Background and Purpose- We compared clinical and radiological predictors of long-term mortality in patients with spontaneous intracerebral hemorrhage (ICH) needing intensive care. Methods- A retrospective multicenter study of adult ICH patients treated in Finnish tertiary hospital's intensive care units during 2003 to 2013. We created 3 multivariable models (clinical, radiological, and combined clinical-radiological) for 12-month mortality prediction and compared their areas under receiver operating characteristic curves (AUCs). We analyzed supratentorial and infratentorial ICHs separately. Results- Of 972 patients (796 supratentorial ICH, 176 infratentorial ICH) included, 43% died within 12 months (42% supratentorial ICH, 49% infratentorial ICH). For all patients, the clinical model (AUC, 0.83; 95% CI, 0.81-0.86) outperformed the radiological model (AUC, 0.73; 95% CI, 0.70-0.77; P<0.001), yet the combined model (AUC, 0.85; 95% CI, 0.83-0.88) outperformed both condensed models (P<0.001). For supratentorial ICH, the combined model outperformed both the clinical and radiological models (AUC, 0.84; 95% CI, 0.81-0.87 versus AUC, 0.82; 95% CI, 0.79-0.85 and AUC, 0.73; 95% CI, 0.69-0.77; P<0.001 for all). For infratentorial ICH patients, the combined model significantly outperformed the radiological model but not the clinical model (AUC, 0.92; 95% CI, 0.88-0.96 versus AUC, 0.76; 95% CI, 0.69-0.83 versus AUC, 0.91; 95% CI, 0.87-0.95; P<0.001 and P=0.433, respectively). Conclusions- Clinical factors were more important than objective radiological factors for 12-month mortality prediction in intensive care unit-treated ICH patients. The effect of clinical and radiological factors on outcome was different for supratentorial and infratentorial ICHs stressing that these should not be treated as one entity.
Keywords: cerebral hemorrhage; critical care; prognosis; risk factors.