Effects of Overnight Intravenous Gamma-Hydroxybutyrate on Sleep Architecture in Uns sedated ICU Patients: A Polysomnographic Study

By
Florian Blanchard
Isabelle Arnulf
Noémie Clavieras
Fanny Charbonnier-Beaupel
Rayan Braik
Arthur James
Jean-Michel Constantin
October 30, 2026

Critical Care

Overview

In this uncontrolled polysomnographic experience involving 13 awake, non-delirious, non-invasively ventilated ICU patients, the eight who received overnight intravenous gamma-hydroxybutyrate (GHB) had higher sleep efficiency and a greater proportion of N3 sleep than the five who did not. REM sleep was less frequent with GHB, and morning natraemia was higher. The small sample, absence of pretreatment recordings, and potential confounding by indication limit interpretation.

Background

Sleep in awake ICU patients is often fragmented, with reduced slow-wave and REM sleep; these disturbances have been linked in the source article to delirium and prolonged liberation from mechanical ventilation. The article describes pharmacological options as limited, noting low-certainty evidence for melatonin and that benzodiazepines and propofol can extend sleep while impairing sleep architecture. GHB, or sodium oxybate, is a GABA-B agonist used to consolidate nocturnal sleep in narcolepsy and is reported to increase slow-wave sleep. Its effects on sleep in critically ill patients had not been described in the article's clinical context.

Data Highlights

Measure	GHB (n=8)	No GHB (n=5)	Reported comparison
Sleep efficiency	82% [73–90]	57% [52–68]	Difference 21; 95% CI 7 to 35
N3 as proportion of total sleep time	58% [28–77]	8% [1–17]	Difference 48; 95% CI 18 to 76
REM as proportion of total sleep time	5%	13%	REM absent in two GHB recipients
Morning sleep-quality NRS	6 [6–7]	4 [2–5]	Scale: 0 (worst) to 10 (best)
Nurse-documented sleep	8/8	2/5	All GHB recipients were judged to have slept
Morning natraemia	140 mmol/L	135 mmol/L	Higher following GHB

Key Findings

Interpretable polysomnographic recordings (at least 4 hours of signal) were available for 13 patients; recordings were obtained on ICU day 4 [3–11]. All patients were calm and awake, had negative CAM-ICU assessments, and were not invasively ventilated.
Eight patients received intravenous GHB, at a median dose of 9.0 g [8.5–9.7], or 126 mg/kg [106–159]; five patients did not receive GHB.
Sleep efficiency and the proportion of sleep classified as N3 were higher in the GHB group, while light sleep was reduced. N3 accounted for more than half the night in four GHB recipients and 99% in one.
REM sleep was less frequent in the GHB group (5% versus 13%) and absent in two GHB recipients.
Subjective assessments favored the GHB group: nurses documented sleep in all eight recipients versus two of five nonrecipients, and median NRS scores were 6 versus 4. Among six GHB recipients with a preceding-night score, the median NRS rose from 3 [2–6] to 7 [6–7], with improvement reported by all six.
No bradycardia, hypotension, desaturation, vomiting, or seizures were observed. Morning natraemia was higher after GHB (140 versus 135 mmol/L); three patients had recurrent arousals from slow-wave sleep, and one became agitated late in the night and received diazepam. No patient was delirious the next day.

Clinical Implications

The authors note that each gram of sodium oxybate contains approximately 8 mmol of sodium and that a 9 g dose supplies roughly 70 mmol, warranting daily electrolyte monitoring. They also caution that AASM scoring cannot distinguish physiological slow-wave sleep from drug-induced delta activity, and that the clinical significance of the observed N3 activity and the effect of repeated administration on REM sleep remain unknown. The authors identify the small, uncontrolled experience, lack of pretreatment recordings, and possible confounding by indication and regression to the mean as limitations.

Conclusion

In this small ICU experience, overnight intravenous GHB was associated with higher sleep efficiency and substantially more scored N3 sleep, alongside reduced REM sleep and higher morning natraemia. The uncontrolled design and uncertainty about the physiological meaning of drug-associated sleep patterns preclude conclusions about efficacy or effects with repeated use.