Heart care's new question is when, not whether. The evidence is thinner than the habit

What changed

Three preprints posted this week, plus a contemporary stroke cohort study, all point at the same shift in heart and stroke care: the decision that matters most is increasingly about timing and about which patient, not about whether to intervene at all. The clearest example is a new heart transplant risk framework, RIGHT-HTx, a three-domain scoring system for early mortality risk after heart transplantation. Its discrimination, a C-index of 0.732, outperforms renal dysfunction alone in predicting one-year mortality after transplant, in a single-centre cohort of 436 patients. m:…5126

But the same paper shows why a good discrimination score is not enough. Temporal validation of RIGHT-HTx showed comparable discrimination, a C-index of 0.738, but poor absolute-risk calibration: 35.4 percent observed mortality against 25.4 percent predicted. A model that ranks patients well while getting the absolute risk wrong by ten points is a model that can mislead a transplant programme about how many beds, organs and follow-up slots it needs. The authors are candid that the framework needs independent multicentre validation before clinical use, and the calibration gap is the reason. m:…5126

The framework also flags a specific high-risk combination: elevated transpulmonary gradient together with a donor-to-recipient weight ratio below 1.0 is independently associated with mortality, with an adjusted hazard ratio of 2.87. That is the kind of concrete pairing a transplant team can actually act on, and it is more useful than the headline score on its own. m:…5126

Earlier is not always better

The valve story makes the same point from the opposite side. A meta-analysis of early transcatheter aortic valve replacement against surveillance in moderate aortic stenosis, drawing on two randomised trials and seven observational cohorts, finds that early TAVR does not detectably change two-year mortality, though it reduces heart failure hospitalisation. It is also associated with a higher rate of new pacemaker implantation compared with clinical surveillance. Fewer hospital stays, no survival gain, and a new device in the chest: that is a trade-off, not a triumph. m:…3409

The meta-analysis adds a warning that should travel well beyond cardiology. Observational cohorts significantly overestimate the mortality reduction of early TAVR compared with randomised trial estimates. The observational literature made early intervention look like a survival win; the randomised evidence says it is not one, at least at two years. Anyone building a case for earlier intervention from registry data should read that sentence twice. m:…3409

A second meta-analysis, on door-to-valve time in acute decompensated severe aortic stenosis, reaches a similarly deflating conclusion. Delayed valve intervention is not significantly associated with long-term all-cause mortality, with a hazard ratio of 1.52 and a confidence interval running from 0.46 to 5.00. Crude short-term mortality is higher with delayed intervention, a relative risk of 4.96, but that association is not robust to sensitivity analyses, and the authors judge the existing observational data to be at serious risk of bias with substantial heterogeneity. The field does not yet have the evidence to define an optimal door-to-valve interval, and this week's paper says so plainly. m:…8098

After the stroke, a different disease

A contemporary cohort study of 1,344 stroke patients treated between 2017 and 2020 extends the timing argument into follow-up care. Cardiac disease is the leading cause of death among ischemic stroke survivors at 31.6 percent, with sudden cardiac arrest accounting for 10.9 percent of deaths, while cancer is the leading cause of death among hemorrhagic stroke survivors at 38.9 percent. Overall survival does not differ significantly between the subtypes, a median of 5.5 versus 5.8 years, so the difference is not how long people live but what they die of. m:…5253

The timing of the risk also flips after discharge. Cerebrovascular causes account for 36.0 percent of deaths during the index stroke hospitalisation but only 15.5 percent of post-stroke discharge deaths. The stroke itself kills in hospital; after discharge, the heart and cancer take over. A follow-up programme built around preventing another stroke is watching the wrong risk for most of the patient's remaining years, and the authors' case for subtype-specific longitudinal surveillance follows directly from the numbers. m:…5253