Albiglutide

A once-weekly GLP-1 receptor agonist built as a tandem GLP-1 dimer fused to human albumin, approved as Tanzeum in 2014 for type 2 diabetes and withdrawn from sale by GSK in 2018 for commercial reasons, months before its cardiovascular outcomes trial read out positive.

Also referenced as: Tanzeum, Eperzan, GSK716155, GSK-716155, Albugon

Also appears in: Weight loss

COA Public product evidence No certificate records are bound to a tracked Albiglutide product page A record counts here only when a vendor publishes a certificate against one of its own tracked Albiglutide product pages. Other certificates naming this compound may still exist — search the wider certificate database. →

---

What is Albiglutide?

Albiglutide is an albumin fusion protein, not a free peptide. It consists of two tandem copies of a modified human GLP-1(7-36) sequence, each carrying a glycine in place of the alanine at position 8 to block cleavage by dipeptidyl peptidase-4, genetically fused to the N-terminus of recombinant human serum albumin. The whole molecule weighs 72,970 daltons and is expressed in a modified strain of the yeast Saccharomyces cerevisiae; it cannot be made by solid-phase peptide synthesis.

The construct began as Albugon at Human Genome Sciences, was co-developed with GlaxoSmithKline under the code GSK716155, and was approved in the United States as Tanzeum on April 15, 2014 and in the European Union as Eperzan on March 21, 2014, in both cases for type 2 diabetes. It was the first albumin-fusion incretin to reach the market; GSK withdrew it in 2018, and a 9,463-patient outcomes trial then showed a 22% reduction in major cardiovascular events. It is not a research-market product.

How it works

• Albumin fusion and half-life. Native GLP-1 is cleared within minutes by DPP-4 and renal filtration. Fusing a DPP-4–resistant GLP-1 dimer to albumin produced a molecule that activated GLP-1 receptor-dependent cAMP signaling, lowered glycemic excursions and stimulated insulin secretion in wild-type but not GLP-1 receptor knockout mice, reduced food intake, and inhibited gastric emptying, although with roughly 100-fold lower potency than exendin-4 in cell-based assays (Baggio et al., Diabetes, 2004;53(9):2492–2500). In the first-in-human study of 39 healthy volunteers, albiglutide showed a terminal half-life of 6 to 8 days and a time to peak concentration of 3 to 4 days, with dose-dependent reductions in weighted mean glucose and fructosamine (Bush et al., Diabetes, Obesity and Metabolism, 2009;11(5):498–505). The approved label gives a half-life of approximately 5 days and an apparent clearance of 67 mL/h, with catabolism presumed to follow the pathway of native albumin in the vascular endothelium.
• Glucose-dependent glycemic lowering across weekly, biweekly and monthly schedules. In a 16-week Phase 2 trial of 356 adults with type 2 diabetes, HbA1c fell by 0.87, 0.79 and 0.87 percentage points with 30 mg weekly, 50 mg every two weeks and 100 mg monthly, respectively, against 0.17 with placebo and 0.54 with twice-daily exenatide, while fasting glucose fell within the first weeks (Rosenstock et al., Diabetes Care, 2009;32(10):1880–1886). As with other GLP-1 receptor agonists, insulin release and glucagon suppression depend on elevated glucose, which kept intrinsic hypoglycemia rates low in trials where albiglutide was not combined with a sulfonylurea or insulin.
• A large molecule with limited central access, and a correspondingly small weight effect. Albiglutide’s effect on body weight in the HARMONY program was close to that of placebo, a finding reviewers attribute to limited penetration of the large albumin fusion into the central nervous system and its appetite circuits (Brønden et al., Clinical Pharmacokinetics, 2017;56(7):719–731). In HARMONY 3, weight change over 104 weeks was −1.21 kg with albiglutide against −1.00 kg with placebo and −0.86 kg with sitagliptin (Ahrén et al., Diabetes Care, 2014;37(8):2141–2148). This is the clearest pharmacological difference between albiglutide and the acylated agonists semaglutide and liraglutide.
• Delayed gastric emptying with comparatively mild gastrointestinal effects, offset by injection-site reactions. The label states that albiglutide delays gastric emptying and may affect absorption of oral drugs taken at the same time. In the head-to-head HARMONY 7 trial, gastrointestinal adverse events were reported by 35.9% of patients on albiglutide versus 49.0% on liraglutide, while injection-site reactions were more than twice as frequent with albiglutide (12.9% versus 5.4%) (Pratley et al., Lancet Diabetes & Endocrinology, 2014;2(4):289–297). Across the pooled placebo-controlled trials, injection-site reactions occurred in 18% of albiglutide patients versus 8% on placebo, and 5.5% of 2,098 exposed patients developed anti-albiglutide antibodies, none of which neutralized the drug in an in vitro bioassay.

