An opinionated take on NEJM highlights for the fourth quarter of 2023

A Chinese study uncovers a new use for an old drug

Bleeding in the small intestine is often due to vascular malformations and is difficult to diagnose and manage because 1) the area is hard to reach with a scope and 2) bleeding tends to be intermittent.  Existing approaches are cumbersome and far from satisfactory, which makes a recent blinded RCT study conducted in China exploring thalidomide use in those patients valuable – not only did it show a significant decrease in bleeding episodes after a course of 4 months of thalidomide, but the effect was sustained well after treatment completion.  This is a significant advance for a frustrating medical problem. It is also a sign of the medical field in China coming into its own in therapeutic innovation. Thalidomide for Recurrent Bleeding Due to Small-Intestinal Angiodysplasia


BiTE therapy in solid tumors

Like the related CAR-T therapies, work on bispecific T-cell engagers (BiTE) has focused nearly entirely on hematological malignancies (surveyed here), with the peculiar exception of uveal melanoma, a rare cancer (with Tebentafusp-tebn/Kimmtrak Immunocore).  This makes a recent phase 2 study of tarlatamab (Amgen) for relapsed small cell lung cancer noteworthy. Though there was no control arm, observed survival was approximately 9 months, a meaningful improvement from typical clinical experience. Tarlatamab for Patients with Previously Treated Small-Cell Lung Cancer


Untangling Aβ in Alzheimer’s disease

So, the emerging story based on lecanemab and donanemab (and perhaps aducanumab) is that by administering an antibody that clears Abeta aggregates and “proto-aggregates”, and acting early in the disease, one can get a small but real clinical effect on Alzheimer’s progression. Unfortunately, this recipe seems to have failed for gantenerumab (Roche) with a couple phase 3 trials that did not meet their endpoints (CDR-SB score). What happened? I spent a few hours parsing the studies for the three agents and this is my analysis.

Comparing gantenerumab and lecanemab, enrollment numbers were similar (though the lecanemab study went 18 months vs. 2 years for gantenerumab); and the effect on amyloid plaque burden (measured by PET) were similar. But there was a difference: although the enrollment criteria were nominally similar, the lecanemab study enrolled patients earlier in their disease: baseline CDR-SB were ~3.1 (lecanemab) vs. ~3.7 (gantenerumab), and MMSE ~25.5 vs. ~23.5 – which probably corresponds to a difference of about 1 year of disease progression.

Comparing gantenerumab and donanemab, the level of impairment (or disease progression) at baseline were similar, but two things were different. First, the donanemab study was twice as large so could pick up smaller effects that would not have been statistically significant in the smaller gantenerumab studies. Second, donanemab is a stronger plaque remover, lowering plaque burden by about 87% vs 57% with gantenerumab.

This seems to confirm that the two factors that are thought to play an important role for therapeutic efficacy: earlier intervention, more plaque removal. Two Phase 3 Trials of Gantenerumab in Early Alzheimer’s Disease; What the Gantenerumab Trials Teach Us about Alzheimer’s Treatment


More on GLP-1

It was well covered in the press, but the SELECT trial of semaglutide (Ozempic/Wegovy, Novo) in an obese non-diabetic population (N=17,604) with preexisting cardiovascular disease (defined for the most part as a previous MI or stroke) is a watershed. Treatment resulted in a ~20% decrease in the cardiovascular event rate compared to placebo, from ~10% to ~8% over 4 years.  Many biomarkers of cardiovascular risk were improved in the treatment arm: blood pressure, inflammation (CRP), cholesterol (lower LDL, higher HDL). To put things in perspective, this effect is only slightly less than that seen for patients in with diabetes (25% risk reduction for semaglutide arm in the SUSTAIN-6 trial), but significantly less than the 40% reduction observed in diabetic patients after bariatric surgery (though these studies were not randomized).  Meanwhile, a similar trial with Lilly’s tirzepatide (Mounjaro) is expected to read out in 2027.

Approximately 90M US adults have a prescription for antihypertensive medication – hard to see how this is not going to be the similar for GLP-1 agonists.  Semaglutide and Cardiovascular Outcomes in Obesity without Diabetes


Screening genomes at the population level

With the growing incidence of genome / exome sequencing ranging from large population studies to individual-driven private sequencing, the question of what findings should be reported to participants can be a thorny one.  The American College of Medical Genetics and Genomics (ACGM) maintains an annually updated list of variants (current list is here) that they believe should be reported based on principles described here, chief among them (1) likelihood of significant morbidity is high (2) something can done about it (note that ApoE4 is not in the list at this time).  But this list is in large part the result of expert opinions rather than systematic data.  This is where Iceland comes in, the first country to have conducted a large-scale genome project (by deCODE Genetics – now Amgen) more than a decade ago. By looking back at those actionable variants on the ACGM list they were able to determine two main things: First, ~4% of the population carried an actionable variant. Second, the lifespan reduction linked to carrying an actionable variant is ~1 year (which is potentially is the order of the benefit possible through intervention). Obviously, this is not uniform across variants and in particular, cancer promoting variants have a much stronger effects than others. Also, death is a crude metric and may not reflect the full burden of a disability linked to a variant. And finally, the Icelandic population is sui generis; we will learn more as the UK Biobank or the All of Us studies mature.  Actionable Genotypes and Their Association with Life Span in Iceland

Meanwhile, genetic prenatal screening, which until now has been focused on chromosomal abnormalities, is increasingly able to identify fetal single base-pair mutations, including de novo mutations, by testing blood sample from the pregnant mother. A couple letters to the editor describe the workflow to achieve that (which is non-trivial since there is strong background of maternal DNA).   High-Resolution and Noninvasive Fetal Exome Screening; Comprehensive Noninvasive Fetal Screening by Deep Trio-Exome Sequencing

What does this all mean? When would one terminate a pregnancy for a disabling mutation? What level of intervention is needed for an incidental finding? A lot to sort out: the profession of genetic counselor has bright days ahead of it (that is, unless AI takes over).


The New England Journal of Medicine is a premier weekly medical journal covering many topics of interest to the health sector. In this series we offer an opinionated perspective on selected highlights that might be of interest to our clients and others.

We use cookies
This website collects cookies to deliver better user experience and to analyze our website traffic and performance; we never collect any personal data or target you with ads.