Head-to-Head Comparison of Clinical Scores for Predicting Stroke Associated Pneumonia after Intracerebral Hemorrhage

Abstract Background: Despite advances in medical knowledge, treatment for ICH remains strictly supportive with not many evidence-based interventions currently available. Stroke associated pneumonia (SAP) is a common medical complication after stroke and has significant impact on stroke outcomes. In the study, we aimed to systematically compare discrimination and calibration of clinical scores with regard to in-hospital SAP after ICH. Methods: The validation cohort was derived from the Beijing Registration of Intracerebral Hemorrhage. The SAP was diagnosed according to the criteria for hospital-acquired pneumonia of Center for Disease Control and Prevention. Five clinical scores were included in the study: the ICH-APS-A, ICH-ASP-B, ISAN, ACCD4 and PASS score. Discrimination was assessed by calculating the area under the receiver operating characteristic curve (AUROC). Pairwise AUROC was compared by using Delong’s method. Calibration was assessed by performing the Hosmer-Lemeshow goodness-of-fit test and plot of observed versus predicted risk according to 10 deciles of the predicted risk. The Cox and Snell R-square and Nagelkerke R-square of the Hosmer-Lemeshow goodness-of-fit test were calculated. Results: A total number of 1964 patients were enrolled. The mean age was 56.8±14.4 and 67.6% were male. The median admission NIHSS was 11 (IQR: 3-21). The median length of stay (LOS) was 16 days (IQR: 8-22). A total number of 575 (29.2%) patients was diagnosed with in-hospital SAP after ICH. The AUROC of five clinical scores ranged from 0.732 to 0.800. In pairwise comparison, the ICH-APS-B (0.800, 95% CI=0.780-0.820, P<0.001) showed statistically better discrimination than other risk models (all P<0.001). All clinical scores performed better among patients with LOS longer than 72 hours. The ICH-APS-B (0.827, 95% CI=0.806-0.848, P<0.001) still showed statistically better discrimination than other risk models in patients with LOS longer than 72 hours (all P<0.001). The ICH-SAP-B had the largest Cox and Snell R-square of the Hosmer-Lemeshow test for in-hospital SAP after ICH.Conclusion: Several models are externally validated to be effective for risk stratification and outcome prediction for in-hospital SAP after ICH, which would be useful tools for personalized care and clinical trials in prevention of SAP after ICH.