Baseline anthropometric and body composition predictors of weight loss response to orlistat in individuals with obesity

Aims: Obesity is a major public health problem that is increasing globally. The ability to predict factors associated with weight loss in the treatment of obesity may contribute to the development of individualized therapeutic approaches. The aim of this study was to investigate parameters related to weight loss response in obese individuals receiving orlistat treatment. Methods: This single-center, retrospective study evaluated the data of 161 patients who were initiated on orlistat therapy due to obesity. Patients were categorized into three groups based on weight loss achieved at 6 months: non-responders (<5% weight loss), moderate responders (5-9.9% weight loss), and responders (≥10% weight loss). In the primary analysis, multiple linear regression with a prespecified enter model was used to identify baseline predictors of percentage weight loss as a continuous outcome. Secondary analyses employed binary logistic regression with enter models to identify predictors of achieving ≥5% and ≥10% weight loss, followed by ROC curve analysis of model-derived predicted probabilities. Exploratory analyses included principal component analysis and factor-based logistic regression. Results: In the primary analysis, baseline muscle mass (MM) was the only independent predictor of percentage weight loss (B=0.585, p<0.001), with the model explaining 18.5% of the total variance. In secondary analyses, waist circumference was a significant negative predictor of achieving both ≥5% (OR=0.897, p=0.009) and ≥10% weight loss (OR=0.868, p=0.018), while MM additionally predicted achievement of ≥10% weight loss (OR=1.857, p=0.037). For the ≥10% threshold, waist circumference and MM were significant in the primary logistic model (OR=0.868 and OR=1.857, respectively); however, bootstrap resampling produced wider bias-corrected and accelerated 95% confidence intervals that included unity for both, and these associations are therefore interpreted as exploratory. ROC curve analysis demonstrated good discriminative ability for both models, with AUC values of 0.749 (95% CI: 0.639-0.860) and 0.765 (95% CI: 0.654-0.877) for the ≥5% and ≥10% weight loss models, respectively. Exploratory factor analysis identified three components reflecting lipid-associated metabolic risk, fat mass and visceral adiposity, and MM with hepatic steatosis, all of which significantly predicted achieving ≥10% weight loss. Conclusion: Baseline MM and waist circumference are independently associated with weight loss response in obese individuals receiving orlistat therapy. Weight loss was more pronounced in individuals characterized by higher MM and lower waist circumference at baseline. However, further long-term comparative randomized controlled studies are needed to determine whether these associations are entirely specific to orlistat treatment and to assess their long-term sustainability in comparison with lifestyle interventions.