To characterize the relationship between the presence and integrity of sleep architecture on electroencephalogram (EEG) with clinical features of patients with non-traumatic subarachnoid hemorrhage (SAH).

While the absence of sleep architecture is associated with poor outcomes in high grade SAH, the incidence and characteristics of sleep architecture across the spectrum of SAH is not well-described.

We conducted a retrospective chart review of sequential patients with acute SAH identified using a specially-designed problem list query integrated into our institution’s electronic health record (covering October 2018- August 2019). Clinical data were obtained in all patients; and EEG characteristics in those patients undergoing monitoring (mean=35 hours). Group comparisons were made between monitored and non-monitored patients, and evidence of EEG sleep architecture recorded. Data are reported as mean±SD for parametric data or median [IQR] for non-parametric data, with significance set at P≤0.05.

 We identified 51 SAH patients, 24% (n=12) of whom underwent any EEG monitoring. Monitored patients had higher Fisher scores (4 [4-4] v. 4[3-4], P=0.05), longer ICU (16±8 v. 9±5 days, P= 0.002) and hospital stays (22±18 v. 12±8 days, P= 0.015), but had no differences in sex, age, or Hunt & Hess scales. Of the 12 subjects with EEG studies to review, stage 2 sleep architecture was found to be absent in 7, fragmented in 3, and intact in 2.  No stage 3 or 4 sleep was not seen in any patient.   

Patients with abnormal sleep architecture had the longest hospital stays (25±18 days) and ICU stays (18±7 days), and only 3/10 discharged to home, compared to 2/2 with normal sleep architecture.

While normal sleep architecture can be observed after SAH and may portend good prognosis, while abnormal or absent sleep architecture is persistently associated with poorer outcomes over a broad range of SAH severity.