Baseline Characteristics Poorly Predict Peak Weight Loss With Semaglutide in SELECT

A prespecified analysis of SELECT assessed whether routinely available baseline characteristics predict peak percentage body weight loss during semaglutide treatment. SELECT evaluated cardiovascular benefits in adults with obesity, preexisting atherosclerotic disease, and no diabetes; mean age was 62 years and 72% were men. This analysis included 7,594 participants continuously treated with semaglutide for at least one year, representing 86.3% of those randomized to the drug. Candidate predictors included demographics, medical history, laboratory variables, and clinical measurements.

Median peak weight loss was 12.5% (interquartile range [IQR] 7.9–17.9), achieved after a median of 19 months (IQR 10–30). The full model explained less than 10% of variability; female sex accounted for 85.3% of the model’s explained variation, or approximately 8% overall. Median peak loss was 16.6% in women versus 12.4% in men. Maximum semaglutide dose was also associated with weight loss: median peak loss was 6.0% with 0.24 mg versus 13.3% with 2.4 mg.

These findings indicate limited utility of baseline clinical characteristics for predicting an individual’s maximum weight-loss response. The analysis was restricted to participants who maintained treatment for at least one year, limiting applicability to those who discontinue earlier. Dose associations do not establish causality. Only the abstract was available for assessment.