, sample size, analysis, statistical methods and control for confounding, inferences for genetics analysis, and inferences drawn from results.For studies with control groups, total scores of ≤35 indicate poor quality studies, >35 and ≤45 indicate studies of moderate quality, and >45 indicate good quality studies.For studies without control groups, total scores of ≤32 indicate poor quality studies, >32 and ≤40 indicate studies of moderate quality, and >40 indicate good quality studies.
Study demographics.