The metabolic and autoregulatory profile of reversible delayed cerebral ischemia in unconscious patients after aneurysmal subarachnoid hemorrhage: a prospective multimodal neuromonitoring cohort study

Abstract

Background: The detection and treatment of delayed cerebral ischemia (DCI) following aneurysmal subarachnoid hemorrhage (SAH) remain challenging. Multimodal neuromonitoring and CT perfusion scanning (CTP) are promising tools for diagnosing DCI in unconscious patients. This study aims to compare the metabolic and autoregulatory characteristics of patients with cerebral hypoperfusion indicative of DCI that either resolves post-treatment or progresses to infarction due to treatment failure.

Methods: In a cohort of 268 consecutive SAH patients, neuromonitoring-comprising intracranial pressure (ICP) and brain tissue oxygen (PtiO₂) measurements, and/or cerebral microdialysis-was implemented in 126 (47%) neurologically unassessable patients. Aberrant neuromonitoring measurements triggered CTP, and in cases of confirmed perfusion deficits, first-tier treatment involved induced hypertension. Non-responsive perfusion deficits were further evaluated with conventional angiography, and spasmolysis or angioplasty was performed if suitable vasospasm was identified. DCI-related infarction was noted on CT imaging at discharge, and clinical outcomes were assessed using the modified rankin scale (mRS) at 12 months. Using a generalized linear mixed-effects model (GLMM), factors associated with the occurrence of DCI-related infarction were assessed.

Results: CTP deficits were identified in 72 (57%) patients, of whom 63 (88%) had neuromonitoring probes near the affected areas. In 24 patients (38%), perfusion deficits progressed to infarction, while in 39 (62%), deficits were successfully reversed through induced hypertension or spasmolysis. In a GLMM, lower pressure reactivity index (PRx-OR 2.70, 95% CI 1.04-4.67; p < 0.001) and lower lactate-to-pyruvate ratio (LPR-OR 1.02, 95% CI 1.01-1.03; p < 0.001) were independently associated with better treatment response and reduced infarction risk, after adjusting for clinical hemorrhage severity. These effects were observed more than 24 h before cerebral hypoperfusion. Pooled PRx and LPR over this time frame were not associated with functional outcome.

Conclusion: Loss of cerebrovascular reactivity and metabolic disturbances precede cerebral hypoperfusion in SAH. Lower PRx and LPR levels are independently associated with improved DCI treatment efficacy. These findings must be interpreted in the context of study limitations, including the small sample size and the focal nature of microdialysis measurements. Nevertheless, the results suggest that invasive neuromonitoring may aid in identifying patients more likely to benefit from treatment.

Trial registration: This project was retrospectivly registered in the German Clinical Trial Register (DRKS00030505) on the third of January 2023.

Keywords: Cerebral autoregulation; Cerebral infarction; Cerebral microdialysis; Delayed cerebral ischemia; Perfusion computed tomography; Subarachnoid hemorrhage.

Fig. 1

Exemplary depiction of the spatial relation between perfusion deficits and invasive neuromonitoring probes as a prerequisite for study inclusion. A Aneurysmal subarachnoid hemorrhage with left-sided blood distribution provided with left-sided dual invasive neuromonitoring (PtiO₂ and cerebral microdialysis) in the frame on the right. B Bifrontal, predominantly left-sided perfusion deficits developed in an area covered by the probes. Patients with this constellation, were included in the study. C The affected area evolves into cerebral infarction despite treatment. Patients with this course of events were included in the study group with infarction. D The affected area is spared from infarction due to successful treatment. These patients were included as the no infarction group. E Patients with perfusion deficits outside the reach of invasive neuromonitoring were excluded from this analysis. PtiO₂, brain tissue oxygen pressure

Fig. 2

Inclusion flow chart. CTP, perfusion computed tomography imaging; DCI, delayed cerebral ischemia; inf., cerebral infarction; MMM, multimodal monitoring; SAH, aneurysmal subarachnoid hemorrhage

Fig. 3

Generalized additive mixed model (GAMM) predicted time courses for A cerebral perfusion pressure (CPP), B brain tissue oxygen pressure (PtiO₂ (mmHg)) pressure and C pressure reactivity index (PRx); 72 h before radiological confirmation of DCI and treatment initiation. Predictions have been split by infarction status. Shaded areas indicated 95% confidence intervals. The superimposed scatterplot represents raw data expressed as hourly mean (CPP & PRx) or hourly median (PtiO₂) values. D Individually averaged PRx over 72 h of time split by infarction status. Reported p-values in panel A, B & C represent Likelihood Ratio Test results for nested models with or without DCI infarction as a predictor. In panel D, the p-value represents the result of an independent T-test performed on patient values averaged over the 72-h time window. DCI, delayed cerebral ischemia; hrs, hours; mmHg, millimeters of mercury

Fig. 4

Generalized Additive Mixed Model (GAMM) predicted time courses of microdialysis analytes A lactate-to-pyruvate ratio (LPR), B individual median LPR over 72 h of time split by infarction status, C glutamate (µmol/L), D individual median glutamate concentrations over 72 h of time split by infarction status. Shaded areas indicated 95% confidence intervals. The superimposed scatterplot represents raw data expressed as one median per hour. Reported p-values in panel A & C represent Likelihood Ratio Test results for nested models with or without DCI infarction as a predictor. In panel B & C, the p-value represents the result of a Mann–Whitney U tests performed on a data consisting of a single median per patient over the 72-h time window. DCI, delayed cerebral ischemia; hrs, hours; LPR, lactate-to-pyruvate ratio