This Week in Endocrinology — Sep 30, 2026

This AudioScholar briefing is generated by artificial intelligence for healthcare professionals and trainees. It is not medical advice.

Welcome to This Week in Endocrinology. This week we're covering 10 notable papers spanning the next generation of obesity and incretin pharmacotherapy, oral GLP-1 receptor agonists and their cardiovascular safety, and a set of studies on special populations, prevention and how we define disease modification. Let's dive in.

We start with the crowded and fast-moving field of obesity pharmacotherapy, where two phase 3 trials of the triple hormone receptor agonist retatrutide landed this week in two different journals. In the New England Journal of Medicine, Jastreboff and colleagues report TRIUMPH-1, in which 2,339 adults with obesity but without diabetes were randomised to once-weekly retatrutide at 4, 9 or 12 milligrams or placebo for 80 weeks [2]. Mean weight loss was about 18 percent at the lowest dose and about 25 percent at 12 milligrams, against roughly 4 percent with placebo — a placebo-adjusted difference of around 21 percentage points, which is the largest reduction yet reported in a phase 3 obesity trial. The trial also carried co-primary outcomes in two embedded comorbidity cohorts, and both moved: among the 574 participants with knee osteoarthritis, pain scores improved by roughly one and a half to two points more than placebo on a ten-point scale, and among the 243 participants with obstructive sleep apnoea, the apnoea-hypopnoea index fell by something over 20 events per hour more than placebo. Adverse events were predominantly gastrointestinal. Alongside that, in The Lancet, Bellido and colleagues report TRIUMPH-2, the companion trial in 1,152 adults with obesity and type 2 diabetes [3]. Here weight loss was attenuated, as it usually is in diabetes: about 19 percent at the 12 milligram dose versus about 5 percent on placebo, a difference of roughly 14 percentage points. Diarrhoea affected about a third of participants on the higher doses and nausea about a quarter, and notably hypotension and dysesthesia were both more frequent with retatrutide than placebo, in the range of 5 to 6 percent at higher doses — a signal worth watching as these agents move toward wider use. Taken together, these two trials establish retatrutide's efficacy ceiling, but neither reports cardiovascular outcomes, and neither has regulatory approval behind it yet.

A contrasting approach appears in The Lancet Diabetes and Endocrinology, where Garvey and colleagues report ZUPREME 1, a phase 2 trial of petrelintide, a long-acting human amylin analogue with a mechanism entirely separate from the incretins [5]. Among 485 adults with obesity and without diabetes randomised across five maintenance doses or placebo, weight fell by about 8 to 10 percent at 28 weeks versus under 2 percent with placebo, with continued decline to nearly 11 percent by week 42. What stands out is tolerability: nausea occurred in about one in five participants on petrelintide compared with about 6 percent on placebo, but vomiting, diarrhoea and constipation rates were essentially the same as placebo. Efficacy also plateaued above 5 milligrams, which has dosing implications. This is phase 2 evidence in a mostly White cohort without diabetes, so it supports further evaluation rather than any conclusion about where amylin analogues sit relative to incretins.

The second theme is the oral GLP-1 story, and here the most clinically consequential paper of the week. In The Lancet, Klein and colleagues report ACHIEVE-4, an event-driven phase 3 trial testing whether orforglipron, an oral non-peptide GLP-1 receptor agonist, is cardiovascularly safe [1]. Nearly 2,750 adults with type 2 diabetes, a body mass index of 25 or above, and established cardiovascular or chronic kidney disease were randomised open-label to orforglipron or titrated insulin glargine across 317 sites. Over a median two years, four-component major adverse cardiovascular events occurred in about 4 percent of the orforglipron group and about 5 percent on glargine, meeting the prespecified non-inferiority margin. This was a safety trial, not a superiority trial, so it should not be read as demonstrating cardiovascular benefit — but the direction is reassuring, and the secondary safety picture is informative: clinically significant or severe hypoglycaemia occurred in about 7 percent on orforglipron versus about 19 percent on insulin, while gastrointestinal adverse events affected about 62 percent of the orforglipron group compared with about 14 percent on insulin and were the commonest reason for stopping. There were 19 deaths in the orforglipron arm and 43 on insulin, all but one judged unrelated to treatment. Complementing this, Nature Medicine carries OUTSTAND-1 from Yu and colleagues, a phase 3 trial of safiglipron, another oral small-molecule GLP-1 receptor agonist taken without fasting or dietary restriction, in 284 adults in China with early type 2 diabetes managed on diet and exercise alone [4]. Placebo-adjusted HbA1c reductions at 32 weeks ranged from about 1.2 to about 1.45 percent, and roughly three quarters of treated participants reached an HbA1c below 7 percent compared with a quarter on placebo. Weight effects were modest, up to about 3.5 percent below placebo, and discontinuation for adverse events reached about 7 percent at the highest dose. So two independent oral agents are now showing peptide-like glycaemic effects, with one of them carrying dedicated cardiovascular safety data — though safiglipron's evidence comes from a single short trial in one country and in a population with very early disease.

