The burden of non-symptomatic cerebral ischemia on MRI and its effect on clinical outcomes in patients with first-ever intracerebral hemorrhage.

Purpose: To compare the burden of non-symptomatic cerebral ischemia (NSCI) detected on magnetic resonance imaging (MRI) and computed tomography (CT), and assess the association of MRI-NSCI with clinical outcomes among patients with first-ever intracerebral hemorrhage (ICH).

Methods: Two thousand three hundred and five consecutive ICH patients admitted to our institution from May 2012 to October 2015 were retrospectively reviewed. Data on clinical characteristics and MRI/CT scans were collected during hospitalization. Information on clinical outcomes at three-month were also obtained.

Results: Three hundred and seventy-seven patients performed MRIs and 1966 had CTs during hospitalization. NSCI was detected in 152 (40.3%) patients with MRIs and in 638 (32.5%) with CTs. Comparing with CT, NSCI detected by MRI was more common (40.3% vs. 32.5%; P = 0.011), more likely to be multiple loci (93.4% vs. 79.6%; P < 0.001) and bilateral hemispheres (84.9% vs. 73.2%; P = 0.003). Furthermore, the presence of NSCI, multiple NSCI, bilateral NSCI and bilateral hematoma combined with bilateral NSCI were associated with poor outcomes (P < 0.001, P < 0.001, P < 0.001, P = 0.041, respectively) in univariate analysis. In multivariable logistic regression, bilateral hematoma combined with bilateral NSCI was still associated with poor outcomes (OR 3.983, 95% CI 1.172-13.539; P = 0.027).

Conclusions: Compared with CT, NSCI on MRI tends to be multiple loci and located in bilateral hemispheres. The results of NSCI in ICH may be underestimated based on CT. In addition, the increased poor outcomes at three-month suggest that NSCI may play an important role in reducing clinical outcomes.