Myocardial strain combined with clinical risk factors in the prediction of in-hospital heart failure among patients with ST-segment elevation myocardial infarction

H Wu,H Wang,X Wang,L Xu,J Wu
DOI: https://doi.org/10.1016/j.crad.2024.07.016
Abstract:Aims: To investigate the predictive value of myocardial strain derived from cardiac magnetic resonance (CMR) combined with clinical indicators for in-hospital heart failure (HF) in STEMI patients. Materials and methods: In all, 139 STEMI patients were included, with 28 in the heart failure group and 111 in the non-HF group, and clinical and laboratory data were collected. Left ventricular (LV) global radial strain (GRS), global longitudinal strain (GLS), global circumferential strain (GCS), left ventricular ejection fraction (LVEF), stroke volume (SV), and infarct size (IS) were assessed by CMR. Results: The HF group had worse GRS, GLS, GCS, LVEF, SV, larger IS, longer symptom to balloon time (SBT) and higher levels of high-sensitivity C-reactive protein (hs-CRP) and neutrophil percentage (N%) than the non-HF group (P<0.05). There was a strong correlation between GRS and LVEF (r=0.741, P<0.001). After adjustment for CMR and clinical risk factors, GRS<15.6%, LVEF<37.7%, SBT>350 min, hs-CRP>11.45 mg/L, and N%>74% were independently associated with HF. Clinical model (SBT>350 min + hs-CRP>11.45 mg/L + N%>74%) were associated with a lower diagnostic accuracy for predicting in-hospital HF than GRS + clinical co-model and LVEF + clinical co-model (P<0.05), respectively. There was no significant difference in the area under the curve (AUC) between GRS + clinical co-model and LVEF + clinical co-model (P=0.620): AUC for clinical model = 0.824, AUC for GRS + clinical co-model = 0.895, and AUC for LVEF + clinical co-model = 0.907. Conclusions: GRS may be effective in predicting in-hospital heart failure after STEMI compared to LVEF, a classical cardiac function parameter, and its combination with clinical risk factors, especially SBT, hs-CRP, and N%, may provide further evidence for early prognostic assessment.
What problem does this paper attempt to address?