PET was performed a median of 6 h from time of angiography in this cohort of 25 patients. Fourteen patients had moderate-severe CVS in at least one vascular territory. Although 12 of the patients had symptomatic DCI at some point in their course, only 3 patients were being actively treated with hypertensive therapy at time of PET. Excluding these patients, mean arterial blood pressure was not different between those with and those without CVS. Twenty-four percent of all brain regions were affected by significant (proximal and/or distal) CVS, involving 34 vascular territories across the 14 affected patients. Global mean CBF tended to be lower in patients with CVS (38.4 ± 9.7 vs. 44.3 ± 14.5 ml/100 g/min in those without CVS, p = 0.07). Regional CBF was also lower (mean 38.6 ± 11.7) in regions affected by CVS compared with 48.7 ± 16.5 ml/100 g/min in brain regions without CVS (p < 0.001). A graded reduction in CBF was seen between mild, moderate, and severe degrees of CVS (Fig. 1). This reduction in perfusion was present whether there was only proximal, only distal, or both proximal and distal CVS present.

More relevant to the question at hand, however, regional hypoperfusion (rCBF < 25 ml/100 g/min) was not confined or even preferentially located within territories affected by CVS. We found regions with low CBF in 10 of 25 patients, including 7 of 14 patients with CVS but also 3 of 11 patients without significant CVS. Nine of these ten patients with hypoperfusion had neurological symptoms during their course, but none were on vasopressors at the time of PET. Interestingly, two of the three patients with hypoperfusion but no CVS still had neurological deficits. A total of 46 brain regions exhibited hypoperfusion (CBF < 25 ml/100 g/min) across ten patients. Regional hypoperfusion did not occur more frequently in territories with significant CVS (15 of 157 affected regions, 10 %) vs. 20 of 217 (9 %) of unaffected regions, even when restricting analysis to only patients with CVS (i.e., comparing territories with and without vasospasm in the same group of patients). Therefore, the majority of regions with hypoperfusion were found outside territories with CVS. In fact, 54 % of hypoperfused regions were located in vascular territories with no CVS anywhere in the ipsilateral carotid (or, if appropriate, vertebrobasilar) circulation. When the 11 patients without CVS were also included, a total of 31 of the 46 (67 %) hypoperfused regions were not in the territory of any affected vessels (Fig. 2). This relationship persisted even after excluding the three patients on vasopressors, in whom treatment could have masked an association between vasospasm and hypoperfusion.

Delayed cerebral infarcts were seen in 20 of the 134 patients (15 %) after exclusion of early/procedural infarcts (which accounted for an equivalent proportion of the total infarcts); moderate-severe angiographic CVS was present in 34 % of this cohort. The rate of delayed infarction was higher in those with CVS compared with those without CVS (31 % vs. 4 %, p < 0.001), although in one of these patients with CVS and infarction, the infarct actually occurred in a vascular territory without CVS. Fifteen of the 20 patients with delayed infarcts had infarcts in territories with CVS, 4 had infarcts unrelated to CVS, and 1 patient had both infarcts related and unrelated to vasospasm. There were no patient characteristics that differentiated these groups, although the number of patients in each group was admittedly low. There were a total of 29 delayed cerebral infarcts, of which 21 were associated with corresponding angiographic CVS, whereas 8 (28 %) occurred in the absence of any proximal CVS. Half of the vasospasm-related infarcts were in the ACA territory whereas no infarcts unrelated to CVS were in the ACA region. Rather, half the vasospasm-unrelated infarcts occurred in watershed territories, compared with 10 % of infarcts seen with CVS (p = 0.03). There was no difference in volume of these two infarct types and they were evenly divided between cortical and subcortical regions.