The Rhythm | Issue 25
Issue 25 | Week of 21 September 2026
Cardiology Intelligence Briefing: knowing where the evidence ends
No flagship cardiology congress fell this week, so the evidence came from journals, and most of it was a second look at established treatments: pooled trials, subgroup analyses and secondary analyses. That is no criticism. Second looks are how cardiology learns where an established result still applies, and where it may not.
Three of the papers push established approaches into new territory: into frailer patients,3 into patients with a history of cancer,4 and from stroke prevention to cognition.2 The certainty was uneven. One large randomised comparison gave a fairly clear answer: opening a narrowed carotid artery did not improve cognition beyond intensive medical therapy.2 The frailty and cancer analyses were harder to interpret, because they rest on subgroups the trials were not designed to compare.3,4
There is a difference between a null result and a silent one. A large randomised trial that finds no effect tells us something. An underpowered subgroup that finds no difference may simply be unable to answer the question, and an interaction test that finds none does not show that a treatment works the same way everywhere. The cancer analysis shows why the distinction matters: the benefit seen in the wider trial was not apparent in this small subgroup, while the safety findings moved in an unfavourable direction.4
The clearest positive signal is a familiar one. Even after recent neutral trials were added, the pooled randomised evidence continued to favour imaging-guided coronary stenting, including for mortality and myocardial infarction.1 Screening for atrial fibrillation (AF) offers a different warning.5 Finding more AF is not the same as preventing harm from it.5
The big picture: this week's lesson is less about extending indications than about being precise about where the evidence for existing ones ends.