Research status

Albiglutide has a complete, sponsor-run Phase 3 program in type 2 diabetes and a positive placebo-controlled cardiovascular outcomes trial. Development stopped in 2018, so the record is closed: there is no obesity program, no pediatric program, and no ongoing trial.

The HARMONY program in type 2 diabetes. Eight Phase 3 trials supported the 2014 approvals. HARMONY 2 (NCT00849017) tested monotherapy in 309 drug-naive adults: at 52 weeks, HbA1c fell 0.84 and 1.04 percentage points more than placebo with 30 mg and 50 mg weekly, with injection-site reactions in 17.8% and 22.2% versus 9.9% (Nauck et al., Diabetologia, 2016;59(2):266–274). HARMONY 3, a 104-week double-blind trial in 1,012 metformin-treated adults, is the strongest of the add-on studies: at two years albiglutide lowered HbA1c by 0.9 points more than placebo, 0.4 more than sitagliptin and 0.3 more than glimepiride, with a hyperglycemic rescue rate of 25.8% against 59.2%, 36.4% and 32.7% (Ahrén et al., Diabetes Care, 2014;37(8):2141–2148). HARMONY 1 added albiglutide to pioglitazone in 310 adults, giving a 0.8-point placebo-adjusted HbA1c reduction at 52 weeks with no weight effect (Reusch et al., Diabetes, Obesity and Metabolism, 2014;16(12):1257–1264). HARMONY 8 compared albiglutide with sitagliptin in 507 adults with mild to severe renal impairment and found a larger HbA1c reduction at 26 weeks (−0.83 versus −0.52 percentage points) with similar tolerability (Leiter et al., Diabetes Care, 2014;37(10):2723–2730).

Two trials pitted albiglutide against insulin. In HARMONY 4 (NCT00838916), 779 adults on metformin with or without a sulfonylurea were randomized 2:1 to albiglutide or insulin glargine; at 52 weeks the HbA1c difference of 0.11 points (95% CI −0.04 to 0.27) met the 0.3-point noninferiority margin, with a 2.61 kg weight advantage for albiglutide and less documented symptomatic hypoglycemia (17.5% versus 27.4%) (Weissman et al., Diabetologia, 2014;57(12):2475–2484). In HARMONY 6, albiglutide added to titrated glargine was noninferior to thrice-daily prandial insulin lispro at 26 weeks (HbA1c −0.82 versus −0.66 points), with weight loss rather than gain and roughly half the rate of symptomatic hypoglycemia (15.8% versus 29.9%) (Rosenstock et al., Diabetes Care, 2014;37(8):2317–2325).

Where it fell short. HARMONY 7 (NCT01128894) was an open-label noninferiority comparison with once-daily liraglutide in 841 adults inadequately controlled on oral drugs. At 32 weeks HbA1c fell 0.78 percentage points with albiglutide and 0.99 with liraglutide; the treatment difference of 0.21 points had an upper confidence limit of 0.34, above the 0.3-point margin, so noninferiority was not shown (noninferiority p=0.0846) (Pratley et al., Lancet Diabetes & Endocrinology, 2014;2(4):289–297). HARMONY 5, in 685 adults already on metformin and glimepiride, found albiglutide 0.87 points better than placebo at 52 weeks but 0.25 points worse than pioglitazone, again failing noninferiority, although with 4.8 kg less weight gain and fewer hypoglycemic events (Home et al., Diabetes, Obesity and Metabolism, 2015;17(2):179–187). Albiglutide thus showed weaker glycemic efficacy than liraglutide and a near-null weight effect, and it required a lyophilized pen that needed reconstitution.

Cardiovascular safety and then benefit. A prospective meta-analysis of the eight Phase 3 trials and the Phase 2b trial, covering 5,107 patients, found no difference between albiglutide and comparators in adjudicated major adverse cardiovascular events plus unstable angina (58 versus 58 events; hazard ratio 1.00, 95% CI 0.68–1.49), while noting a numerical excess of atrial fibrillation or flutter in the albiglutide group (Fisher et al., Lancet Diabetes & Endocrinology, 2015;3(9):697–703). The dedicated outcomes trial, Harmony Outcomes (NCT02465515), randomized 9,463 adults aged 40 or older with type 2 diabetes and established cardiovascular disease at 610 sites in 28 countries to weekly albiglutide 30 to 50 mg or placebo. After a median 1.6 years, the primary composite of cardiovascular death, myocardial infarction or stroke occurred in 338 patients (7%) on albiglutide and 428 (9%) on placebo (hazard ratio 0.78, 95% CI 0.68–0.90; p=0.0006 for superiority). Acute pancreatitis (10 versus 7 cases), pancreatic cancer (6 versus 5) and medullary thyroid carcinoma (none) did not differ (Hernandez et al., Lancet, 2018;392(10157):1519–1529). The trial was stopped as planned once 611 primary events had accrued, with the last patient visit on March 12, 2018; by then GSK had already announced the withdrawal of the product.

