Design: Post hoc subgroup analysis of a randomised trial (TARGET)
Population: 1,216 patients with 30-90% coronary stenosis on CT angiography
Primary outcome: Angiography without obstructive disease, or no intervention despite obstructive disease, within 90 days
Effect: Two-year major adverse cardiovascular events: adjusted hazard ratio 0.48 (95% CI 0.27-0.87) in women, 0.89 (0.59-1.33) in men; no significant sex-by-treatment interaction

The TARGET trial randomised 1,216 patients with 30-90% stenosis on coronary computed tomography angiography to on-site CT-derived fractional flow reserve (CT-FFR) guidance or standard care. This post hoc analysis split the result by sex, on the long-standing observation that women reach the catheter laboratory with less obstructive disease and worse outcomes than men.

The two effects separated. In men, CT-FFR guidance did what it is meant to do — fewer angiograms that found no obstructive disease, and more early revascularisation. In women, it did not change either. Yet over two years the women managed with CT-FFR had fewer major adverse cardiovascular events (adjusted hazard ratio 0.48, 95% CI 0.27-0.87) while the men did not (0.89, 0.59-1.33).

The authors do not overclaim it, and neither should a reader: the interaction between sex and strategy was not statistically significant, which is the test that decides whether two subgroup results genuinely differ. Read as a reason to keep asking the question in trials designed to answer it, not as a reason to triage access to CT-FFR by sex.

Do not change who is offered CT-FFR on the basis of this analysis
Note the mechanism gap: the outcome benefit in women was not explained by fewer angiograms or more revascularisation
Expect this question to recur in prospectively powered analyses
Keep recording sex-disaggregated outcomes in local CT-FFR audit

Why it matters

It raises the possibility that a diagnostic strategy helps two groups through different mechanisms — which trials powered on the whole cohort would never detect.

The statistics, in plain English

Two subgroups can show different hazard ratios by chance alone, which is why the interaction test matters more than either estimate. Here it was not significant, so the honest statement is that the trial cannot distinguish the two effects. The women's hazard ratio also came with a wide interval (0.27-0.87) on a small number of events, so even taken at face value its size is uncertain.