This Week in Endocrinology — Aug 18, 2026

Welcome to This Week in Endocrinology. This week we're covering 10 notable papers spanning new frontiers in type 2 diabetes pharmacotherapy, precision genetics and personalised care across the diabetes spectrum, and long-term risk in survivors — from childhood-onset type 1 diabetes to young people treated for thyroid cancer. Let's dive in.

We start with incretin-based therapy, where two phase 3 trials in The Lancet Diabetes and Endocrinology push in different directions — one toward simplifying injectable therapy, the other toward avoiding injections altogether. In COMBINE 4, reported by Ji and colleagues, 485 insulin-naive adults with type 2 diabetes and a baseline HbA1c of around nine and a half percent were randomised, open-label, to once-weekly IcoSema — the fixed combination of insulin icodec and semaglutide — or to once-daily glargine U100, both titrated to target over 40 weeks. IcoSema lowered HbA1c by about three and a third percentage points versus just under two and a half with glargine, a treatment difference of roughly nine tenths of a percentage point in favour of the combination. The weight story was arguably more striking: bodyweight fell slightly on IcoSema while it rose by nearly four kilograms on glargine, a between-group difference of about four and a half kilograms. Clinically significant or severe hypoglycaemia was also roughly halved, at about 0.3 versus 0.6 episodes per person-year. Gastrointestinal side effects were the main tolerability issue, as you would expect from the semaglutide component. For a patient failing oral agents who needs insulin, this is a once-weekly injection that outperforms daily basal insulin on glucose, weight and hypoglycaemia simultaneously — though remember this was open-label, and the titration target was tight.

Alongside that, Yabe and colleagues report ACHIEVE-J, a 52-week open-label safety study of the oral GLP-1 receptor agonist orforglipron in just over 400 Japanese adults with type 2 diabetes, randomised to 3, 12 or 36 milligrams daily on top of diet and exercise or one or two oral agents. The primary endpoint here was safety, not efficacy, and the signal is dose-dependent tolerability rather than any new hazard. Around 85 percent of participants had at least one treatment-emergent adverse event, most mild or moderate, but discontinuation for adverse events climbed from about one in twenty at the lowest dose to about one in seven at 36 milligrams, driven by gastrointestinal symptoms. Level 2 hypoglycaemia was rare — around two percent in the higher-dose groups — and there was no severe hypoglycaemia. The practical message is that an oral GLP-1 receptor agonist appears acceptably safe over a year in an east Asian population, but that the top dose carries a meaningful attrition cost, which argues for slow titration and honest counselling about nausea.

Staying with drug classes, the European Journal of Endocrinology published a large target-trial emulation from Wang and colleagues that widens the lens on SGLT-2 inhibitors well beyond glucose. Using linked English primary care, hospital and mortality data from 2012 to 2023, they compared more than 364,000 adults over 40 with type 2 diabetes starting either an SGLT-2 inhibitor or a DPP-4 inhibitor, and tracked 59 prespecified long-term conditions across 17 organ systems. SGLT-2 inhibitor initiation was associated with roughly a quarter lower all-cause mortality, about a 14 percent lower risk of a first hospitalisation, and fewer hospitalisations overall. Risk was lower for 28 of the 59 conditions — including dementia, cancer, steatotic liver disease, diabetic foot ulcer, epilepsy and rheumatoid arthritis — and higher for four, notably candidiasis and diabetic ketoacidosis. This is observational, and a benefit spanning epilepsy and rheumatoid arthritis should make any of us think about residual confounding and healthier-user effects. But the direction of travel is consistent with the trial evidence, and the ketoacidosis and genital mycotic infection signals are real and actionable at the point of prescribing.

That liver finding connects neatly to a review in Diabetologia from Yki-Järvinen and Roden, who argue that metabolic dysfunction-associated steatotic liver disease should be treated as a core feature of type 2 diabetes rather than an incidental comorbidity. Their central number is uncomfortable: up to 95 percent of MASLD cases remain undiagnosed, partly because there is no accepted standard for assessing steatosis, while non-invasive tests tend to overestimate progression in type 2 diabetes. They emphasise that the combination of type 2 diabetes and metabolic syndrome features strongly predicts liver outcomes, yet cardiovascular disease remains the leading cause of death in non-cirrhotic MASLD — so finding fibrosis does not displace cardiometabolic risk management. They also flag genetic heterogeneity, with the PNPLA3 I148M variant raising liver risk while possibly protecting against cardiovascular disease. With drugs now approved for steatohepatitis, the bottleneck has shifted from treatment to case-finding, and that responsibility sits substantially with endocrinologists.

