The authors reported that 1,978 of 4,737 patients with first ST-segment elevation myocardial infarction (STEMI) developed heart failure (HF) during the index hospitalization, corresponding to an incidence of 41.8%. The finding emerges from the Henan STEMI registry, a multicentre prospective observational study in central China. The registry excluded patients with prior myocardial infarction or prior HF, so these rates apply to first-ever STEMI presentations without known antecedent HF. The source did not provide details in the abstract about timing of HF onset relative to reperfusion or specific in-hospital therapies administered to patients who developed HF.
Median age in the analytic cohort was 62.8 years (interquartile range 52.4–70.8). Women comprised 1,158 patients, or 24.4% of the cohort. Investigators classified HF phenotypes using the last in-hospital echocardiogram left ventricular ejection fraction (LVEF). Among the 1,978 patients with in-hospital HF, 268 (13.5%) were categorized as HF with reduced ejection fraction (HFrEF, LVEF ≤40%), 309 (15.6%) as HF with mildly reduced ejection fraction (HFmrEF, LVEF 41–49%), and 707 (35.7%) as HF with preserved ejection fraction (HFpEF, LVEF ≥50%). A substantial subgroup—694 patients (35.1% of HF cases)—had missing LVEF on the last in-hospital echocardiogram and were analyzed separately. The abstract did not specify reasons for missing LVEF data or whether missingness reflected logistical barriers, severity of illness, or incomplete echocardiography documentation.
Follow-up at one year was completed in 3,967 patients, representing 83.7% of the initial cohort. The one-year mortality among patients who developed in-hospital HF was 23.3% (391/1,680 reported in the survival analysis), compared with 5.8% (132/2,287) among patients who did not develop HF during hospitalization; the Kaplan-Meier comparison yielded log-rank P <0.001. Multivariable Cox frailty modeling, with the non-HF group as reference, identified in-hospital HF following myocardial infarction as an independent risk factor for one-year mortality (hazard ratio 2.13, 95% confidence interval 1.70–2.67). The abstract does not enumerate which covariates were included in the multivariable adjustment beyond this statement about model type.
Absolute one-year mortality varied substantially by HF subtype. The rates reported in the abstract were: HFrEF 29.6% (67/226), HFmrEF 20.8% (53/255), HFpEF 11.7% (72/615), and HF with missing LVEF 34.1% (199/584). In adjusted analyses, HFpEF carried the lowest adjusted risk among the LVEF-classified HF groups (hazard ratio 1.59, 95% CI: 1.16–2.17, versus the non-HF reference). The highest adjusted risk was observed in the HF group with missing LVEF (hazard ratio 2.49, 95% CI: 1.91–3.26). The authors reported these subgroup hazard ratios from multivariable Cox frailty models; the abstract did not list individual model covariates or present interactions between HF subtype and other clinical variables.
Survival estimates were generated with Kaplan-Meier analysis and group differences tested by the log-rank statistic. The study team used multivariate Cox frailty models to assess associations between HF status or HF LVEF subgroups and one-year mortality, using the non-HF group as the reference category. The use of frailty models indicates adjustment for clustering at some level (for example, by center), although the abstract does not state which clustering variable was specified. The abstract reports point estimates and 95% confidence intervals for hazard ratios but does not provide full model covariate lists, proportional hazards diagnostics, or handling strategies for missing data beyond the separate classification of cases with missing LVEF.
This observational registry analysis documents a high incidence of in-hospital HF after first STEMI and a markedly elevated one-year mortality among those patients. The source abstract notes the substantial proportion of HF cases with missing LVEF and reports the worst outcomes in that group; however, it did not explain the reason for the missing echocardiographic data. The abstract does not provide granular information on acute management (timing and type of reperfusion, guideline-directed medications during hospitalization, device therapies), which limits causal inference about drivers of HF incidence and subsequent mortality. Other unanswered items in the abstract include the specific covariates used in multivariable adjustment, how centers were accounted for in frailty models, or whether cause-specific mortality (cardiovascular versus noncardiovascular) was adjudicated. These gaps identify areas where the full text would need to be consulted to inform clinical interpretation and potential changes in in-hospital practice for first STEMI patients at risk of HF.