Semaglutide evidence atlas and traceability

What the evidence supports

• STEP 1 established substantial 68-week weight loss in adults without diabetes.
• STEP 2 confirmed benefit in adults with type 2 diabetes, with a smaller average response than in populations without diabetes.
• STEP 5 demonstrated persistence of benefit through two years during continued treatment.
• STEP 6 supplied randomized evidence in an East Asian population.
• OASIS 1 demonstrated efficacy of a high-dose oral formulation in adults without diabetes.
• STEP 10, SELECT analyses, and the osteoarthritis trial extend evidence to prediabetes, long-term cardiovascular populations, and obesity-related functional disease.

Conflicting evidence and interpretive limits

• Trial efficacy exceeds many routine-care estimates because adherence, support, access, dose escalation, and persistence differ.
• Lean-mass reductions occur alongside fat loss; available studies use heterogeneous methods and do not establish a single clinically meaningful lean-mass effect.
• Head-to-head observational comparisons can be confounded by treatment selection, shortages, dose achieved, and discontinuation.

Geographic applicability

The STEP program was multinational, and STEP 6 directly improves East Asian applicability. Representation remains incomplete for many African, Latin American, Indigenous, low-income, and medically complex populations.

Regulatory interpretation

Weight-management indications exist in multiple jurisdictions, but eligible populations, formulations, cardiovascular-risk language, prescribing restrictions, and reimbursement differ. Current local labels and health-technology guidance control clinical use.

Priority research gaps

• Long-term outcomes beyond the major trial periods
• Optimal maintenance strategies after substantial weight loss
• Body-composition and functional outcomes in older or frail adults
• Comparative effectiveness across diverse health systems
• Access, affordability, and discontinuation effects

What the evidence supports

• SUSTAIN 1 demonstrated monotherapy efficacy against placebo.
• The broader SUSTAIN program compared semaglutide with active glucose-lowering treatments and insulin strategies.
• PIONEER 1 established oral monotherapy efficacy; the PIONEER program expanded oral evidence across treatment settings.
• SUSTAIN 6, FLOW, and SOUL provide outcome evidence in cardiovascular and kidney high-risk populations.

Conflicting evidence and interpretive limits

• Average glycemic and weight effects vary by baseline HbA1c, background therapy, achieved dose, adherence, and comparator.
• Rapid glycemic improvement can complicate interpretation of early retinopathy events in susceptible patients.
• Outcome-trial findings should not be generalized to every person with type 2 diabetes or to primary prevention without qualification.

Geographic applicability

The development program included multinational populations, including dedicated Japanese and broader Asian studies, but trial access and representation remain uneven across regions.

Regulatory interpretation

Semaglutide is authorized for type 2 diabetes in many jurisdictions. Product, formulation, cardiovascular-risk, kidney-risk, and combination-therapy language differs by regulator and changes over time.

Priority research gaps

• Durability under routine-care adherence
• Comparative sequencing with SGLT2 inhibitors and newer incretin therapies
• Outcomes in underrepresented ethnic and socioeconomic populations
• Long-term microvascular outcomes

What the evidence supports

• SUSTAIN 6 established cardiovascular safety and suggested benefit with injectable semaglutide in high-risk type 2 diabetes.
• PIONEER 6 established cardiovascular safety for oral semaglutide but was not powered as a definitive superiority trial.
• SELECT demonstrated reduced major adverse cardiovascular events in adults with established cardiovascular disease and overweight or obesity without diabetes.
• SOUL demonstrated cardiovascular outcome benefit with oral semaglutide in high-risk type 2 diabetes.

Conflicting evidence and interpretive limits

• Evidence is strongest for secondary prevention and high-risk diabetes populations, not universal primary prevention.
• Component outcomes and subgroup estimates should not be interpreted as independently definitive when interaction testing is absent or underpowered.
• Mechanisms may include weight, glycemic, inflammatory, hemodynamic, and direct pathways; mediation analyses do not prove a single causal route.

Geographic applicability

Large multinational outcome trials improve generalizability, but participants still reflect trial eligibility, specialist access, and regional recruitment patterns.

Regulatory interpretation

Cardiovascular-risk-reduction indications differ by product and jurisdiction. Regulatory wording must be displayed product-by-product rather than inferred from the drug class.

Priority research gaps

• Primary-prevention populations
• Long-term effects after treatment interruption
• Comparative cardiovascular effectiveness versus dual agonists
• Implementation in lower-resource health systems

What the evidence supports

• FLOW is the pivotal dedicated kidney-outcomes trial in type 2 diabetes and chronic kidney disease.
• SELECT kidney analyses extend evidence to people with overweight or obesity and established cardiovascular disease without diabetes.
• A randomized trial in obesity-related CKD without diabetes found improvements in albuminuria and related measures.
• FLOW subgroup analysis did not show evidence that concomitant SGLT2 inhibitor use eliminated benefit, but subgroup precision is limited.

