Effects of Sodium Oxybate on Sleep Physiology and Sleep/Wake-related Symptoms in Patients with Fibromyalgia Syndrome: A Double-blind, Randomized, Placebo-controlled Study

Patients

Men or women 18 years of age or older who met the 1990 ACR classification criteria for FM¹ were recruited and enrolled after providing informed consent.

Inclusion criteria required that the patient’s average visual analog scale score for pain (PVAS) be > 4 on a 0 to 10 point VAS, based on patient diary records for the week prior to randomization. In addition, patients agreed to adhere to the following rules during the study: discontinue opiates, antidepressants, cyclobenzaprine, and tramadol; continue with any preexisting nonpharmacologic regimen; restrict rescue analgesic therapies to the use of acetaminophen ≤ 4000 mg/day, ibuprofen ≤ 1200 mg/day, naproxen ≤ 660 mg/day, or ketoprofen ≤ 75 mg/day; forgo ingestion of alcohol; and for women who were not surgically sterile or postmenopausal ≥ 2 years, use a medically accepted method of birth control.

Patients were excluded if any of the following were found to apply: an inflammatory rheumatic disease; a painful disorder other than FM; hyper- or hypothyroidism; a medical or psychological condition that might compromise participation in the study; an Apnea Hypopnea Index > 15 per hour on a screening PSG (exempted if using satisfactory continuous positive airway pressure therapy that controlled the apneas/hypopneas); a seizure disorder; history of head trauma resulting in loss of consciousness; migraine headaches; intracranial surgery; current or recent history (within one year) of any substance abuse disorder (including alcohol); succinic semialdehyde dehydrogenase deficiency; taking SXB or any investigational therapy in the 30 days prior to the screening visit; any past use of anticonvulsants for epilepsy; unwillingness to stop use of anticonvulsants used for pain, any antidepressant (exempted if discontinued for at least 5 half-lives), sleep aids (such as hypnotics, tranquilizers, and antihistamines; nonsedating antihistamines were exempt), or benzodiazepines; serum creatinine > 2.0 mg/dl; abnormal liver function tests (transaminase at least twice the upper limit of normal or serum bilirubin ≥ 1.5 times the upper limit of normal); a positive pregnancy test; an electrocardiogram that disclosed a clinically significant arrhythmia or an atrioventricular conduction delay greater than first-degree block; pending worker’s compensation litigation or other monetary settlements; or an occupation that required night shift work.

Clinic sites

The study was conducted at 21 clinical sites in the continental US, with personnel experienced in diagnosing and caring for patients with FM. Collectively, the healthcare providers at these sites were identified as “the Oxybate SXB-26 FM Study Group” (see Acknowledgment).

Study design and treatment

The trial was an 8-week, randomized, double-blind, placebo-controlled study of SXB in the treatment of FM. After being screened for eligibility (Visit 1) and initiation of a washout period of up to 39 days (Visit 2), eligible subjects entered a 2-week baseline period in which they recorded their PVAS scores 3 times a day (Visit 3). At Visit 4, those who had PVAS scores ≥ 4 on a 0–10 scale during the baseline period were randomized equally to one of 3 treatment groups: placebo, SXB 4.5 g/night, or SXB 6 g/night. All treatments were given nightly at bedtime and 2.5–4 hours later in 2 evenly divided doses. Treatment was 8 weeks in duration with 3 treatment visits: Visit 5 (2 weeks), Visit 6 (4 weeks), and Visit 7 (8 weeks).