Precision approaches ran through several papers this week. In the Journal of Clinical Endocrinology and Metabolism, Clément and colleagues present a European expert consensus on rare monogenic obesity — POMC, PCSK1 and LEPR deficiencies and Bardet-Biedl syndrome — arguing that with melanocortin-4 receptor agonists now approved for conditions affecting as few as one in a million people, care should be concentrated in designated centres of expertise. Their framework is worth internalising: early and precise genetic diagnosis, equitable access for eligible patients, structured monitoring with a willingness to stop therapy when there is no measurable benefit, and real-world data collection to fill the long-term evidence gap. In Diabetes Care, Murray Leech and colleagues take a complementary genomic approach in over 1,100 insulin-treated people referred for MODY testing, adding a type 1 diabetes genetic risk score to the gene panel. Roughly one in five of those referred actually had genetic signatures consistent with type 1 diabetes, and among those whose MODY testing came back unsolved, the score reclassified about 16 percent as probable type 1 diabetes. As a pre-testing triage tool it discriminated poorly overall, but performed well in children under 10, where more than half of tests could have been avoided without missing MODY. So in young children the score can spare unnecessary panel testing; in adults it is better used to interpret a negative panel than to decide whether to order one.

Two further papers in Diabetes Care address type 1 diabetes management and outcomes. Nielsen and colleagues linked a New South Wales registry of over 5,200 children diagnosed with type 1 diabetes before age 16 to hospital, emergency, insurance and death records. Within 20 years of onset, incidence ranged from about nine per ten thousand person-years for cardiac complications to about 53 per ten thousand for kidney complications. Diagnosis before age seven, compared with diagnosis at 13 to 16, was associated with substantially lower 20-year risk across almost every outcome — peripheral neuropathy risk was around a sixth, severe retinopathy a quarter, cardiac complications about a third, and kidney complications roughly 60 percent lower — with lower-limb infections the exception. Importantly, risk did not differ by attained age, supporting the idea that prepubertal-onset disease progresses to complications more slowly. That is reassuring for families of very young children, but the corollary is that adolescent-onset diabetes deserves earlier and more intensive complication surveillance. Then Runchey and colleagues report a randomised trial of time-restricted eating in adults with type 1 diabetes and overweight or obesity — 32 participants assigned for six months to an eight-hour eating window without calorie counting, to 25 percent daily calorie restriction, or to no intervention. The primary outcome was negative: bodyweight did not change significantly with either strategy relative to control or between the two diets. HbA1c did fall modestly more with time-restricted eating than with calorie restriction, by about half a percentage point, and reassuringly there was no excess of ketoacidosis, severe hypoglycaemia or severe hyperglycaemia. With only 32 participants this is a safety and feasibility signal, not a weight-loss endorsement.

Rounding out the week, two papers address longer-term risk after intervention. In Diabetologia, Rogers and colleagues report a non-randomised phase 2 multicentre study of belatacept plus sirolimus as maintenance immunosuppression in islet transplantation for type 1 diabetes with hypoglycaemia unawareness, comparing nine recipients with a contemporaneous cohort of 24 on tacrolimus and mycophenolate. Eight of the nine on belatacept and sirolimus met the composite primary outcome of freedom from hypoglycaemia with detectable C-peptide and HbA1c under seven percent at 12 months, compared with about half of the tacrolimus group, and renal function remained unchanged only in the belatacept arm — with higher regulatory T cell proportions. Nine patients is a very small, non-randomised comparison, so treat this as a rationale for a proper trial rather than a change of standard. And in Thyroid, Bello and colleagues followed more than 5,200 survivors of differentiated thyroid cancer diagnosed before age 40, matched four-to-one to cancer-free controls over four decades. Second primary malignancies were about 50 percent more common in survivors and appeared after a shorter latency. The radioactive iodine gradient is the clinically important part: compared with untreated survivors, one course of radioactive iodine was associated with roughly double the risk of a second cancer, and two or more courses with nearly a tripling. Overall mortality was low and similar to controls, but was higher among those who developed second cancers or received radioactive iodine. That is a strong argument for risk-adapted, parsimonious radioactive iodine use in young patients and for surveillance that extends well beyond a decade.

If you only have time for one paper this week, make it COMBINE 4 in The Lancet Diabetes and Endocrinology [1]. A once-weekly injection that beats daily basal insulin on HbA1c, weight and hypoglycaemia at the same time changes how you frame the conversation about starting insulin.

Here are the key takeaways from this week in Endocrinology. First, once-weekly insulin icodec combined with semaglutide outperformed daily glargine across glucose, weight and hypoglycaemia in insulin-naive type 2 diabetes, with gastrointestinal effects the main trade-off [1]. Second, oral orforglipron looks acceptably safe over a year in Japanese adults, but discontinuation rises steeply at the 36 milligram dose, so titrate slowly [2]. Third, large-scale English data associate SGLT-2 inhibitors with lower mortality and hospitalisation and reduced risk across 28 long-term conditions, alongside genuine signals for candidiasis and ketoacidosis [3]. Fourth, steatotic liver disease is massively underdiagnosed in type 2 diabetes and case-finding now sits with us, since effective drugs exist [4]. Fifth, adolescent-onset type 1 diabetes carries substantially higher 20-year complication risk than prepubertal onset, and time-restricted eating did not produce significant weight loss in type 1 diabetes though it appeared safe [7,9]. And finally, cumulative radioactive iodine dose tracks with second primary malignancy risk in young thyroid cancer survivors, reinforcing risk-adapted use and long surveillance [10].

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 a quick rating on Apple Podcasts or Spotify helps other physicians discover the show.

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