GLP-1 Agonists in Cushing's Syndrome: Efficacy, Safety, Data

THE PROTOHUMAN PERSPECTIVE#

Cushing's syndrome is one of the cruelest metabolic traps in endocrinology. Chronic hypercortisolism drives insulin resistance, visceral fat accumulation, muscle wasting, and a cascade of downstream metabolic dysfunction that conventional interventions barely dent. The central problem: you're fighting metabolic fires while the endocrine arsonist — excess cortisol — keeps pouring fuel. For anyone interested in metabolic optimization, Cushing's represents an extreme stress test for any therapeutic agent.

GLP-1 receptor agonists have become the most discussed pharmacological class in metabolic medicine. But until now, their use in hypercortisolemic states has been a clinical blank spot. The data from Iglesias et al. is the first multicenter attempt to fill that gap — and while it's small, the signal is real. Weight loss, glycemic improvement, and critically, no destabilization of the hypothalamic-pituitary-adrenal axis. That last point matters enormously. If GLP-1RAs can operate effectively even under the metabolic sabotage of cortisol excess, their downstream utility across stress-related metabolic pathology expands considerably.

THE SCIENCE#

What Are GLP-1 Receptor Agonists — And Why Cushing's Is the Hard Test#

GLP-1 receptor agonists are synthetic analogues of the endogenous incretin hormone glucagon-like peptide-1, which binds GLP-1 receptors and triggers Gs-mediated signaling cascades that promote glucose-dependent insulin secretion, suppress glucagon, delay gastric emptying, and modulate appetite via hypothalamic pathways^[3]. The currently approved agents — semaglutide, liraglutide, dulaglutide, and the dual GIP/GLP-1RA tirzepatide — have established efficacy in type 2 diabetes (HbA1c reductions of 1.5–2.0%) and obesity (weight loss of 7–24%), with cardiovascular risk reductions of 14–20% for major adverse events^[6].

The question nobody had properly asked: do these agents still work when cortisol is actively dismantling metabolic homeostasis?

Cushing's syndrome creates a uniquely hostile metabolic environment. Chronic cortisol excess drives hepatic gluconeogenesis, impairs peripheral glucose uptake, promotes lipolysis and paradoxical visceral lipogenesis, and disrupts insulin signaling at the receptor level^[1]. The metabolic derangements — obesity, insulin resistance, dyslipidemia, hypertension — cluster with near-universality in CS patients. And the usual pharmacological approaches (metformin, sulfonylureas, insulin intensification) offer limited traction against this degree of metabolic disruption.

The Cushing's-Specific Data: Iglesias et al. (2026)#

The multicenter retrospective study by Iglesias, Nobre, Hanzu et al. enrolled 20 adults with biochemically confirmed Cushing's syndrome and concurrent T2D and/or obesity, all treated with GLP-1RAs (liraglutide, semaglutide, tirzepatide, or dulaglutide) for at least three months^[1].

After a median follow-up of 13 months:

• Body weight decreased by 10.3 ± 8.8 kg (p < 0.001)
• BMI dropped by 4.2 ± 2.8 kg/m² (p < 0.001)
• Fasting glucose fell by 12.1 ± 35.3 mg/dL (p = 0.015)
• HbA1c decreased by 0.5 ± 0.9% (p = 0.011)
• Liver enzymes improved: ALT dropped 3.9 U/L (p = 0.005), GGT dropped 25.9 U/L (p = 0.012)
• ACTH and cortisol levels remained stable — no perturbation of adrenal axis dynamics

Only 26.3% of patients reported gastrointestinal side effects, all mild. No serious adverse events were attributed to GLP-1RA therapy.

The Catch, Though#

I need to be direct about the limitations here, because the enthusiasm around GLP-1RAs can blind people to what data actually does and does not prove.

n = 20. Twenty patients. Retrospective design. No control arm. The standard deviations on some of these outcomes are enormous — look at fasting glucose: −12.1 ± 35.3 mg/dL. That's a standard deviation nearly three times the mean effect. Some patients clearly responded dramatically; others barely moved. We don't know who got semaglutide vs. liraglutide vs. tirzepatide in what proportions, and the metabolic differences between those agents are not trivial.

