The study, known as DUET, was a prospective, multicenter, single-arm, open-label trial that enrolled 48 adults with narcolepsy type 1 or type 2 who were already taking the maximum recommended nightly dose of 9 grams of low-sodium oxybate, a medication sold under the brand name Xywav. Rather than simply observing these patients, investigators invited them to titrate upward — slowly, cautiously, and always under the watchful eye of overnight polysomnography — to a dose of up to 12 grams per night, split into two doses taken at bedtime and again roughly four hours later.

The pharmacological logic behind the trial is straightforward. Oxybate, the active moiety shared by both high-sodium and low-sodium formulations, is a central nervous system depressant that consolidates fragmented nighttime sleep and suppresses the REM intrusion that defines narcolepsy. The 9-gram ceiling was set in 2002 with the original FDA approval of sodium oxybate, based on dose cohorts of 3, 6, and 9 grams in a phase 3 registration study that showed a strong dose-dependent reduction in cataplexy of up to 69 percent at the highest dose. But real-world data from the REMS database and the Nexus Narcolepsy Registry revealed that roughly 1 percent and 4 percent of patients, respectively, were quietly taking doses above 9 grams — a signal that the ceiling might be an artifact of the registration studies rather than a true pharmacological limit.

DUET was designed to test that signal rigorously. Participants completed a 9-day baseline assessment while still on 9 grams per night, then spent between 2 and 8 weeks titrating upward in increments of no more than 0.5 grams per week. Every dose escalation was performed in a sleep clinic under overnight polysomnographic monitoring — a safety precaution that reflects the drug’s status as a CNS depressant with the potential to slow breathing during sleep. Once investigators judged that a participant had reached an optimal dose, that dose was held stable for two weeks before a final 9-day end-of-treatment assessment, mirroring the baseline procedures exactly.

The results, published in the Journal of Clinical Sleep Medicine, are striking. The mean stable dose achieved was 11.1 grams per night, and nearly 44 percent of participants titrated all the way to 12 grams — the maximum allowed. Against that backdrop, the Epworth Sleepiness Scale, the gold-standard patient-reported measure of daytime sleepiness, fell from a mean of 11.4 at baseline to 8.3 at end of treatment. That 3.1-point drop is not merely statistically significant; it exceeds the 2-point threshold that the American Academy of Sleep Medicine considers clinically meaningful, and it pushed the cohort mean below the score of 10 that defines excessive daytime sleepiness in the first place.

The improvements extended well beyond sleepiness. For participants with narcolepsy type 1, the Narcolepsy Severity Scale clinical trials version — a 15-item instrument that captures the full pentad of narcolepsy symptoms — fell from 17.3, indicating moderate severity, to 9.4, indicating mild severity. For those with type 2, the modified scale dropped by 5.5 points, well beyond the minimal clinically important difference of roughly 3.4 points established in longitudinal studies. Every component of the scale improved, including daytime sleepiness, cataplexy, hypnagogic and hypnopompic hallucinations, sleep paralysis, and disrupted nighttime sleep.

Cataplexy, the sudden loss of muscle tone triggered by emotion, showed particularly dramatic gains. Based on the Narcolepsy Severity Scale, the mean number of weekly cataplexy episodes fell from 7.3 to 2.6 — a 36 percent reduction that comfortably exceeds the 25 percent clinical significance threshold set by the AASM. The electronic diary data, which captured daily attacks in a subset of participants, showed an even steeper decline from 4.1 to 0.4 attacks per week. Perhaps most tellingly, the proportion of participants rating their overall disease as severe on the Patient Global Impression of Severity scale plummeted from 61.4 percent to 29.5 percent, and 92.3 percent reported that their disease had improved by the end of the study. Not a single participant said their disease had worsened.

The safety findings are arguably as important as the efficacy ones. Because oxybate is a CNS depressant, the central clinical concern with pushing doses higher is respiratory depression — the possibility that the drug could suppress breathing during sleep, particularly in patients with underlying sleep-disordered breathing. The trial excluded anyone with an untreated apnea-hypopnea index above 10, and participants with adequately treated obstructive sleep apnea were required to maintain their positive airway pressure or oral appliance therapy throughout. Against that backdrop, the polysomnographic respiratory measures remained essentially unchanged: the apnea-hypopnea index held steady at roughly 1.6 to 1.8 events per hour, the central apnea index remained near zero, and mean oxygen saturation stayed at 95.8 percent. All 43 participants who completed the end-of-treatment polysomnogram had an apnea-hypopnea index of 10 or below.

