Hyperventilation and breath-holding test with indocyanine green kinetics predicts cerebral hyperperfusion after carotid artery stenting

Abstract

Cerebral hyperperfusion syndrome (CHS) is a serious complication following carotid artery stenting (CAS), but definitive early prediction of CHS has not been established. Here, we evaluated whether indocyanine green kinetics and near-infrared spectroscopy (ICG-NIRS) with hyperventilation (HV) and the breath-holding (BH) test can predict hyperperfusion phenomenon after CAS. The blood flow index (BFI) ratio during HV and BH was prospectively monitored using ICG-NIRS in 66 patients scheduled to undergo CAS. Preoperative cerebrovascular reactivity (CVR) and the postoperative asymmetry index (AI) were also assessed with single-photon emission computed tomography before and after CAS and the correlation with the BFI HV/rest ratio, BFI BH/rest ratio was evaluated. Twelve cases (18%) showed hyperperfusion phenomenon, and one (1.5%) showed CHS after CAS. A significant linear correlation was observed between the BFI HV/rest ratio, BFI BH/rest ratio, and preoperative CVR. A significant linear correlation was observed between the BFI HV/rest ratio and postoperative AI (r = 0.674, P < 0.0001). A BFI HV/rest ratio of 0.88 or more was the optimal cut-off point to predict hyperperfusion phenomenon according to receiver operating characteristic curve analyses. HV and BH test under ICG-NIRS is a useful tool for detection of hyperperfusion phenomenon in patients who underwent CAS.

Keywords: Carotid artery stenting; hyperperfusion syndrome; hyperventilation test; indocyanine green; near-infrared spectroscopy.

Figure 1.

Schematic drawing of the hyperventilation and breath-holding protocol. After 10 minutes of normal breathing at rest, the patient performed voluntary hyperventilation or breath-holding, EtCO₂ and PaCO₂ were determined, and ICG was injected. BFI and TTP were calculated from the ICG time-intensity curve with N200NX ICG Analyze software after each procedure.

Figure 2.

Preoperative cerebrovascular reactivity (CVR) and postoperative asymmetry index (AI) the day after stent placement. A CVR value lower than the mean + 2 SD (i.e. 23.8%) was defined as reduced CVR (vertical line), and an AI value higher than the mean + 3 SD (i.e. 104.6%) was defined as an increase in AI (horizontal line). The black closed circle indicates the patient with hyperperfusion syndrome, and the gray closed circles indicate patients with hyperperfusion phenomenon after CAS.

Figure 3.

Typical changes in the ICG time-intensity curve and SPECT findings a patient with and a patient without CHS after the CAS procedure. In non-CHS case, preoperative ¹²³I-IMP SPECT in the resting state showed no decrease in rCBF in the left hemisphere before (a) and one day after CAS (b). The time-intensity curve from ICG-NIRS monitoring at rest (c), just after hyperventilation (d), and just after breath-holding (e) in a patient without CHS. A remarkable change in the BFI calculated from the time-intensity curve just after hyperventilation and breath-holding was observed. In CHS case, preoperative ¹²³I-IMP SPECT in the resting state showed a decrease in rCBF in the right hemisphere (f). One day after CAS, hyperperfusion was seen in the ipsilateral hemisphere (g). One week after CAS, hyperperfusion in the ipsilateral hemisphere had normalized (h). The time-intensity curve from ICG-NIRS monitoring at rest (i), just after hyperventilation (j), and just after breath-holding (k) in a patient with CHS. The BFI values were nearly unchanged during rest, hyperventilation, and breath-holding.

Figure 4.

(a) The relationship between the BFI hyperventilation (HV)/rest ratio and preoperative cerebrovascular reactivity (CVR). (b) The relationship between the BFI breath-holding (BH)/rest ratio and preoperative CVR. A significant linear correlation was observed between the BFI HV/rest ratio and preoperative CVR, and between the BFI BH/rest ratio and preoperative CVR (r = −0.703 and 0.576, respectively P < 0.0001 for both). (c) The relationship between the TTP HV/rest ratio and preoperative CVR. (d) The relationship between the TTP BH/rest ratio and preoperative CVR. A significant linear correlation was observed between the TTP HV/rest ratio and preoperative CVR, and between the TTP BH/rest ratio and preoperative CVR (r = 0.731 and −0.509, respectively, P < 0.0001 for both). The black closed square indicates the patient with hyperperfusion syndrome, and the gray closed square indicates patients with the hyperperfusion phenomenon after CAS.

Figure 5.

(a) The relationship between the BFI hyperventilation (HV)/rest ratio and the postoperative asymmetry index (AI). A significant linear correlation was observed between the BFI HV/rest ratio and postoperative AI (r = 0.674, P < 0.0001). (b) The relationship between the BFI breath-holding (BH)/rest ratio and the postoperative asymmetry index (AI). A significant linear correlation was observed between the BFI BH/rest ratio and postoperative AI (r = −0.413, P < 0.01). (c) Receiver operating characteristic (ROC) curves for the BFI HV/rest ratio, the BFI BH/rest ratio, and the preoperative CVR. The ROC curve suggested that a BFI HV/rest ratio of 0.88 or more was the optimal cut-off point for predicting hyperperfusion phenomenon (area under the curve, 0.88; sensitivity, 0.83; specificity, 0.75; 95% confidence interval, 0.78–0.97; P < 0.0001).