Standard research · Standard sources

Evidence for GLP-1 medications and long-term maintenance

Completed July 22, 2026

Evidence for GLP-1 Medications and Long-Term Maintenance

Synthesis Date: July 22, 2026

Executive Summary

GLP-1 receptor agonists (semaglutide, liraglutide, tirzepatide) are effective for long-term weight maintenance only with continuous use. Discontinuation leads to substantial weight regain (60–75% within 12–18 months) and loss of cardiometabolic benefits. Semaglutide 2.4 mg reduces major cardiovascular events by 20% in patients with established CVD and overweight/obesity without diabetes over ~4 years. Real-world discontinuation rates approach 65% at 1 year, creating a large efficacy-effectiveness gap. Optimal maintenance dosing, class-effect generalizability, and long-term safety beyond 4 years remain uncertain.

Key Findings

1. Weight Maintenance Requires Continuous Therapy

  • STEP 4 (semaglutide, JAMA 2021): After 20-week run-in, continued semaglutide lost additional 7.9% vs. placebo regained 6.9% at week 68 (net difference 14.8 percentage points, p<0.001). Two-thirds of placebo arm regained ≥50% of run-in loss.
  • SURMOUNT-4 (tirzepatide, JAMA 2024): 36-week lead-in produced 20.9% mean loss. Continued drug lost additional 5.5% vs. placebo regained 14.0% at week 88 (net difference 19.4 percentage points).
  • STEP 1 extension (2022): One year off drug after 68 weeks of semaglutide: two-thirds of lost weight regained.
  • Meta-analyses: 60–75% weight regained within 12–18 months of stopping across agents (Lancet eClinicalMedicine 2026, BMJ 2026, Obesity Reviews 2025).

2. Cardiovascular Benefit

  • SELECT trial (NEJM 2023, n=17,604): Semaglutide 2.4 mg over mean 39.8 months reduced MACE by 20% (HR 0.80, p<0.001) in patients with CVD and BMI ≥27 without diabetes. Weight loss sustained at 10.2% throughout.
  • Class effect uncertain: Tirzepatide CVOT (SURMOUNT-MMO) is ongoing; oral semaglutide at lower doses shows attenuated effect. Expert disagreement persists (JMCP 2026, Frontiers 2026).

3. Safety Profile

  • GI side effects: Most common; 10–16% discontinuation due to adverse events in trials (SELECT: 16.6% semaglutide vs 8.2% placebo). Nausea, vomiting, diarrhea peak during titration.
  • Pancreatitis: Large meta-analyses and multi-year data do not confirm early concern; FDA label retains warning. Cleveland Clinic (2025) suggests denying to patients with pancreatitis history "unwarranted."
  • Thyroid C-cell tumors: FDA black box warning remains for MTC/MEN2. Human data from Mayo Clinic and multi-database cohorts suggest detection bias rather than causation; not definitive.
  • Gallbladder disorders: Small but significant increase (2.8% vs 2.3% in SELECT).
  • Aspiration risk: Documented during anesthesia; FDA label updated.

4. Population Differences

  • Sex: Women lose ~11% vs men ~7% of starting weight (JAMA IM 2026, 64 RCTs).
  • Age, race/ethnicity, baseline BMI/A1c: No significant differences in efficacy (same meta-analysis).
  • Older adults ≥65: Higher AE-related discontinuation; efficacy preserved.
  • Adolescents 6–17: Effective for BMI reduction; GI AEs more common; no signal for suicidality (JAMA Pediatrics 2025, 18 RCTs).

5. Real-World Effectiveness Gap

  • Discontinuation rates: 46.5% at 1 year in T2D patients; 64.8% in obesity-only patients (JAMA Network Open 2024/2025, n=125,474). Only ~1 in 3 obesity patients persistent at 1 year; ~27% adherent (PDC≥80%).
  • Efficacy vs effectiveness: Trial maintenance effects (15–19 percentage point difference vs placebo) likely overestimate real-world outcomes due to enriched populations, intensive lifestyle counseling, and monitoring not replicated in practice.