That is not to say the trial was entirely without respiratory signals. A handful of participants showed elevated apnea-hypopnea indices during sequential titration safety polysomnograms — predominantly central apneas in some cases, obstructive apneas in others — though these normalized by the end-of-treatment visit or led to discontinuation. One participant experienced a single instance of narrow complex tachycardia during sleep during the first titration polysomnogram, with a maximum heart rate of 158 beats per minute, though subsequent recordings showed heart rates in the 80 to 90 beats per minute range. The most common treatment-emergent adverse events were nausea, vomiting, and headache, occurring in roughly 10 to 12 percent of participants — a profile consistent with the known safety profile at lower doses. One serious adverse event, severe syncope during a morning blood draw, was judged unrelated to the study drug.

The trial’s design has important implications for how oxybate is prescribed going forward. The low-sodium formulation contains 92 percent less sodium than high-sodium oxybate, a distinction that matters enormously at higher doses: a 12-gram nightly dose of the high-sodium formulation would deliver roughly 2,200 milligrams of sodium, approaching the recommended daily maximum of 2,300 milligrams, whereas the same dose of the low-sodium version delivers only about 175 milligrams. That pharmacological distinction becomes clinically significant precisely because the DUET data suggest that some patients genuinely need doses above 9 grams to achieve optimal symptom control — and the sodium burden of the older formulation would have made such titration far riskier from a cardiovascular standpoint.

The study is not without limitations. It was open-label, with no control group, meaning that placebo effects and regression to the mean cannot be entirely excluded. The cohort was small — 48 participants, of whom 44 completed the study — and the criteria for enrollment were based on investigator judgment that a patient would benefit from higher doses, introducing a selection bias toward those with residual symptoms. The dosing regimen was constrained to twice-nightly administration with a maximum first dose of 7.5 grams and a maximum second dose of 4.5 grams, leaving open the question of whether other regimens, such as thrice-nightly dosing, might offer additional flexibility.

Still, the findings represent a meaningful shift in the narcolepsy treatment landscape. The 9-gram ceiling has stood for more than two decades, largely because no one had formally studied what lies beyond it. DUET demonstrates that the ceiling is not a pharmacological wall but a regulatory artifact — and that for a substantial minority of patients, the extra grams translate into fewer cataplexy attacks, less daytime sleepiness, and a disease that they no longer describe as severe. The data support what clinicians have long suspected: that narcolepsy, like the patients who live with it, is heterogeneous, and that treatment must be individualized rather than capped by a one-size-fits-all limit set in a different era of sleep medicine.

For the patients who enrolled in DUET, many of whom had lived with narcolepsy for years and had already exhausted the standard dosing options, the trial offered something that has been in short supply: permission to keep titrating. The results suggest that for those who still struggle despite maximal recommended doses, the answer may not be to add another alerting agent or accept residual symptoms, but simply to keep going — cautiously, under polysomnographic surveillance, and always in partnership with a clinician who understands that the optimal dose is the one that works for the individual patient, not the one printed on the label.

Subject of Research: Symptom improvement with dosing above 9 g of low-sodium oxybate: effectiveness and safety in participants with narcolepsy in the DUET study

Article Title: Symptom improvement with dosing above 9 g of low-sodium oxybate: effectiveness and safety in participants with narcolepsy in the DUET study

Article References: Simmons, J. H., Schneider, L. D., Ruoff, C. M., Plante, D. T., Nichols, D. A., Steininger, T. L., Whalen, M., Kirby, M. T., Dai, J., Cairns, A., & Bogan, R. K. (2026). Symptom improvement with dosing above 9 g of low-sodium oxybate: effectiveness and safety in participants with narcolepsy in the DUET study. Journal of Clinical Sleep Medicine, 22(1), Article 198. https://doi.org/10.1007/s44470-026-00193-6

Image Credits: AI Generated

DOI: 10.1007/s44470-026-00193-6

Keywords: Symptom, improvement, dosing, above, low-sodium, oxybate, effectiveness, safety, participants, narcolepsy, DUET, scientific research