This Week in General Medicine — Sep 21, 2026

Welcome to This Week in General Medicine. This week we're covering 10 notable papers spanning how we weigh benefit against harm in individual patients, a run of novel therapeutics tested in adolescents and in refractory autoimmune disease, and a cluster of work on cancer risk stratification and population-level prevention. Let's dive in.

We'll start with two papers that sharpen decisions you make face to face with a patient. In The BMJ, Hellemans and colleagues used registry data from five European countries and regions to emulate a randomised trial of deceased donor kidney transplantation versus staying on dialysis, following more than 64,000 waitlisted patients on dialysis, of whom just over 39,000 were transplanted between 2000 and 2019 [1]. For standard criteria donor kidneys, whether donation followed brain death or circulatory death, five year survival was better than continued dialysis at every recipient age, which is the reassuring half of the story. The benefit from expanded criteria donor kidneys, though, shrank steadily as recipients got older, and it essentially disappeared for expanded criteria kidneys donated after circulatory death. Among 75 year olds receiving that combination, five year survival was statistically indistinguishable from remaining on dialysis, around 59 percent versus 55 percent, and that group carried a threefold higher early post-transplant mortality. Diabetes and cardiovascular disease status did not change the picture. This is observational, and the estimates in the oldest patients are uncertain, but it gives you concrete language for a conversation with an older patient offered a marginal kidney: the early perioperative hit is real and the long-run survival dividend may not be there. Staying with what patients actually experience, JAMA published a multicentre, double-blind, placebo-controlled trial from Sweden in which 370 nulliparous women aged 18 to 31 received either intrauterine mepivacaine or saline instilled through a hydrosonography catheter two minutes before intrauterine device placement [2]. Pain during placement fell by about 15 millimetres on a 100 millimetre scale, roughly a quarter lower than placebo, and almost everyone in the mepivacaine group described the pain as tolerable compared with just over nine in ten on placebo, which works out to about one in fifteen women treated gaining a tolerable experience. Given that fear of pain demonstrably steers women away from intrauterine contraception, this is a cheap, practical addition worth raising with your gynaecology colleagues.

The second theme is novel therapeutics moving into younger and harder-to-treat populations. The headline trial is LATA, in The Lancet, reported by Bwakura-Dangarembizi and colleagues, which randomised 476 virologically suppressed adolescents with HIV across Kenya, South Africa, Uganda and Zimbabwe to either long-acting injectable cabotegravir and rilpivirine every eight weeks or daily oral dolutegravir, lamivudine and tenofovir [3]. Nearly all had acquired HIV vertically and had been on antiretrovirals for a median of close to twelve years. By week 96, confirmed viraemia occurred in two participants on injectables versus fifteen on daily oral therapy, roughly 1 percent against 6 percent. Injectable therapy was not just non-inferior, it was statistically superior. Seven participants permanently stopped injections, including two for drug reactions, and serious adverse events were comparable between groups. For an age group we know struggles with daily adherence, this is a strong signal that eight-weekly injections belong in the conversation, including in resource-limited settings. In Nature Medicine, a phase 1 trial from China led by Xue enrolled eleven adolescents aged 12 to 18 with severe or moderately severe haemophilia B and gave a single adeno-associated virus gene therapy encoding factor IX Padua [4]. Over 52 weeks there was no dose-limiting toxicity, with one case of transient liver enzyme elevation that normalised after immunosuppression. Mean factor IX activity at one year was about 42 international units per decilitre, and the annualised bleeding rate fell from roughly 14 episodes a year to half an episode. Eleven patients and one year of follow-up is a safety signal, not a durability claim, but it extends gene therapy below adulthood. And in Cell, Wang and colleagues report a first-in-human experience with a lipid nanoparticle messenger RNA encoding a CD19-targeting T cell engager, given to three patients with refractory secondary immune thrombocytopenia [5]. All three had rapid and complete peripheral B cell depletion with sustained depletion in marrow, durable platelet recovery through six months, and only grade 1 and 2 adverse events with no cytokine release syndrome. Three patients is a proof of concept, but it suggests a transient, non-cellular route to deep B cell depletion in autoimmune disease.

