Semaglutide for Type 2 Diabetes: The Evidence Base

Semaglutide is approved for type 2 diabetes in major jurisdictions globally (including via the TGA in Australia for this indication). The evidence base comprises the SUSTAIN trial program (SUSTAIN 1–6 and SUSTAIN Plus substudies), enrolling >13,000 participants across diverse treatment contexts (monotherapy, add-on to metformin, add-on to thiazolidinediones, add-on to insulin, in patients with renal impairment, and in high-risk cardiovascular populations). This comprehensive evidence demonstrates semaglutide's efficacy for glycaemic control, cardiovascular protection, and renal benefit.

The SUSTAIN program employed a modular design allowing evaluation of semaglutide across the entire spectrum of type 2 diabetes care, from newly diagnosed to extensively pre-treated populations. This design provides evidence of real-world applicability and versatility as a glucose-lowering agent. SUSTAIN-6, the largest trial powered for cardiovascular outcomes, demonstrated cardiovascular benefit—a major advantage over many diabetes medications.

Across the SUSTAIN program, semaglutide consistently reduced HbA1c by 1.0–1.8 percentage points depending on baseline HbA1c, dose (0.5–1.4 mg weekly), and concomitant therapy. SUSTAIN 1, in treatment-naïve participants, showed HbA1c reduction of 1.5–1.8 percentage points with doses 1.0 mg and 1.4 mg, respectively. In SUSTAIN 2 (metformin-treated), similar efficacy was achieved. These reductions are among the largest seen with any single oral or injectable diabetes medication and compare favourably to insulin initiation or intensification.

Target HbA1c achievement rates were high: 50–70% of participants achieved HbA1c <7% (<53 mmol/mol) with semaglutide, compared to 20–35% with comparators (placebo, sitagliptin, or dulaglutide). Fasting plasma glucose was reduced by 20–30 mg/dL (1.1–1.7 mmol/L), and postprandial glucose excursions were substantially reduced. This comprehensive glycaemic benefit positions semaglutide as one of the most potent glucose-lowering agents available.

SUSTAIN-6 enrolled 3,297 participants with type 2 diabetes and established cardiovascular disease or high cardiovascular risk, randomising them to semaglutide 1.0 mg or placebo over 104 weeks. The trial demonstrated a 26% relative risk reduction in the composite primary endpoint (cardiovascular death, nonfatal myocardial infarction, nonfatal stroke), with hazard ratio 0.74 (95% CI 0.58–0.95). Individual components showed benefits: nonfatal myocardial infarction and nonfatal stroke reductions were statistically significant, while cardiovascular death showed a non-significant trend toward benefit.

SUSTAIN-6's cardiovascular benefit was particularly notable given that randomisation was stratified by cardiovascular risk, meaning all participants had either prior cardiovascular events or multiple risk factors. This demonstrates that semaglutide's cardioprotection is clinically meaningful in populations with substantial baseline cardiovascular burden. The benefit was consistent across subgroups (age, gender, baseline HbA1c, prior cardiovascular event status), supporting broad applicability.

Secondary outcomes in SUSTAIN trials included weight loss, blood pressure, and lipid parameters. Semaglutide-induced weight loss ranged from 2–4 kg across trials—modest compared to STEP obesity trials but meaningful in diabetes populations and contributing to cardiovascular benefit. Systolic blood pressure decreased by 2–4 mmHg, diastolic blood pressure by 1–2 mmHg. LDL cholesterol decreased by approximately 5–10%, and triglycerides by approximately 10–15%. These improvements in cardiovascular risk factors beyond glycaemic control likely contribute significantly to the observed cardiovascular benefit.

Notably, SUSTAIN 5 enrolled participants with moderate renal impairment and demonstrated that semaglutide remained efficacious for glycaemic control without worsening renal function. Urine albumin-to-creatinine ratio decreased with semaglutide, suggesting possible renal protective effects independent of glycaemic control. These findings support use of semaglutide in diabetes populations with comorbid renal disease.

The most common adverse events in SUSTAIN trials were gastrointestinal: nausea occurred in 25–35% of semaglutide recipients (vs. ~10% placebo), vomiting in 5–10%, and diarrhoea in 10–20%. Most events were mild-to-moderate and transient, resolving within 4–8 weeks despite continued treatment. Serious gastrointestinal events (acute pancreatitis, acute cholecystitis) occurred at similar frequencies in semaglutide and placebo arms, not suggesting a causal relationship. Heart rate increased by approximately 2–3 bpm with semaglutide; clinically significant arrhythmias were rare.

No increased incidence of thyroid cancer, thyroid nodules, or other malignancies was observed despite the thyroid C-cell tumour signal in preclinical rodent studies. This lack of clinical thyroid toxicity despite rodent signals suggests species-specific toxicology not translating to humans. Serious adverse events (hospitalizations, severe infections) occurred at similar frequencies in semaglutide and comparator arms.

While the SUSTAIN program provides robust efficacy evidence, several gaps remain. Long-term (>5 years) durability of glycaemic control is incompletely characterized; some observational data suggest gradual HbA1c increases over time despite continued treatment ('glucose creep'), though this requires confirmation. Additionally, trials enrolled predominantly white, middle-income populations; generalisability to more diverse populations and lower-resource settings remains uncertain.

Real-world effectiveness studies from health systems and registries show somewhat lower HbA1c reductions and weight loss than trials, reflecting lower adherence, higher discontinuation, and broader inclusion of patients with contraindications or intolerance. These real-world data temper expectations relative to trial results. Finally, whether semaglutide can prevent diabetes progression or remission (reversal) in early-stage disease remains incompletely explored; most SUSTAIN trials enrolled established diabetes populations.