The cortisol stability finding is genuinely important — and honestly, this is the data point I care about most. In a population where HPA axis disruption is the core disease, showing that GLP-1RAs don't further destabilize ACTH or cortisol is clinically meaningful even from a small cohort. But the glycemic and weight effects need replication in a controlled setting before anyone should treat this as definitive.

Orforglipron: The Oral GLP-1RA Pipeline#

While the Cushing's data addresses a niche but critical population, the broader GLP-1RA landscape is shifting toward oral formulations. Jamal, Khan, Qadri et al. published a GRADE-assessed meta-analysis of orforglipron — a non-peptide oral GLP-1RA — pooling five RCTs with 2,672 participants^[2].

The pooled results showed orforglipron reduced HbA1c by 1.07% (sensitivity analysis: −1.45%), body weight by 5.25 kg, BMI by 1.99 kg/m², and systolic blood pressure by 3.90 mmHg versus placebo. Fasting plasma glucose dropped significantly with clear dose-response patterns across the 3, 6, 24, and 36 mg dose groups. Heart rate increased by 6.38 bpm — a class effect that deserves more long-term scrutiny than it currently gets. Hypoglycemia and serious adverse events did not differ from placebo^[2].

The heterogeneity numbers, however, are striking. I² = 100% for HbA1c, 96% for weight, 97% for BMI. The sensitivity analyses clean this up substantially, but this level of between-study variance means the pooled estimates should be interpreted as directional, not precise.

The Broader Safety Landscape#

The umbrella review by Nature Communications — covering 123 meta-analyses, 464 outcomes, and 5,617 articles — provides the most extensive safety mapping available for the GLP-1RA class^[3]. GLP-1RAs showed improvements across endocrine, metabolic, cardiovascular, renal, and respiratory outcomes, with potential reductions in fracture risk and all-cause mortality. But they also flagged increased risks of diabetic retinopathy, ketoacidosis, GI events, and treatment discontinuation.

The VA database study by Nature Medicine, encompassing 215,970 GLP-1RA initiators compared across multiple control groups, adds real-world texture^[4]. Reduced risks of substance use disorders, seizures, neurocognitive disorders including Alzheimer's, and cardiometabolic events emerged. But so did increased gastrointestinal disorders, hypotension, syncope, arthritic conditions, and nephrolithiasis. This is the honest complexity of the class — it's not all upside.

The state-of-the-art safety review by Fiorentino et al. in Drugs reinforces that GI side effects remain the dominant tolerability barrier, generally emerging during dose titration and resolving with continued use^[5]. Cholelithiasis, cholecystitis, gastroparesis, and bowel obstruction are rarer but clinically relevant, especially in susceptible individuals. Suicidality signals have not been confirmed but surveillance continues. Safety data in pregnancy, pediatrics, and advanced organ impairment remain insufficient.

COMPARISON TABLE#

THE PROTOCOL#

For clinicians or patients considering GLP-1RA therapy in the context of Cushing's syndrome or severe cortisol-driven metabolic dysfunction — based on current evidence, which I want to stress is preliminary:

Step 1: Confirm biochemical status. Before initiating GLP-1RA therapy in suspected CS, ensure Cushing's syndrome is biochemically confirmed via 24-hour urinary free cortisol, late-night salivary cortisol, or low-dose dexamethasone suppression test. GLP-1RA therapy should not delay or replace treatment of the underlying cortisol source.

Step 2: Select the appropriate agent and initiate at the lowest dose. In the Iglesias et al. cohort, semaglutide, liraglutide, tirzepatide, and dulaglutide were all used. Start with the lowest available dose (e.g., semaglutide 0.25 mg weekly, liraglutide 0.6 mg daily). Titrate slowly — hypercortisolemic patients may have altered GI motility and gastroparesis risk is real in this population.

Step 3: Monitor HPA axis parameters alongside metabolic markers. Track ACTH, morning cortisol, fasting glucose, HbA1c, body weight, BMI, and liver enzymes (ALT, AST, GGT) at baseline, 3 months, and every 6 months thereafter. The Iglesias data showed cortisol stability, but individual responses in active CS may vary.