Post-hoc analyses have continued to appear. The benefit was driven by myocardial infarction: first MIs were reduced by 25% (hazard ratio 0.75, 95% CI 0.62–0.91), consistently across ST-elevation and non-ST-elevation events (Krychtiuk et al., European Heart Journal — Cardiovascular Pharmacotherapy, 2024;10(4):279–288). In the 1,922 patients with a history of heart failure, albiglutide had no effect on cardiovascular death or heart-failure hospitalization (hazard ratio 1.06, 95% CI 0.79–1.43), against 0.73 in those without heart failure (Ferreira et al., European Journal of Heart Failure, 2022;24(10):1792–1801). The atrial fibrillation question raised by the 2015 meta-analysis was revisited in the full trial: albiglutide was associated with numerically fewer atrial fibrillation events (108 versus 131; hazard ratio 0.82, 95% CI 0.63–1.06), and its effect on the primary endpoint did not differ by baseline atrial fibrillation history (Krychtiuk et al., European Journal of Preventive Cardiology, 2026;33(1):30–41).

What was never done. There is no trial of albiglutide for chronic weight management, which is unsurprising given a weight effect indistinguishable from placebo in most HARMONY trials. No pediatric study was completed. A ready-to-use liquid formulation, intended to replace the reconstitution pen, was shown noninferior to the lyophilized product in a 26-week Phase 3 comparison of 308 adults (NCT02683746, completed May 15, 2017; Shaddinger et al., Diabetes Research and Clinical Practice, 2019;152:125–134) and had a long-term extension registered, but the program was halted before any liquid product was filed. As of November 2026 no company has revived the molecule, and PubMed shows no new interventional albiglutide trials since the Harmony Outcomes analyses.

Common dosage forms

• Dual-chamber pen, 30 mg and 50 mg. Single-dose dual-chamber pens holding lyophilized albiglutide, the only presentation approved in the United States; no longer manufactured.
• European Union presentation. The same 30 mg and 50 mg pens under the Eperzan name.
• Investigational liquid autoinjector. A liquid prefilled autoinjector tested in a 2016–2017 Phase 3 study that never reached approval.
• No other presentation. No vial, tablet, or multi-dose pen exists in any approved or investigational form.

Key considerations

• Regulatory status. The FDA Approved status reflects the April 15, 2014 approval of Tanzeum (BLA 125431) for glycemic control in adults with type 2 diabetes. GSK withdrew Tanzeum from sale in 2018 for commercial reasons; the BLA remains on Drugs@FDA as Discontinued with no license revocation, and the European authorization for Eperzan was withdrawn on October 29, 2018.
• A positive cardiovascular trial after discontinuation. Harmony Outcomes showed a 22% relative reduction in major adverse cardiovascular events, in the same range as liraglutide in LEADER and semaglutide in SUSTAIN-6, but several caveats apply. Median follow-up was 1.6 years, the benefit was driven by myocardial infarction with no significant reduction in cardiovascular death or stroke, patients with existing heart failure saw no reduction in heart-failure outcomes, and because weight loss was negligible the trial is often cited as evidence that GLP-1 receptor agonist cardiovascular protection does not depend on weight reduction.
• Labeled safety signals. Tanzeum carries the GLP-1 class boxed warning and the usual class warnings, and pancreatitis adjudicated as likely drug-related occurred in 0.3% of albiglutide patients across the eight Phase 3 trials versus 0% on placebo. Injection-site reactions (18% versus 8% on placebo) and a 5.5% anti-drug antibody rate were the practical tolerability costs of the albumin carrier.
• A fusion protein, not a synthesizable peptide. Albiglutide is a 72,970-dalton recombinant protein with an albumin domain, produced in yeast, which no peptide-synthesis route can produce and standard peptide purity assays cannot characterize. It is sometimes grouped with dulaglutide as a fusion GLP-1, but the two differ in carrier (albumin versus IgG4 Fc), expression system (yeast versus mammalian cells) and weight effect, and dulaglutide remains approved with its own cardiovascular indication.
• Naming. Albugon was the Human Genome Sciences name for the construct, GSK716155 the GSK development code, Tanzeum the US brand, and Eperzan the EU brand. All refer to the same molecule.