Sami Al Kasab, MD
Intracerebral hemorrhage (ICH) is a known complication of intravenous alteplase. The rates of symptomatic intracerebral hemorrhage following intravenous alteplase administration have varied between 1-4% depending on the definition used and the study.
In this study, Drs. Connor et al analyze the association between multimodal CT imaging parameters, including NCCT hypo attenuation degree, vlCBV, impaired blood-brain barrier permeability surface product, clinical and laboratory data at baseline, early reperfusion status, and development of parenchymal hemorrhage (PH) on follow-up imaging. All patients received NCCT, CT angiography, and CT perfusion at baseline. A 24 to 48 hour scan (either NCCT or MRI) was obtained. Imaging analyses were performed by readers blinded to other imaging and patient outcomes. All NCCT were scored for ASPECTS scores, the degree of hypo attenuation within the ischemic region using a 3-point grading system. Functional parametric maps of cerebral blood flow (CBF), CBV, Tmax, and a modified CTP algorithm for permeability calculations were used.
One-hundred and eighteen subjects (22 patients with parenchymal hemorrhage (PH) versus 96 without) were included in the final analysis. Median NIHSS was 15, median ASPECT was 8, and median stroke onset to imaging time was 112 minutes. Median baseline NIHSS was higher among patients that experienced PH (19 versus 15, P=0.01), and median ASPECTS score was lower among patients that experienced PH (7 versus 8, P=0.01). The proportion of PH versus no PH in subjects with NCCT hypo attenuation grade 2 or 3 was 4x that in patients with hypo attenuation grade 1 (29.5% versus 7%, P <0.001).
On logistic regression models, NCCT hypo attenuation grade (P=0.004) was the only significant variable associated with PH on follow-up imaging for model 1 (area under the curve, 0.69, 95%CI, 0.61-0.77), whereas NCCT hypo attenuation grade (p=0.006) and rvlCBV (p=0.04) were significant variables associated with PH on follow-up imaging for model 2 (area under the curve, 0.73; 95% CI, 0.63-0.83; variance inflation factor, 1.09).
The study has several limitations. It is a case-control study. The sample size is also small. The authors report all hemorrhagic transformations and not only clinically significant hemorrhagic transformations (PH2).
In conclusion, hypo attenuation on NCCT and very low CBV on CTP can help determine the risk of hemorrhagic transformation in patients receiving intravenous alteplase.
This study is the largest sample study in our field that has looked at multi-modality imaging and its utility in predicting ICH in patients receiving intravenous alteplase. Since ICH post alteplase is a rare event, we deliberately used a case control design as otherwise classic statistical models that tend to predict rare events are unstable. Moreover, our study finds that NCCY hypo attenuation is the most robust predictor of ICH in these patients; an important message for our field as it is one modality that is available everywhere. The message from our paper is for clinicians to look at the NCCT more carefully while estimating treatment risk.
Dr.Menon, thank you for your input and for explaining the rationale behind the statistical approach. The study is certainly of a clinical significance, given that non-contrast head CT is underutilized despite its availability. This would be even more important when the decision to administer tPA is made in patients with mild stroke symptoms which many clinicians tend to treat, in these cases the use of NCCT could be used to make the decision of whether or not to administer tPA.