The novel cystatin C, lactate, interleukin-6, and N-terminal pro-B-type natriuretic peptide (CLIP)-based mortality risk score in cardiogenic shock after acute myocardial infarction

Abstract

Abstract Background  Cardiogenic shock (CS) complicating acute myocardial infarction (AMI) still reaches excessively high mortality rates. This analysis is aimed to develop a new easily applicable biomarker-based risk score. Methods and results  A biomarker-based risk score for 30-day mortality was developed from 458 patients with CS complicating AMI included in the randomized CULPRIT-SHOCK trial. The selection of relevant predictors and the coefficient estimation for the prognostic model were performed by a penalized multivariate logistic regression analysis. Validation was performed internally, internally externally as well as externally in 163 patients with CS included in the randomized IABP-SHOCK II trial. Blood samples were obtained at randomization. The two trials are registered with ClinicalTrials.gov (NCT01927549 and NCT00491036), are closed to new participants, and follow-up is completed. Out of 58 candidate variables, the four strongest predictors for 30-day mortality were included in the CLIP score (cystatin C, lactate, interleukin-6, and N-terminal pro-B-type natriuretic peptide). The score was well calibrated and yielded high c-statistics of 0.82 [95% confidence interval (CI) 0.78–0.86] in internal validation, 0.82 (95% CI 0.75–0.89) in internal-external (temporal) validation, and 0.73 (95% CI 0.65–0.81) in external validation. Notably, it outperformed the Simplified Acute Physiology Score II and IABP-SHOCK II risk score in prognostication (0.83 vs 0.62; P < 0.001 and 0.83 vs. 0.76; P = 0.03, respectively). Conclusions  A biomarker-only score for 30-day mortality risk stratification in infarct-related CS was developed, extensively validated and calibrated in a prospective cohort of contemporary patients with CS after AMI. The CLIP score outperformed other clinical scores and may be useful as an early decision tool in CS.