Abstract
INTRODUCTION: In recent years, gamma-glutamyl transpeptidase to platelet ratio (GPR) has been proposed as a new inflammatory marker. We aimed to evaluate the association between GPR and outcomes after cardiac arrest (CA). METHODS: A total of 354 consecutive patients with CA were included in this retrospective study. Patients were divided into three groups according to tertiles of GPR (low, n = 119; middle, n = 117; and high, n = 118). To determine the relationship between GPR and prognosis, a logistic regression analysis was performed. The ability of GPR to predict the outcomes was evaluated by receiver operating characteristic (ROC) curve analysis. Two prediction models were established, and the likelihood ratio test (LRT) and the Akaike Information Criterion (AIC) were utilized for model comparison. RESULTS: Among the 354 patients (age 62 [52, 74], 254/354 male) who were finally included in the analysis, those in the high GPR group had poor outcomes. Multivariate logistic regression analysis revealed that GPR was independently associated with the three outcomes, for ICU mortality (odds ratios (OR) = 1.738, 95% confidence interval (CI): 1.221-2.474, P = 0.002), hospital mortality (OR = 1.676[1.164 - 2.413], P = 0.005), and unfavorable neurologic outcomes (OR = 1.623[1.121 - 2.351], P = 0.010). The area under the ROC curve was 0.611 (95% Cl: 0.558-0.662) for ICU mortality, 0.600 (95% CI: 0.547-0.651) for hospital mortality, and 0.602 (95% CI: 0.549-0.653) for unfavorable neurologic outcomes. Further, the LRT analysis showed that compared with the model without GPR, the GPR-combined model had a higher likelihood ratio χ (2) score and smaller AIC. CONCLUSION: GPR, as an inflammatory indicator, was independently associated with outcomes after CA. GPR is helpful in estimating the clinical outcomes of patients with CA.