Step 4: Titrate based on tolerability, not just efficacy targets. Increase dose every 4–6 weeks if GI side effects remain manageable. In the CS cohort, 26.3% experienced mild GI effects — lower than the general population rate reported in broader reviews, which may reflect cautious titration or smaller sample noise.

Step 5: Assess hepatic response. The liver enzyme improvements (ALT −3.9 U/L, GGT −25.9 U/L) in the CS cohort are encouraging and align with data from non-CS populations showing GLP-1RA hepatoprotective effects. If liver enzymes worsen, investigate independently — CS itself drives hepatic steatosis.

Step 6: Evaluate at 6 and 12 months for continuation decisions. If weight loss exceeds 5% of baseline and HbA1c improves by ≥0.3%, continuation is justified. If no response at 6 months despite adequate titration, reassess whether underlying cortisol control is sufficient to permit metabolic improvement.

What is the evidence for using GLP-1 receptor agonists in Cushing's syndrome?#

The evidence is early-stage. One multicenter retrospective study by Iglesias et al. (2026) evaluated 20 CS patients on GLP-1RAs and found significant weight loss (−10.3 kg), BMI reduction (−4.2 kg/m²), and improved glycemic control without destabilizing cortisol levels. This is the first formal evaluation in this population, so I'd characterize the evidence as promising but requiring prospective confirmation.

How does orforglipron differ from injectable GLP-1 receptor agonists?#

Orforglipron is a non-peptide oral GLP-1RA taken daily, unlike injectable agents such as semaglutide (weekly) or liraglutide (daily). A meta-analysis by Jamal et al. (2026) across five RCTs showed orforglipron reduced HbA1c by 1.07% and weight by 5.25 kg versus placebo. The oral route could dramatically expand accessibility, but it's still in Phase III trials and not yet approved for clinical use.

Why do GLP-1 receptor agonists increase heart rate?#

GLP-1 receptor activation in the sinoatrial node and autonomic nervous system appears to drive a modest chronotropic effect. In the orforglipron meta-analysis, heart rate increased by approximately 6.38 bpm. Whether this translates to long-term cardiovascular risk remains unresolved — the cardiovascular outcome trial data for injectable GLP-1RAs is actually protective, which suggests the heart rate increase alone doesn't tell the full risk story.

Who should avoid GLP-1 receptor agonists?#

Patients with a personal or family history of medullary thyroid carcinoma or MEN2 syndrome should not use GLP-1RAs. Caution is also warranted in those with a history of pancreatitis, severe gastroparesis, or bowel obstruction. Safety data in pregnancy, lactation, and advanced hepatic or renal impairment remain insufficient, as the narrative review by Fiorentino et al. (2025) explicitly flags.

How long do gastrointestinal side effects typically last?#

GI side effects — primarily nausea, vomiting, diarrhea, and constipation — generally emerge during dose escalation and tend to resolve within 4–8 weeks of reaching a stable dose. Slow titration is the single most effective mitigation strategy. In the Cushing's cohort, only 26.3% reported mild GI symptoms, and none discontinued therapy for this reason.

VERDICT#

6.5/10.

Here's my honest assessment. The Cushing's-specific data from Iglesias et al. fills a genuine clinical gap — nobody had formally evaluated GLP-1RAs in this population, and the signal is encouraging. Weight loss of 10.3 kg and HbA1c reduction of 0.5% in a population where metabolic interventions routinely fail is worth paying attention to. The cortisol stability finding is clinically important.

But twenty patients, no control group, and retrospective design — this is hypothesis-generating, not practice-changing. I've seen too many small studies generate excitement that evaporates on replication. The broader GLP-1RA evidence base is strong and getting stronger, particularly with the oral pipeline represented by orforglipron, and the safety profile is increasingly well-characterized across massive cohorts. But for the specific claim — GLP-1RAs in Cushing's syndrome — we need a prospective trial with proper controls before this moves from "interesting signal" to "standard practice." I'd want at least n=100 with stratification by CS etiology before I'd change my protocol recommendations.