Association between coagulation disorder scores and in-hospital mortality in ARF patients: a retrospective analysis from the MIMIC-IV database.

ABSTRACT

Introduction: Acute respiratory failure (ARF) has a high mortality rate, and currently, there is no convenient risk predictor. The coagulation disorder score was proven to be a promising metric for predicting in-hospital mortality, but its role in ARF patients remains unknown. Methods: In this retrospective study, data were extracted from the Medical Information Mart for Intensive Care IV (MIMIC-IV) database. Patients diagnosed with ARF and hospitalized for more than 2 days at their first admission were included. The coagulation disorder score was defined based on the sepsis-induced coagulopathy score and was calculated by parameters, namely, additive platelet count (PLT), international normalized ratio (INR), and activated partial thromboplastin time (APTT), based on which the participants were divided into six groups. Results: Overall, 5,284 ARF patients were enrolled. The in-hospital mortality rate was 27.9%. High levels of additive platelet score, INR score, and APTT score were significantly associated with increased mortality in ARF patients (P < 0.001). Binary logistic regression analysis showed that a higher coagulation disorder score was significantly related to the increased risk of in-hospital mortality in ARF patients (Model 2 coagulation disorder score = 6 vs. coagulation disorder score = 0 OR, 95% CI 7.09, 4.07-12.34, P < 0.001). The AUC of the coagulation disorder score was 0.611 (P < 0.001), which was smaller than that of sequential organ failure assessment (SOFA) (De-long test P = 0.014) and simplified acute physiology score II (SAPS II) (De-long test P < 0.001) but larger than that of additive platelet count (De-long test P < 0.001), INR (De-long test P < 0.001), and APTT (De-long test P < 0.001), respectively. In subgroup analysis, we found that in-hospital mortality was markedly elevated with an increased coagulation disorder score in ARF patients. No significant interactions were observed in most subgroups. Of note, patients who did not administrate oral anticoagulant had a higher risk of in-hospital mortality than those who administrated oral anticoagulant (P for interaction = 0.024). Conclusion: This study found a significant positive association between coagulation disorder scores and in-hospital mortality. The coagulation disorder score was superior to the single indicators (additive platelet count, INR, or APTT) and inferior to SAPS II and SOFA for predicting in-hospital mortality in ARF patients.