White Matter Lesions Double the Risk of Post-Thrombolytic Intracerebral Hemorrhage

Abstract

Background and Purpose-Cerebral white matter lesions (WMLs), a surrogate for small-vessel disease, are common in patients with stroke and may be related to an increased intracranial bleeding risk after intravenous thrombolysis in acute ischemic stroke. We aimed to investigate the risk of symptomatic intracerebral hemorrhage (sICH) in the presence of WMLs in a large cohort of ischemic stroke patients treated with intravenous thrombolysis. Methods-We included 2485 consecutive patients treated with intravenous thrombolysis at the Helsinki University CentralHospital. WMLs were scored according to 4 previously published computed tomography visual rating scales from all baseline head scans. A sICH was classified according to the European Cooperative Acute Stroke Study II criteria. The associations of sICH with nominal, ordinal, and continuous variables were analyzed in a univariate binary regression model and adjusted in multivariate binary regression models. Results-In univariate and multivariate regression analyses, all 4 tested visual WML rating scales (as continuous variables or dichotomized at different cutoff points) were associated with increased risk of sICH. In binary analyses, WML doubled the bleeding risk: the odds ratios of all 4 visual rating scales ranged from 2.22 (95% confidence interval, 1.49-3.30) to 2.70 (1.87-3.90) in univariable and from 2.00 (1.26-3.16) to 2.62 (1.71-4.02) in multivariable analyses. The multivariableadjusted odds ratio for the association of high load of WMLs with remote parenchymal hemorrhage was 4.11 (2.38-7.10). Conclusions-WMLs visible on computed tomography are associated with a more than doubled risk of sICH in patients treated with intravenous thrombolysis for acute ischemic stroke.

Leukoaraiosis, Cerebral Hemorrhage, and Outcome After Intravenous Thrombolysis for Acute Ischemic Stroke

“…4, 6, 14, 15, 18, 19 The pooled incidence of poor outcome (defined as mRS>2 in three studies, 6, 14, 18 mRS≥2 in two studies 15, 19 ) was 47% (95%CI: 34%–60%) in the leukoaraiosis negative group compared to 64% (95%CI: 55%–74%) in the leukoaraiosis positive group. Pooled analysis of these studies demonstrated the odds of poor outcome at 3–6 month post-stroke to be significantly elevated in those with evidence of leukoaraiosis (OR: 2.02; 95%CI: 1.54–2.65, p<0.0001, Figure 3A), with significant statistical heterogeneity between studies (I 2 :73.5%, p=0.002).…”

“…Six of the included studies (n=4976 patients, 71% of the whole included patient population in the systematic review) provided adjusted effect sizes of the independent association between leukoaraiosis and sICH in multivariable analyses (adjusted for different co-variates depending on each study’s results). 6, 13, 14, 16–18 In a pooled sensitivity analysis of these studies, there was a significant association between leukoaraiosis and sICH risk with an adjusted OR of 1.75 (95%CI: 1.35–2.27; p<0.0001) and moderately high degree of statistical heterogeneity (Figure 2C). …”

“…4–6, 12–19 For the ECASS-1/−2 report (published only as an abstract), we contacted the lead authors to provide relevant data for the meta-analysis. 18 A summary of the characteristics of included studies, methodological key issues and quality indicators are noted in Table 1.…”

“…As detailed below, not all studies provided the same range of outcome measures, resulting different total number of patients contributing in each meta-analysis. Six studies provided data on severe leukoraiosis (Table 1): 6, 13–15, 17, 18 all used the Van Swieten scale 7, 8 and severe leukoaraiosis was defined as >2 (or >4 if each hemisphere was rated separately). There was no evidence of publication bias in the funnel plot (data not shown).…”

“…Estimates in separate sensitivity analyses indicated that the effect size is greater in observational studies than in RCTs (but still in the same direction), raising a concern about bias and inflation of the overall results by the non-RCTs. For example, the study by Curtze et al, 5, 6 the largest prospective observational study on the topic, provided similarly conservative effect estimates to RCTs, 4, 15, 18 in contrast to retrospective small studies. Finally, the main limitation of our analysis is that it is not fully adjusted for other important variables that can potentially affect the risk for thrombolysis-related sICH and outcome, such as time from stroke onset to IV thrombolysis, baseline serum glucose, infarct volume, early ischemic CT signs, pre-stroke functional status etc.…”