Conflicting evidence and interpretive limits

• Earlier cardiovascular trials relied heavily on albuminuria-driven composites rather than kidney failure outcomes.
• Evidence in non-diabetic CKD remains smaller and shorter than FLOW.
• Subgroup analyses should not be treated as head-to-head combination-therapy trials.

Geographic applicability

FLOW was multinational, but applicability is strongest for patients meeting its diabetic CKD criteria and receiving comparable background care.

Regulatory interpretation

Kidney-risk language and approved populations vary by jurisdiction and product. Clinical interpretation should follow current regulator-approved labeling and kidney guidelines.

Priority research gaps

• Non-diabetic CKD outcomes
• Advanced kidney failure and dialysis populations
• Optimal integration with SGLT2 inhibitors and finerenone
• Kidney outcomes after discontinuation

What the evidence supports

• The phase 2 NASH trial demonstrated steatohepatitis resolution but did not establish a significant fibrosis-stage improvement in its primary analysis.
• ESSENCE phase 3 interim evidence strengthened support for histologic benefit in non-cirrhotic MASH with fibrosis.
• The compensated-cirrhosis phase 2 trial did not demonstrate significant fibrosis improvement or NASH resolution.

Conflicting evidence and interpretive limits

• Positive non-cirrhotic findings must not be generalized to established cirrhosis.
• Histologic surrogate improvement is not identical to demonstrating fewer liver failures, transplantations, or deaths.
• Weight loss contributes to liver improvement, and the proportion attributable to direct hepatic mechanisms remains uncertain.

Geographic applicability

Trials were multinational but biopsy requirements, specialist centers, and exclusion criteria constrain generalizability to routine liver care.

Regulatory interpretation

MASH authorization and product branding differ by jurisdiction and have changed rapidly. The page must present current regulator-specific status and must not infer global approval from one agency.

Priority research gaps

• Hard liver outcomes
• Durability of fibrosis benefit
• Decompensated or advanced cirrhosis
• Comparative and combination therapy
• Long-term post-treatment outcomes

What the evidence supports

• STEP-HFpEF demonstrated improved health status and function in obesity-related HFpEF without diabetes.
• STEP-HFpEF DM replicated clinically important symptom and functional benefits in participants with type 2 diabetes.
• SELECT and FLOW prespecified analyses support reduced worsening-heart-failure events in broader high-risk populations.
• A pooled analysis across SELECT, FLOW, and the STEP-HFpEF trials strengthens consistency.

Conflicting evidence and interpretive limits

• The STEP-HFpEF trials were not designed as definitive mortality trials.
• Benefits cannot automatically be generalized to HFrEF, non-obesity-related HFpEF, or unstable heart failure.
• Weight loss may mediate part of the effect, but symptom and event pathways remain incompletely resolved.

Geographic applicability

Multinational recruitment supports broad relevance, but enrolled participants represent selected obesity-related HFpEF phenotypes.

Regulatory interpretation

Heart-failure evidence may inform guidelines and labeling differently across jurisdictions; the page must distinguish trial evidence from an approved heart-failure indication.

Priority research gaps

• Mortality and hospitalization as primary outcomes
• HFrEF populations
• Non-obesity-related HFpEF
• Long-term functional durability

What the evidence supports

• Randomized programs consistently show increased nausea, vomiting, diarrhea, constipation, and related discontinuation.
• Gallbladder events occur more often in some weight-loss trials, potentially influenced by rapid weight loss.
• Large outcome trials have not established broad excess cardiovascular or all-cause mortality risk.
• Regulators continuously evaluate pancreatitis, thyroid, psychiatric, anesthesia/aspiration, ocular, counterfeit, and compounded-product concerns.

Conflicting evidence and interpretive limits

• Rare-event estimates are imprecise in individual trials.
• Spontaneous-reporting systems identify signals but cannot establish incidence or causality.
• Animal thyroid C-cell findings and jurisdiction-specific warnings should not be converted into unsupported claims about human cancer risk.
• Compounded or counterfeit products are not interchangeable evidence for authorized semaglutide products.

Geographic applicability

Safety reporting differs substantially across regulatory systems, prescribing environments, formulations, and access pathways.

Regulatory interpretation

The page must display current product-specific FDA, EMA, and other regulator warnings separately and retain the date of each assessment.

