The Predictive Role of Systemic Inflammation Response Index in the Prognosis of Traumatic Brain Injury: A Propensity Score Matching Study.

Baojie Mao,Lei Feng,Dongdong Lin,Yanfei Shen,Jiangchun Ma,Yuning Lu,Rui Zhang,Ming Wang,Shu Wan
DOI: https://doi.org/10.3389/fneur.2022.995925
IF: 3.4
2022-01-01
Frontiers in Neurology
Abstract:BackgroundWe aimed to evaluate the predictive power of systemic inflammation response index (SIRI), a novel biomarker, to predict all-cause mortality in patients with traumatic brain injury (TBI) in the intensive care unit (ICU).MethodsClinical data were retrieved from the Medical Information Mart for Intensive Care-IV (MIMIC-IV) database. Kaplan-Meier (KM) methods and cox proportional hazard models were performed to examine the association between SIRI and all-cause mortality. The predictive power of SIRI was evaluated compared to other leukocyte-related indexes including neutrophils, lymphocytes, monocytes and white blood cells (WBC) by the Receiver Operating Characteristic (ROC)curve for 30-day mortality. In addition, propensity score matching (PSM) was conducted to reduce confounding.ResultsA total of 350 TBI patients were enrolled overall in our study. The optimal cutoff point of SIRI was determined at 11.24 × 109/L. After 1:1 PSM, 66 matched pairs (132 patients) were generated. During the 30-day, in-hospital and 365-day follow-up periods, patients with low SIRI level were associated with improved survival (p < 0.05) compared with patients with high SIRI level. Cox regression analysis identified that higher SIRI values was an independent risk factor for all-cause mortality and results were stable on multiple subgroup analyses. Furthermore, ROC analysis indicated that the area under the curve of SIRI [0.6658 (95% Confidence Interval, 0.5630–0.7687)] was greater than that of neutrophils, monocytes, lymphocytes and WBC. The above results were also observed in the matched cohort.ConclusionIt was suggested that TBI patients with high SIRI level would suffer from a high risk of 30-day, in-hospital and 365-day mortality. SIRI is a promising inflammatory biomarker for predicting TBI patients' prognosis with relatively better predictive power than other single indicators related to peripheral differential leukocyte counts.
What problem does this paper attempt to address?