Abstract
OBJECTIVE: Cardiac microvascular obstruction (CMVO) remains a severe complication in non-ST elevation myocardial infarction (NSTEMI) patients with reperfusion therapy. We aimed at developing and validating the nomogram to predict the possibility of CMVO after primary percutaneous coronary intervention (PCI) by integrating clinical and laboratory-based information. METHODS: A total of 325 patients undergoing primary PCI for NSTEMI were recruited and divided into the training cohort (n=226) and the validating cohort (n = 99). The development of the nomogram was based on independent predictors of CMVO, and these variables were selected by multivariable logistic regression analysis. RESULTS: Independent predictors contained in nomogram were identified by multivariable logistic regression analysis, and these independent predictors included neutrophils (OR 1.166, 95% CI 1.044-1.303, P<0.01), hemoglobin (OR 1.037, 95% CI 1.013-1.062, P<0.01), triglyceride (OR 1.343, 95% CI 1.059; 1.704, P=0.015), Killip grade (OR 2.190, 95% CI 1.065-4.503, P=0.033), high thrombus load (OR 3.146, 95% CI 1.424-6.952, P<0.01), no-reflow (OR 3.142, 95% CI 1.419-6.955, P<0.01) and ischemic postconditioning (OR 0.445, 95% CI 0.209-0.944, P=0.035). The nomogram accurately predicted the presentation of CMVO in both the training set and validating set (AUC, 0.835 and 0.881, respectively). The results predicted by nomogram were confirmed to be highly consistent with the results of DE-CMR, both the training and validating cohorts, by Calibration plot and Hosmer-Lemeshow test. Decision curve analysis (DCA) also suggested that the nomogram was applicable in the clinic. CONCLUSION: The nomogram showed good performance in predicting CMVO, and it could help clinicians optimize the clinical treatments to improve the prognosis of NSTEMI patients.