Weight regain after discontinuation of semaglutide and tirzepatide: development and validation of a risk-prediction model
Calculator | Citation | For research use only.
About the GLP-1 Weight-Regain Risk Calculator
This model was developed in a Mayo Clinic cohort of 21,882 adults who discontinued semaglutide and 14,513 who discontinued tirzepatide. Drug-specific models use six routinely available variables — with an optional refinement from early weight-loss pace when a 3-month weight is entered — to estimate the timing of clinically meaningful weight regain after stopping treatment.
Outcome. Time to regain ≥50% of the total body weight lost on treatment (TR50), measured from the discontinuation date.
Predictors. Weight lost on treatment (%TBWL), treatment duration, age, sex, baseline BMI, and diabetes — the six routinely available variables. Early weight-loss pace (deceleration) is an optional refinement applied when a 3-month weight is supplied.
Performance. Discrimination (C-index) 0·77 (semaglutide) / 0·82 (tirzepatide); cohort median TR50 6 months for both. In an external cohort, the semaglutide model preserved its risk ordering. Entering a 3-month weight applies an independently estimated early-pace (deceleration) effect as an offset on top of this published model: the six core effects, baseline curve, and quartile cut-points are unchanged; an average-pace weight leaves TR50 unchanged, faster early loss shifts it later and a slow start earlier. Deceleration is a significant, independent association in the study.
How to read the result
The risk level places the patient in one of four equal-sized groups (quartiles) of the development cohort, ranked by predicted risk of early regain: low, intermediate-low, intermediate-high, and high, from slowest to fastest regain. The time shown is the predicted time to regain at least half of the weight lost (measured from the lowest weight reached during treatment), measured from the day treatment stops; the weight and BMI shown are what regaining half of that loss corresponds to for this patient. The time is a typical estimate for patients with similar characteristics, rounded to the nearest month. A shorter time means earlier weight regain; a longer time means the loss tends to be maintained for longer.
Disclaimer
This tool is intended for research use by clinicians and researchers only and must not be used for clinical decision-making. It does not provide medical advice, treatment recommendations, or clinical diagnoses, and must not be used as a substitute for professional clinical judgement. It does not establish causal relationships between input variables and outcomes. Outputs are estimates based on population-level data and may not accurately reflect individual risk. Simplified descriptions of input variables are provided for usability.
Research underlying the model
Discontinuation of GLP-1 receptor agonists is common, and weight regain after stopping is frequent but its timing varies widely. In a Mayo Clinic cohort of 21,882 adults who discontinued semaglutide and 14,513 who discontinued tirzepatide, we developed drug-specific models to predict the time to regain ≥50% of the weight lost (TR50) from six routinely available variables — weight lost on treatment, treatment duration, age, sex, baseline BMI, and diabetes — with an optional refinement from early weight-loss pace when a 3-month weight is available. The models discriminated the timing of regain well (held-out C-index 0·77 for semaglutide and 0·82 for tirzepatide), and the semaglutide model preserved its risk ordering in an external cohort.
Citation. Weight regain after discontinuation of semaglutide and tirzepatide: development and validation of a risk-prediction model. Romanos M et al. Journal under review, 2026 · ObesityWeek oral presentation, Nov 15, 2026