CT hypodensities unreliable marker of cerebral vasospasm
MedWire News: Even experienced clinicians rarely agree on whether hypodensities on computed tomography (CT) scans of patients with aneurysmal subarachnoid hemorrhage relate to vasospasm, research shows.
"We also find that CT hypodensities are not a robust marker of angiographic vasospasm severity, even when expert consensus exists that their presence represents angiographic vasospasm-related DIND [delayed ischemic neurological deficit]," say the researchers.
For the study, two experienced neurosurgeons and a neuroradiologist independently reviewed CT scans of 413 patients who participated in the Clazosentan to Overcome Neurological iSChemia and Infarction OccUrring after Subarachnoid hemorrhage (CONSCIOUS-1) trial.
The researchers found considerable variability between the reviewers regarding whether CT hypodensities related to vasospasm. Agreement was moderate to substantial, with kappa values ranging from 0.51 to 0.78, where 1.00 represents perfect agreement.
The reviewers unanimously agreed that hypodensities in 21 patients were caused by vasospasm-induced DIND. Although 43% of these patients had severe vasospasm on angiography, 19% had little or no vasospasm. The specificity of CT hypodensities for severe vasospasm was 93%, but sensitivity was just 41%. The sensitivity for moderate or severe vasospasm was lower still, at 28%.
The reviewers were unanimous in deciding that hypodensities in 164 patients had causes not related to DIND, and 74% of these patients did indeed have mild or no vasospasm. However, 8% had severe vasospasm.
The reviewers were more likely to attribute hypodensities to DIND than other causes if patients had ventriculostomy or shunt on postprocedure CT (61 vs 45%), but less likely if they had new infarcts on the postprocedure scan (26 vs 48%).
"It would appear that differentiating between hypodensities related to treatment complications and angiographic vasospasm is indeed difficult," R Loch Macdonald (St Michael's Hospital, Toronto, Ontario, Canada) and team comment in the journal Stroke.
The reviewers' decisions were not associated with the method of aneurysm treatment (clipping or coiling), the presence of intracerebral hemorrhage on baseline scans, or the use of rescue therapies for vasospasm.
"The results of this study are important, because treatment of angiographic vasospasm can be associated with morbidity, particularly in patients with cardiovascular disease who may not tolerate hemodynamic therapy," say the researchers.
Although the decision to treat vasospasm is not based purely on CT findings, these may have an important influence, they say. Based on the current study, around 20% of patients may have mild or no vasospasm, despite their CT findings indicating the presence of DIND.
"This 'false-positive' group is at risk of overtreatment with pharmacologically induced hypertension and hypervolemia, assuming these measures are not effective for DIND not associated with angiographic vasospasm, which is not known," say Macdonald et al.
By Eleanor McDermid