Statistical analysis
Distributions of baseline characteristics were compared among five eGFR groups, including sex, age, BMI, smoking, comorbidity, clinical characteristics, medication use prior to admission, and the 1-month, and 1-year mortality rates. Data of continuous variables were expressed as median and 25th percentile to 75th percentile and examined using the Kruskal-Wallis test for continuous variables that were not normally distributed. Categorical variables were expressed as frequencies and examined using the Chi-square test.
The cumulative incidence curves of mortality were computed for the 5 subgroups of eGFR levels by the Kaplan–Meier method and tested using the log-rank test. The 1-month and 1-year mortality rates (per 1000 person-days) after intracerebral hemorrhagic were calculated by the eGFR levels. We further used the Cox proportional hazards regression analysis to estimate the hazard ratio (HR) of mortality and 95% confidence interval (CI) associated with eGFR levels at admission for patients with intracerebral hemorrhage, using patients with eGFR ≥ 90mL/min/1.73 m2 as references. We also used the receiver operating characteristic (ROC) curve to assess the predictive performance of eGFR levels with regard to one-month mortality and one-year mortality after intracerebral hemorrhage. All statistical analyses were performed using SAS statistical software, version 9.4 (SAS Institute Inc., Cary, NC). Statistical significance was defined using two‐tailed tests (P < 0.05).