Hyperglycemia Is Associated With Computed Tomography Perfusion Core Volume Underestimation in Patients With Acute Ischemic Stroke With Large-Vessel Occlusion.

BACKGROUND

CT Perfusion (CTP) predictions of infarct core play an important role in the determination of treatment eligibility in large vessel occlusion (LVO) acute ischemic stroke (AIS). Prior studies have demonstrated that blood glucose can affect cerebral blood flow (CBF). Here we examine the influence of acute and chronic hyperglycemia on CTP estimations of infarct core.

METHODS

From our prospectively collected multi-center observational cohort, we identified patients with LVO AIS who underwent CTP with RAPID (IschemaView, Stanford, CA) post-processing, followed by endovascular therapy with substantial reperfusion (TICI 2b-3) within 90 minutes, and final infarct volume (FIV) determination by MRI 48-72 hours post-treatment. Core volume over- and under-estimations were defined as a difference of at least 20 mL between CTP-RAPID predicted infarct core and DWI FIV. Primary outcome was the association of presentation glucose and HgbA1c with underestimation (UE) of core volume and was measured using multivariable logistic regression adjusted for comorbidities and presentation characteristics. Secondary outcomes included frequency of overestimation (OE) of infarct core.

RESULTS

Among 256 patients meeting inclusion criteria, median age was 67 [IQR 57-77], 51.6% were female, and 132 (51.6%) and 93 (36.3%) had elevated presentation glucose and elevated HgbA1c, respectively. Median CTP-predicted core was 6 mL [IQR 0-30], median DWI FIV was 14 mL [IQR 6-43] and median difference was 12 mL [IQR 5-35]. Twenty-eight (10.9%) patients had infarct core OE and 68 (26.6%) had UE. Compared to those with no UE, patients with UE had elevated blood glucose (median 119 [103-155] vs 138 [117-195], p=0.002) and HgbA1c (median 5.80 [5.40-6.40] vs 6.40 [5.50-7.90], p=0.009). In multivariable analysis, UE was independently associated with elevated glucose (aOR 2.10, p=0.038) and HgbA1c (aOR 2.37, p=0.012). OE was associated with lower presentation blood glucose (median 109 [ 99-132] in OE vs 127 [107-172] in no OE, p=0.003) and HgbA1c (5.6 [IQR 5.1 - 6.2] in OE vs 5.90 [5.50-6.70] in no OE, p=0.012).

CONCLUSIONS

Acute and chronic hyperglycemia were strongly associated with CTP UE in patients with LVO AIS undergoing EVT. Glycemic state should be considered when interpreting CTP findings in patients with LVO AIS.