EEG Monitoring to Detect Vasospasm after Subarachnoid Hemorrhage
ABSTRACT & COMMENTARY
Source: Vespa P, et al. Early detection of vasospasm after acute subarachnoid hemorrhage using continuous EEG ICU monitoring. EEG Clin Neurophys 1997;103:607-615.
Cerebral vasospasm after aneurysmal subarachnoid hemorrhage (SAH) is a significant source of co-morbidity, occurring in 40-70% of patients. Early detection of vasospasm would permit prompt institution of appropriate therapy, hopefully thereby reducing long-term morbidity. Vespa and colleagues now report that continuous quantitative EEG monitoring is 100% sensitive for cerebral vasospasm. Furthermore, in nearly half of the patients studied, EEG monitoring showed abnormalities before transcranial Doppler (TCD) studies or cerebral angiograms.
Vespa et al studied 32 consecutive patients with aneurysmal SAH admitted to the intensive care unit at UCLA Medical Center. Four channels of EEG signal were acquired continuously and subjected to computerized quantitative analysis. The percent of EEG activity in the "alpha" (6-14 Hz) frequency band ("relative alpha," "RA") and the variability of RA were plotted as histograms vs. time. These histograms were qualitatively analyzed visually and independently by two electroencephalographers, both of whom were blinded as to other clinical and neuroimaging findings. Also, the authors studied a quantitative measure of RA variability. The continuously acquired quantitative EEG data were compared with the results of daily neurological examinations and TCD studies, as well as cerebral angiograms and CT scans. If cerebral vasospasm was diagnosed by TCD or angiography, hypertensive hypervolemic therapy with phenylephrine and albumin was instituted. Treatment for vasospasm was not instituted based on the EEG findings.
Nineteen of 32 patients developed vasospasm. Fifteen of those developed visibly reduced RA variability coincident with the development of vasospasm. The quantitative RA variability scores declined from a mean of 0.45 to 0.17 during vasospasm (P < 0.0016). In the other four patients, RA variability was abnormally low at the outset. With resolution of vasospasm, the EEG recovered its baseline variability. Thus, Vespa et al found the sensitivity of reduced or low RA variability to be 100% for the detection of cerebral vasospasm after SAH.
The decreased RA variability preceded transcranial Doppler or angiographic documentation of vasospasm by two or more days in 10 of 19 patients. Confounding alternative causes of altered EEG in the vasospasm patients were seen in five, including intraparenchymal cerebral hemorrhage (1), elevated intracranial pressure (1), and hydrocephalus (3).
Thirteen patients did not have vasospasm. Seven of these also developed qualitatively visible reduced RA variability (i.e., these were "false positives"). However, the decline in quantitative RA variability scores did not reach statistical significance. The following neurological events accompanied these qualitative EEG changes: increased intracranial pressure (1), embolic stroke during an endovascular procedure (1), and recurrent SAH (2). The remaining three developed hemiparesis, accompanied by EEG worsening, but without documented vasospasm on TCD or angiogram. Thus, Vespa et al found the specificity of reduced RA variability to be 50% for cerebral vasospasm after SAH.
COMMENTARY
We previously reported a similar study of continuously acquired quantitative EEG monitoring for patients with SAH (Labar D, et al. EEG Clin Neurophys 1991;78:325-332). We studied 11 SAH patients, using continuous trend analyses plots over time of the following derived EEG parameters: 1) total EEG power (amplitude), 2) "alpha ratio" (7.5-15 Hz activity/1-7 Hz activity), and 3) percent delta (1-3.5 Hz activity/1-30Hz activity). For all 11 ischemic events that occurred during our study, we found 91% sensitivity for the total EEG power, 64% sensitivity for the alpha ratio, and 45% sensitivity for the percent delta. In four instances, various EEG monitoring features changed before a clinical neurological event occurred. In five instances, asymptomatic unilateral focal abnormalities on neuroimaging were accompanied by unilateral changes in total EEG power on the long-term bedside ICU monitoring.
These papers illustrate that quantitatively analyzed EEG data acquired continuously at the bedside in the neurological ICU are a powerful tool for assessing and treating patients with subarachnoid hemorrhage. Compact, portable monitoring machines capable of the analyses described, on up to eight channels of EEG signal, currently can be purchased commercially for less than $40,000. Availability of this type of monitoring should be considered a standard of care for ICUs treating SAH patients, and this type of monitoring should be used for all such cases. -drl
You have reached your article limit for the month. Subscribe now to access this article plus other member-only content. TESTING
- Award-winning Medical Content
- Latest Advances & Development in Medicine
- Unbiased Content