Expert Disagreements

IssuePro – Lifetime TherapyCon – Potential Taper/Stop
Continuation vs. cessationWithdrawal trials show rapid regain; indefinite use required to maintain effectSURMOUNT-4 lower-dose arm (5 mg tirzepatide) preserved ~17% loss vs 22% on full dose; some suggest tapered maintenance possible
CV class effectSemaglutide proven; class effect plausible due to shared mechanismNo CVOT for tirzepatide in obesity; oral semaglutide lower dose less effective; class effect not proven
Pancreatitis riskLarge meta-analyses show no class-wide signalFDA label still warns; case reports in real-world use
Optimal maintenance dosingMax tolerated dose for maximal effectStep-down to lower dose may preserve most loss; reduces cost/side effects (SURMOUNT-MAINTAIN)
Switching agents for maintenanceOral orforglipron maintained weight loss after switching from injectables over 52 weeks (RCT 2025)Durability beyond 1 year unknown; limited generalizability

Trade-offs Decision Table

StrategyProsCons
Continuous full-dose therapyMaximal weight loss maintenance (15–24% total), proven CV benefit, sustained metabolic improvementsHigh cost ($900–$1,500/month US list), high GI side effect burden, unknown beyond ~4 years
Continuous lower-dose therapyReduced cost, fewer side effects, potentially better tolerability for maintenanceLess total weight loss maintained (e.g., ~17% vs 22% with tirzepatide 5 mg in SURMOUNT-4; evidence limited)
Drug holiday/discontinuationNo drug cost, no side effects, freedom from daily injections~60–75% weight regain within 12–18 months, loss of CV benefit, rebound of BP, lipids, glycemic control
Structured tapering with lifestyleCould reduce regain rate; lower cost and side effect burdenNo established protocol; SURMOUNT-MAINTAIN is first step; behavioral bridging not yet tested

Practical Recommendations

  1. Indication: GLP-1 RAs should be presented as chronic therapy; patients should understand that discontinuation will likely lead to weight regain.
  2. Shared decision-making: Discuss tolerance of GI side effects, cost, and commitment before initiation.
  3. Dose selection: Start with standard titration; if full dose not tolerated, consider lower maintenance dose (e.g., tirzepatide 5 mg) based on SURMOUNT-MAINTAIN data—benefit may be partial but still superior to placebo.
  4. Lifestyle co-intervention: All trials included diet/exercise counseling; maximize lifestyle support to enhance real-world outcomes.
  5. Monitoring: Check for gallbladder symptoms; educate about aspiration risk before procedures requiring anesthesia.
  6. CV risk: For patients with established CVD and BMI ≥27 without diabetes, semaglutide 2.4 mg has proven benefit; consider as part of comprehensive risk reduction.
  7. Discontinuation: If patient chooses to stop, provide structured support including gradual dose tapering and intensive lifestyle counseling to potentially attenuate regain (no RCT evidence yet).

Unresolved Uncertainties

  1. Durability beyond 5 years: No RCT data beyond ~4 years (SELECT).
  2. Primary prevention CV benefit: SELECT enrolled only secondary prevention patients.
  3. Optimal dose-reduction or tapering protocols for maintenance.
  4. Intermittent or cycling dosing strategies (no evidence).
  5. Long-term thyroid cancer risk: Detection bias hypothesis not confirmed.
  6. Adherence interventions: No proven strategy to improve real-world persistence.

Caveats & Limitations

  • All pivotal trials are industry-sponsored (Novo Nordisk, Eli Lilly), raising potential for publication bias.
  • Withdrawal trial designs (STEP 4, SURMOUNT-4) use enriched populations (run-in tolerators/responders), limiting generalizability.
  • Lifestyle co-intervention in trials not routinely replicated in real-world care.
  • Meta-analyses pool largely same manufacturer-funded trials; no formal assessment of publication bias presented.
  • FDA black box warning for MTC remains despite conflicting human evidence.