Our third theme is safety and effectiveness in populations that trials usually exclude. In Diabetologia, Collier and colleagues used the French national health data system to examine first-trimester exposure to second-line glucose-lowering drugs across just over 19,000 pregnancies in women with pharmacologically treated pre-existing type 2 diabetes [7]. Compared with insulin alone, exposure to second-line agents overall — including GLP-1 receptor agonists, SGLT2 inhibitors, DPP-4 inhibitors and sulfonylureas — was not associated with an increased risk of major congenital malformations. However, stricter exposure definitions in sensitivity analyses raised the possibility of a small excess of cardiac malformations with GLP-1 receptor agonists and with sulfonylureas, and the authors are explicit that teratogenicity cannot be excluded. Body mass index and HbA1c were unavailable in the database, so residual confounding by diabetes severity is a real limitation; this is the largest dataset of its kind, and it is broadly reassuring without being definitive. Then in the Journal of Clinical Endocrinology and Metabolism, Oktavian and colleagues pooled 23 studies and 967 participants to assess automated insulin delivery in older adults and other high-risk groups with type 1 diabetes, including those with impaired hypoglycaemia awareness or cognitive or functional impairment [9]. Time in range improved by about 11 percentage points, equivalent to roughly two and three quarter hours a day, with the largest gains overnight, and time in hypoglycaemia fell modestly. Commercial and non-commercial systems performed comparably. Two caveats matter: psychosocial measures, including hypoglycaemia fear, did not improve significantly, and the certainty of evidence was rated very low, with substantial heterogeneity and only 11 randomised trials in the pool.

Finally, three papers about prevention, trajectory and how we measure success. In The Lancet Diabetes and Endocrinology, Tuomilehto and colleagues report 34 years of oral glucose tolerance test-based surveillance in Mauritius, across more than 36,000 survey attendances in seven national surveys [8]. Age and sex-standardised diabetes prevalence rose from about 15 percent in 1987 to a peak of about 23 percent in 2009, then fell significantly to about 20 percent in 2021, with intermediate hyperglycaemia falling further, to about 16 percent. The proportion of diabetes that was newly diagnosed at survey dropped from about two thirds to about a third, implying much better case-finding. This is the first national population to show a declining prevalence measured by the gold-standard tolerance test, and while the design is repeated cross-sectional and cannot attribute causality to any specific policy, it is a counterweight to the assumption that national diabetes burden only moves in one direction. On the question of what we should be measuring, Khunti and colleagues, also in The Lancet Diabetes and Endocrinology, report a modified Delphi process that produced a core outcome set for disease modification in type 2 diabetes — 35 outcomes across 11 domains, agreed across two rounds involving clinicians, researchers and people living with diabetes [6]. This is a consensus framework for trial design rather than clinical evidence, but it signals a shift in emphasis from glycaemic control toward altering long-term disease trajectory. And in Diabetologia, Samuelsson and colleagues report long-term follow-up of the DiAPREV-IT trial, in which 50 autoantibody-positive children received GAD-alum or placebo [10]. The original trial was negative, and over a median of nearly 13 years there was still no overall effect on progression to stage 3 type 1 diabetes. But among the 27 children carrying the HLA-DR3-DQ2 haplotype, GAD-alum cut the hazard of progression to roughly a third, with a median time to clinical diabetes of nine years versus about three and a half years on placebo, while children without that haplotype appeared to progress faster on treatment. These are small, post hoc genotype subgroups from a trial that missed its primary endpoint, so the finding is hypothesis-generating — but it is a concrete argument for genotype-stratified design in type 1 diabetes prevention trials.

If you only have time for one paper this week, make it ACHIEVE-4 in The Lancet [1]. It settles the open question of whether a non-peptide oral GLP-1 receptor agonist carries cardiovascular risk relative to insulin in high-risk type 2 diabetes — and it does so for the agent closest to reaching clinic.

Here is what this week's evidence adds up to in Endocrinology. First, retatrutide extends the efficacy frontier in obesity to around a quarter of body weight in people without diabetes, with parallel benefits in knee osteoarthritis pain and sleep apnoea severity, though cardiovascular outcome data and the hypotension and dysesthesia signals in the diabetes trial remain unresolved. Second, oral GLP-1 receptor agonism now has both glycaemic efficacy and dedicated cardiovascular safety evidence behind it, with the caveat that non-inferiority against insulin is not a demonstration of benefit, and that gastrointestinal intolerance drove most discontinuations. Third, amylin analogues look like a genuinely distinct option with notably placebo-like gastrointestinal tolerability apart from nausea, but that evidence is phase 2 only. Fourth, in pregnancy the largest dataset to date does not show an overall malformation excess with second-line glucose-lowering agents compared with insulin, while leaving a possible small cardiac signal open. And fifth, the pooled automated insulin delivery data support glycaemic benefit in older and high-risk adults with type 1 diabetes, but the certainty is very low and psychosocial gains were not demonstrated.

That's your roundup for This Week in Endocrinology. The full transcript and references are available on the episode page in your AudioScholar library. This is an AI-curated summary — for clinical decisions, always consult primary sources and current guidelines. See you next week.

If this weekly briefing is useful, follow the show in your podcast app so new episodes arrive automatically. And if it saves you time, a quick rating in your podcast app helps other physicians find the show.

This is an automated summary generated by artificial intelligence, which can make mistakes. Always review the original source materials.