MATERIALS AND METHODS: Searches of electronic databases (PubMed, Embase, CINAHL, Cochrane) were undertaken for randomized controlled trials describing weight loss outcomes in adults at 5 years postoperatively. Where sufficient data was available to undertake meta-analysis, the Hartung-Knapp-Sidik-Jonkman estimation method for random effects model was utilized. The review was registered with PROSPERO and reported following in accordance with Preferred Reporting Items for Systematic Reviews and Meta-Analyses.
RESULTS: Five studies met the inclusion criteria totaling 1028 patients (LVSG=520, LRYGB=508). Moderate but comparable levels of bias were observed within studies. Statistically significant body mass index loss ranged from -11.37 kg/m (range: -6.3 to -15.7 kg/m) in the LVSG group and -12.6 kg/m (range: -9.5 to -15.4 kg/m) for LRYGB at 5 years (P<0.001). Systematic review suggested that LRYGB produced a greater weight loss expressed as percent excess weight and percent excess body mass index loss than LVSG: this was not corroborated in the meta-analysis.
CONCLUSIONS: Five year weight loss outcomes suggest both LRYGB and LVSG are effective in achieving significant weight loss at 5 years postoperatively, however, differences in reporting parameters limit the ability to reliably compare the outcomes using statistical methods. Furthermore, results may be impacted by large dropout rates and per protocol analysis of the 2 largest included studies. Further long-term studies are required to contradict or validate the results of this meta-analysis.
METHODS: Electronic databases (Pubmed, EMBASE, CINAHL) were searched for RCTs conducted in adults (>18y) that compared the 5-year- outcomes of LVSG to LRYGB and described comorbidity outcomes were included. Where data allowed, effect sizes were calculated using the Hartung-Knapp-Sidik-Jonkman estimation method for random effects model. Presence of bias was assessed with Cochrane Risk of Bias 2.0 and funnel plots, and certainty of evidence evaluated by GRADE. The study prospectively registered with PROSPERO (CRD42018112054).
RESULTS: Three RCTs (LVSG=254, LRYGB=255) met inclusion criteria and reported on chronic disease outcomes. Improvement and/or resolution of hypertension favoured LRYGB (odds ratio 0.49, 95% CI 0.29, 0.84; P =0.03). Trends favoring LRYGB were seen for type 2 diabetes and dysplidemia, and LVSG for sleep apnea and back/joint conditions ( P >0.05). The certainty of evidence associated with each assessed outcome ranged from low to very low, in the setting of 'some' to 'high' bias assessed as being present.
CONCLUSION: Both LRYGB and LVSG are effective in providing long-term improvements in commonly experienced obesity-related comorbidities; however, the limited certainty of the evidence does not allow for strong clinical conclusions to be made at this time regarding benefit of one procedure over the other.