Priority research gaps

• Very rare adverse events
• Long-term use across decades
• Pregnancy and reproductive outcomes
• Peri-procedural aspiration risk
• Risks from compounded and counterfeit products
• Causal interpretation of emerging ocular signals

What the evidence supports

• SUSTAIN 6 identified more adjudicated diabetic-retinopathy complications, concentrated among participants with pre-existing disease and substantial early HbA1c reduction.
• Post hoc analyses support an early-worsening mechanism rather than proof of direct retinal toxicity.
• Multiple observational analyses have examined NAION, with some reporting elevated relative risk.

Conflicting evidence and interpretive limits

• NAION studies use different populations, comparators, exposure definitions, and outcome ascertainment.
• Absolute risk is low and confounding by diabetes, obesity, sleep apnea, vascular disease, and healthcare contact remains material.
• Observational associations do not prove causality.

Geographic applicability

Available ophthalmic datasets are concentrated in selected health systems and specialty populations; absolute risks may differ by baseline disease and access to eye care.

Regulatory interpretation

Regulatory conclusions and label language should be tracked separately by jurisdiction. Patients with diabetic eye disease require individualized clinical monitoring rather than generalized alarm.

Priority research gaps

• Prospective adjudicated NAION data
• Mechanistic evidence
• Risk by baseline retinopathy severity
• Effect of rate of glycemic improvement
• Absolute risk across diverse populations

What the evidence supports

• STEP 4 randomized continued therapy against withdrawal after run-in and demonstrated divergent weight trajectories.
• The STEP 1 extension documented substantial regain during one year off treatment.
• Real-world studies show high discontinuation and variable reinitiation, although reasons and outcomes are incompletely captured.

Conflicting evidence and interpretive limits

• Trial withdrawal does not model every tapering, behavioral, or alternative-medication strategy.
• Extension cohorts are vulnerable to selection and missing follow-up.
• Real-world discontinuation may reflect cost, shortages, adverse effects, goal attainment, or access rather than pharmacology alone.

Geographic applicability

Maintenance and discontinuation are strongly shaped by insurance, national reimbursement, supply, and clinical follow-up.

Regulatory interpretation

Regulatory approval does not define an evidence-based stopping strategy. Local prescribing guidance and individualized clinical care remain necessary.

Priority research gaps

• Tapering versus abrupt discontinuation
• Lower-dose maintenance
• Behavioral and resistance-training maintenance
• Switching strategies
• Long-term outcomes after repeated stopping and restarting

What the evidence supports

• Large electronic-health-record analyses demonstrate clinically meaningful weight loss with semaglutide in practice.
• Comparative cohorts often find greater weight loss with tirzepatide, but are not randomized.
• Utilization studies identify discontinuation, reinitiation, access, and supply as major determinants of outcomes.
• Real-world data can detect rare or delayed safety signals not well measured in trials.

Conflicting evidence and interpretive limits

• Medication exposure, formulation, dose achieved, adherence, and source of medication may be misclassified.
• Confounding by indication and treatment selection persists even after statistical adjustment.
• Commercially insured and integrated-system cohorts may not represent uninsured or lower-resource populations.

Geographic applicability

Much current evidence comes from the United States and other high-income health systems; worldwide reimbursement and access differences materially affect effectiveness.

Regulatory interpretation

Observational effectiveness and safety signals complement but do not replace randomized evidence or regulator adjudication.

Priority research gaps

• Comparable international cohorts
• Verified dose and adherence
• Long-term persistence
• Low-resource settings
• Outcomes after shortages and forced switching

What the evidence supports

• STEP TEENS demonstrated substantial BMI reduction and improved several cardiometabolic measures over 68 weeks.
• The adverse-event pattern was broadly consistent with adult experience, with gastrointestinal and gallbladder events requiring attention.
• Regulatory authorization in adolescents exists in several jurisdictions under defined criteria.

Conflicting evidence and interpretive limits

• The randomized sample was modest for uncommon harms.
• The trial duration cannot resolve long-term growth, pubertal, bone, reproductive, or psychosocial outcomes.
• Adult maintenance and cardiovascular-outcome evidence cannot be assumed to apply to adolescents.

Geographic applicability

The pivotal evidence was multinational but too small to characterize many regional, developmental, and socioeconomic subgroups.

Regulatory interpretation

Age, weight, BMI, comorbidity, formulation, and continuation criteria differ by regulator and payer. Current local pediatric labeling controls use.

Priority research gaps

• Multi-year growth and developmental outcomes
• Bone health and body composition
• Mental-health and eating-disorder outcomes
• Maintenance and discontinuation
• Younger children
• Transition from pediatric to adult care