Our third theme is cancer risk, from the primary care consultation to the genome. In PLOS Medicine, Nicholson and colleagues examined more than 275,000 English primary care patients presenting with unexpected weight loss, of whom about one in twenty were diagnosed with cancer [6]. They asked whether the trend in a patient's blood tests over the preceding one to ten years discriminates cancer better than a single abnormal result. The most important finding is simpler than the question: adjusting blood test results for age and sex improved discrimination substantially for both approaches. Historical trend added further discrimination for some specific test-and-cancer combinations, with the best performance reaching an area under the curve of about 0.82, but trend analysis required at least two prior tests and the cohort is subject to testing bias. The practical message is that an albumin or platelet count should be interpreted against what is normal for that patient's age and sex, and against their own previous values, rather than against a single laboratory threshold. Science carried a striking genetic counterpart: LoPiccolo and colleagues analysed more than 3.3 million individuals and found that carrying the germline EGFR T790M variant significantly raises lung cancer risk, with no increased risk across seventeen other cancers and no interaction with polygenic risk [7]. The risk exceeds that conferred by smoking and is several-fold higher in never-smokers. Prevalence is higher in the United States than in British and Irish descendant populations, tracing to a Southern Appalachian founder event around two hundred years ago, affecting people of British, Irish and African descent in that region. If you practise in that part of the country and see familial or never-smoker lung cancer, targeted germline testing may become relevant. Tying these threads together, The Lancet published a framework piece from Rebbeck arguing for precision prevention and early detection [8], noting that only about 13 percent of cancers are picked up through guideline-based screening while between a fifth and half first present to an emergency department. The argument is that de-escalating prevention in genuinely low-risk people matters as much as escalating it in high-risk people, and that the real bottleneck is health system infrastructure and workforce, not the science.

Finally, two papers on population-level determinants of risk. In The Lancet Global Health, Gaye and colleagues pooled individual data from 109 nationally representative surveys across 76 countries, covering more than 315,000 adults, to map the care cascade for hypertension, diabetes and high cholesterol by household wealth [9]. Inequalities widened at every step of the cascade and were worst at the point of disease control. The regional patterns diverge sharply: in the Americas, control consistently favoured wealthier people, whereas in the African region coverage was uniformly low and the largest absolute gaps actually favoured the least wealthy, particularly for lipid treatment. Crucially, baseline cardiovascular risk was higher among the poorest, but achievable risk reduction tracked baseline risk rather than treatment gaps, and was greater in men than in women everywhere. In other words, the populations with the biggest treatment gaps are not necessarily those who stand to gain the most, which complicates how we target global interventions. And in PLOS Medicine, Luo and colleagues applied a sleep staging algorithm to wrist accelerometer data from nearly 96,000 UK Biobank participants followed for about nine years, mapping sleep against more than a thousand health outcomes [10]. More REM sleep was associated with lower incidence of 83 conditions and more deep sleep with seven, while greater night-to-night irregularity and more wakefulness after sleep onset tracked with higher risk. Minimum risk clustered in the six to eight hour window, and those sleeping under five hours accounted for the overwhelming majority of the adverse associations. It is observational and cannot establish causation, but it reinforces the six to eight hour message with objective data.

If you only have time for one paper this week, make it the target trial emulation of deceased donor transplantation versus dialysis in The BMJ [1]. It directly reshapes how you counsel an older patient offered a marginal kidney, and that conversation happens in general medicine clinics far more often than in transplant centres.

Here are the key takeaways from this week in General Medicine. First, standard criteria kidney transplants beat dialysis at every age, but for expanded criteria kidneys donated after circulatory death in the elderly, the survival advantage may not exist and early mortality is higher. Second, intrauterine local anaesthetic meaningfully reduces pain at intrauterine device placement in nulliparous women, and that matters for contraceptive uptake. Third, eight-weekly injectable cabotegravir and rilpivirine outperformed daily oral therapy in African adolescents with suppressed HIV. Fourth, when a primary care patient presents with unexpected weight loss, interpret their bloods against age, sex, and their own prior values rather than a single laboratory cut-off. And fifth, objective accelerometer data reinforce six to eight hours of regular sleep, with sleep under five hours carrying the broadest risk.

That's your roundup for This Week in General Medicine. 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 think of one colleague — in any specialty — who never has time to keep up with the literature. Tell them about AudioScholar: a free ten-minute weekly for every specialty, to listen to in any podcast app, or to read at audioscholar dot C C.

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