Induction-Phase rSO2–MAP Behaviour and Cross-Clamp Desaturation in NIRS-Guided Selective Carotid Endarterectomy: A Retrospective Cohort Study

Objective: The objectives of this study were to characterise induction-phase regional cerebral oxygen saturation (rSO2)–mean arterial pressure (MAP) dynamics during near-infrared spectroscopy (NIRS)-guided selective carotid endarterectomy (CEA) and to examine whether the Awake→Intubated pressure–oxygenation pattern may represent an early adjunctive physiological signal of subsequent cross-clamp-related ipsilateral cerebral desaturation. Methods: In this retrospective observational cohort study, 322 consecutive elective CEAs managed with an NIRS-guided selective shunting protocol between October 2019 and February 2025 were analysed, after excluding patients considered for routine pre-emptive shunting because of contralateral internal carotid artery occlusion or ≥70% stenosis. Standardised MAP and bilateral rSO2 values were extracted at the Awake, Intubated, and Clamp stages, defined as 3 min after carotid cross-clamping. Awake→Intubated ipsilateral ΔrSO2/ΔMAP was evaluated as a continuous, exploratory pressure–oxygenation index, with MAP–rSO2 directional change classified as concordant or discordant. Clamp-related desaturation was defined as a ≥20% ipsilateral rSO2 decrease from Awake to Clamp. Discrimination and adjusted associations were evaluated using receiver operating characteristic analysis and multivariable logistic regression, respectively. Results: Clamp-related ≥20% ipsilateral rSO2 desaturation occurred in 43 patients (13.4%). The Awake→Intubated ipsilateral ΔrSO2/ΔMAP ratio differed significantly between patients with and without ≥20% desaturation and showed significant discrimination on receiver operating characteristic analysis, with an area under the curve (AUC) of 0.799 (95% confidence interval [CI] 0.723–0.876; p < 0.001). Concordant pressure–oxygenation change was more frequent among patients with ≥20% desaturation (31/43, 72.1%), whereas discordant change predominated among those without desaturation (228/279, 81.7%; p < 0.001). In multivariable analysis, Awake→Intubated ipsilateral ΔrSO2/ΔMAP remained associated with clamp-related ≥20% desaturation after adjustment (adjusted odds ratio [OR] 1.63, 95% CI 1.15–2.33; p = 0.006), along with symptomatic presentation and 50–69% contralateral stenosis. Postoperative stroke occurred in 4/322 patients (1.2%), and no 30-day mortality occurred. Conclusions: During NIRS-guided selective CEA, induction-phase rSO2–MAP dynamics were associated with subsequent cross-clamp-related ipsilateral cerebral desaturation. As the outcome was a NIRS-defined desaturation rather than an independent clinical, neurological, or imaging endpoint, these findings indicate association with a surrogate marker rather than prediction of clinically relevant cerebral ischaemia. The Awake→Intubated ΔrSO2/ΔMAP ratio and directional pressure–oxygenation pattern may represent early adjunctive physiological signals associated with clamp-related desaturation. These findings are hypothesis-generating and require prospective validation with systematic multimodal monitoring.