Study population and design. This was a retrospective, single-center, observational study. Patients who were admitted to the neurosurgical ICU in a tertiary referral hospital (Samsung Medical Center, Seoul, Republic of Korea) from January 2013 to December 2019 were eligible. This study was approved by the Institutional Review Board of Samsung Medical Center (approval number: SMC 2020-09-082). The requirement for informed consent was waived by the Institutional Review Board of Samsung Medical Center due to its retrospective nature. Included criteria were: (1) patients who were hospitalized in the neurosurgical ICU due to postoperative management or neurocritical illness, and (2) those whose serum cTnI levels were obtained within seven days after ICU admission. Exclusion criteria were: (1) those with insufficient medical records, (2) those who had ‘do not resuscitation’ order, (3) those who were admitted to departments other than neurosurgery, and (4) those who were transferred to other hospitals or with unknown prognoses (Fig. 1).
Definitions and endpoints. In this study, baseline characteristics such as comorbidities, ICU management, and laboratory data were collected retrospectively using Clinical Data Warehouse. Our center constructed the “Clinical Data Warehouse Darwin-C” designed for investigators to search and retrieve de-identified medical records from electronic archives. It contains data pertaining to more than four million patients. Clinical and laboratory data were extracted from the Clinical Data Warehouse Darwin-C after finalizing the patient list in this study. Risk of surgery was defined according to the 2014 European Society of Cardiology/European Society of Anesthesiology (ESC/ESA) guidelines14. Perioperative management of patients followed institutional protocols based on current guidelines6,14. According to the institutional guideline, perioperative cTnI was measured for patients with more than moderate risk or undergoing moderate- to high-risk surgeries14. It was also measured at the discretion of attending clinician for patients with mild risks6,14. An automated analyzer (Advia Centaur XP; Siemens Healthcare Diagnostics, Erlangen, Germany) with a highly sensitive immunoassay was used for cTnI measurement. The lowest limit of detection was 6 ng/L. In this study, cTnI elevation was defined as an increase in cTnI above 0.0 6µg/L within 7 days after ICU admission15. Acute Physiology and Chronic Health Evaluation (APACHE) II score was calculated based on the worst value recorded during the initial 24 h in the ICU admission16,17. If the patient was intubated, the verbal score of Glasgow Coma Scale (GCS) was estimated using eye and motor scores as reported previously18. MACEs were defined as non-fatal cardiac arrest, emergent coronary revascularization, acute coronary syndrome, stroke, congestive heart failure, atrial fibrillation (new onset or destabilization of pre-existing atrial fibrillation), major arrhythmia, cardiovascular death, and rehospitalization for cardiovascular reasons1. The primary endpoint was in-hospital mortality and the secondary outcome was MACE.
Statistical analyses. All data are presented as means ± standard deviations for continuous variables and frequencies and proportions for categorical variables. Data were compared using Student’s t-test for continuous variables and Chi-square test or Fisher’s exact test for categorical variables. Propensity score matching was used to control the selection bias and the confounding factor detected in this observational study. Each patient with cTnI elevation was matched to one control patient with the nearest neighbor matching within calipers determined by the propensity score. A caliper width of 0.2 of the standard deviation of the logit of the propensity score was used for the matching19. To determine the effectiveness of propensity score matching for controlling the differences between patients with and without cTnI elevation, standardized mean differences (SMDs) were calculated for each variable before and after matching. SMDs less than 10% indicated successful propensity scores matching and balancing between the two groups. To evaluate whether there is a difference in in-hospital mortality and MACEs according to the cTnI elevation, we performed multiple logistic regression with stepwise variable selection in the overall and matched population. In the overall population, we tried to obtain the result of correcting confounding through regression adjustment, and in the matching dataset, we perform doubly robust estimation to additionally correct the bias that still exits after propensity score matching. The variables included in the multiple analyses were age, sex, comorbidities, cause of ICU admission, utilization of organ support modalities, including mechanical ventilators, continuous renal replacement therapy and vasopressors, ICP monitoring devices, hyperosmolar therapy, GCS, and APACHE II score on ICU admission. Cumulative mortality was calculated by Kaplan–Meier estimate and compared using a log-rank test. All tests were two-sided and p values less than 0.05 were considered statistically significant. All statistical analyses were performed with R Statistical Software (version 4.0.2; R Foundation for Statistical Computing, Vienna, Austria).