← All articles
A plain ceramic cup and a folded linen cloth on a wooden table in warm evening light

Early TB treatment and an esophageal cancer drug — plus 2 stories

ResearchDigest · 4 storiesMedicine and treatment

4 stories this week: severe pneumonia in infants with HIV, esophageal cancer after a first treatment line, blood pressure and memory in type 2 diabetes, and a heart-rhythm ablation. Each note names who was studied and where the result stops.

Lonevi

In infants with HIV and severe pneumonia, tuberculosis treatment from admission did not lower the share who died

A folded cotton blanket on a wooden chair beside a sunlit hospital window

On 14 November 2026 The Lancet published the tuberculosis comparison from EMPIRICAL, an open-label factorial superiority trial in 19 hospitals in Cote d'Ivoire, Malawi, Mozambique, Uganda, Zambia and Zimbabwe. Infants aged 28 to 365 days were admitted with severe HIV-associated pneumonia and had no current tuberculosis diagnosis and no known contact. The trial is registered as NCT03915366 and is complete.

Investigators enrolled 563 infants and analysed 558. The comparison was 276 infants who started tuberculosis treatment on top of standard pneumonia care against 282 who did not start it on admission. Median age was 4.4 months, and 49% were girls. In the comparison arm, 130 of 282 infants (46%) later started tuberculosis treatment anyway, at a median of 4.0 days: 55 (20%) after a microbiological result and 75 (27%) on clinical grounds. The authors write that this later start may have narrowed the gap between the two strategies.

By 12 months, 132 of 276 infants (48%) in the early-treatment groups had died, and 121 of 282 (43%) in the comparison groups. The adjusted hazard ratio for death was 1.12 (95% CI 0.87 to 1.43, p=0.3868). At day 15 the odds ratio was 1.24 (0.83 to 1.84). Serious adverse events did not differ, and the investigators reported no treatment-related deaths. The interval crosses 1, so the trial shows neither a survival gain nor a rise in deaths. A separate valganciclovir comparison from the same trial is a different result and is not reported here.

After esophageal cancer grew through immunotherapy, an antibody-drug conjugate extended time without growth and time until death

An unmarked amber glass bottle with a black cap on a dark stone bench

On 12 November 2026 JAMA published PANKU-Esophagus01, a phase 3 open-label trial at 80 hospitals in China (NCT06304974). Adults had recurrent or metastatic esophageal squamous-cell carcinoma that had progressed after a first line of a PD-1 or PD-L1 inhibitor plus platinum chemotherapy. This report is an interim analysis, not a regulatory approval. Izalontamab brengitecan is a bispecific antibody-drug conjugate aimed at EGFR and HER3.

Of 704 people screened, 497 were randomized and 493 received treatment: 249 received izalontamab brengitecan and 248 received a physician-chosen chemotherapy (irinotecan, paclitaxel or docetaxel). Median age was 61 years (range 34 to 76), and 448 of 497 (90.1%) were men.

Median progression-free survival, by blinded central review, was 4.17 months (95% CI 3.61 to 4.53) with the new drug and 1.97 months (1.61 to 2.73) with chemotherapy, hazard ratio 0.50 (0.40 to 0.63). Median overall survival was 9.79 months (7.89 to 11.01) versus 7.20 months (6.18 to 7.82), hazard ratio 0.63 (0.48 to 0.82). Grade 3 or higher treatment-related adverse events occurred in 212 of 249 patients (85.1%) on the new drug and 147 of 248 (59.3%) on chemotherapy, mostly blood-count effects. The result belongs to this second-line population in the trial. It does not describe people who have not already received immunotherapy plus platinum, and it is not an approval.

A lower blood-pressure target did not change dementia or mild cognitive impairment in type 2 diabetes

A plain navy fabric band coiled on a pale wooden table

On 12 November 2026 The BMJ published prespecified cognitive outcomes from BPROAD, an open-label trial with masked outcome assessors at 145 centres in seven regions of China (NCT03808311). Adults aged 50 or older had type 2 diabetes, raised systolic pressure and higher cardiovascular risk. The parent trial included 12,821 people. This analysis included 12,434 without probable dementia at the start. Intensive treatment aimed at systolic pressure below 120 mm Hg. Standard treatment aimed below 140. Treatment could continue for up to five years.

Mean age was 63.7 years, and 5,567 participants (44.8%) were women. The assigned groups were 6,219 and 6,215. After one year, mean systolic pressure was 121.6 mm Hg with intensive treatment and 133.2 mm Hg with standard treatment. Over a median of 3.8 years, dementia or mild cognitive impairment occurred in 747 people (3.69 per 100 person-years) and in 783 people (3.89 per 100 person-years). The hazard ratio was 0.96 (95% CI 0.86 to 1.06, P=0.38).

In an MRI subset with results for 428 people on intensive treatment and 434 on standard treatment, white-matter lesion volume rose less with the lower target. The between-group difference was −0.09 asinh-transformed cubic centimetres (95% CI −0.15 to −0.02, P=0.01). That imaging difference is not a difference in dementia cases. The trial does not show a change in dementia or mild cognitive impairment in this population.

Adding low-voltage ablation to pulmonary-vein isolation did not cut rhythm recurrences overall

A thin catheter coil on a blue sterile drape under operating-room light

On 8 November 2026 JAMA published COAST-AF, a single-blind trial at 14 ablation centres in Canada (NCT03347227). Adults had persistent atrial fibrillation for more than three months. Enrolment ran from August 2018 to September 2024, with follow-up through April 2026. Of 508 patients, 253 were assigned to pulmonary-vein isolation plus ablation of low-voltage atrial areas, and 255 to isolation alone. Mean age was 65.2 years, 386 (76%) were men, and 268 (54%) had low-voltage areas.

The primary outcome was an atrial arrhythmia returning more than 90 days after a single procedure. After a mean follow-up of 21 months, that outcome occurred in 124 patients (49.4%) assigned to the extra ablation and in 146 patients (57.5%) assigned to isolation alone. The hazard ratio was 0.82 (95% CI 0.65 to 1.04, P=0.11). The overall comparison is not statistically significant. The abstract pairs those counts and percentages with the randomized groups. The primary analysis is described as modified intention-to-treat among patients who received the procedure.

A prespecified subgroup of 268 patients who did have low-voltage areas showed a different pattern: recurrence in 68 of 138 (49.3%) with the extra ablation versus 90 of 130 (69.2%) with isolation alone, hazard ratio 0.77 (0.66 to 0.91, P=0.002). Among patients without those areas, recurrence was similar, 56 of 113 (49.6%) versus 51 of 113 (45.1%), hazard ratio 1.09 (0.90 to 1.32). Procedures were longer when the extra ablation was added, and safety outcomes were similar. A subgroup result does not replace the overall finding. For persistent atrial fibrillation as a whole, the trial does not show a clear drop in recurrence.

This article reports published trials. It does not replace a consultation with a doctor, and it is not a recommendation to start, stop or change treatment.

Sources

Articles in this section are educational and are not medical advice, a diagnosis, or a prescription. Consult a qualified professional before acting on anything you read here.

Read next