At the end of the baseline period and prior to randomization at Visit 4, subjective sleep, fatigue, and daytime functioning were assessed using a series of questions eliciting sleep-related history, the Jenkins Scale for Sleep (JSS; a 4-item self-report questionnaire for sleep disturbance, including 3 questions on nighttime sleep and one on daytime tiredness, range 0–20: higher values indicate worse sleep)³⁴, the Epworth Sleepiness Scale (ESS; an 8-item self-report questionnaire for daytime sleepiness assessing the likelihood of falling asleep during different daytime situations; range 0–24: higher values indicate more sleepiness)³⁵, Fatigue VAS (FVAS; morning, afternoon, and evening daily diary; range 0–100: 0 = no fatigue, 100 = worst imaginable fatigue), and several measures of daytime functioning, including the Functional Outcome of Sleep Questionnaire (FOSQ; a series of questions evaluating the effects of daytime sleepiness on daily activities and function; range 5–20: lower values indicate worse functioning)³⁶, Short Form-36 Health Survey (SF-36) Vitality domain (evaluating the subjects’ energy level; range 0–100: higher values indicate better function, normalized value for the US is 50)³⁷, and the Fibromyalgia Impact Questionnaire (FIQ) general and morning tiredness subscales (range 0–10: higher scores indicate more tiredness)³⁸. The JSS, ESS, FVAS, FOSQ, SF-36 Vitality domain, and FIQ general and morning tiredness were collected at timepoints throughout the study and at study end.

PSG was performed at the end of screening period (Visit 2), at the randomization visit, defined as baseline (Visit 4), and at Weeks 4 and 8 of treatment (Visits 6 and 7). PSG data were centrally scored using described methods³⁹. PSG scorers were all certified PSG technologists and were required to have interrater reliability > 90%. PSG measures included sleep onset latency (SOL), wake after sleep onset (WASO), and total sleep time, all of which were measured in minutes, as well as percentage sleep efficiency, number of EEG arousals, number of apneas/hypopneas and periodic limb movements per hour of sleep, and time spent in each sleep stage. Baseline alpha-intrusion was calculated as the percentage of time in alpha frequency during the first 4 hours of sleep, where the majority of time is spent in non-REM sleep. Patients received water at bedtime and 4 hours later during Visit 2 and Visit 4 PSG as a control for study drug administration during treatment. During Visit 6 and Visit 7 PSG, patients took study drug at bedtime and 4 hours later.

CAP was also assessed using a validated computerized automatic detection methodology (Somnologica)^40,41. The total CAP rate is the sum of CAP A1 and CAP A2/3. Subgroup analyses of CAP characteristics included measures of CAP phase A1 and CAP phases A2 and A3. For technical reasons, not all PSG machines allowed data conversion to European Data Format, which is required for the calculation of the CAP. Consequently, baseline CAP data were available for 88 of 195 randomized patients and for 47 of 151 patients who completed the study.

Other self-assessed outcome measures were used in the study, including the PVAS, FIQ, SF-36 Health Survey, and Patient Global Impression of Change, and have been reported separately⁴². Clinician administered assessments included the tender point count and index, the clinician global impression of severity at baseline, and the clinical global impression of change at study end.

Tolerability and safety assessments

Safety and tolerability assessments for this study included spontaneously reported or observed treatment-emergent adverse events, 12-lead electrocardiographic measurements, clinical laboratory tests (hematology, chemistry, and urinalysis), vital sign measurements, and physical examination findings. Safety results from this study have been reported⁴² and are summarized here.

Statistical analysis

Randomization was achieved by computer-generated code using a block size of 6. Outcome measures analyzed for this report include PSG data, FVAS, JSS, ESS, SF-36 Vitality, FOSQ, and FIQ (general and morning tiredness), which were secondary endpoints of the study protocol. No adjustments for multiplicity were made. For completeness, previously reported mean and median PVAS data are also included⁴². Baseline data were summarized for all randomized patients and for the population of patients who completed the study. Patients who completed the study were exposed to SXB or placebo for 8 weeks. Changes from baseline data were analyzed on the completed population, thereby avoiding imputation of data including the PSG data. To compare the treatment groups, analysis of variance was employed with treatment as the only factor. Mean change from baseline to study endpoint (Week 8) was reported. In the case of missing data, no imputation methods were applied. All statistical testing was 2-sided with a significance level of 5%. SAS Version 9.1 for Windows (SAS, Cary, NC, USA) was used throughout. Correlations were analyzed between measures of change in PSG and change in subjective sleep and also between measures of change in subjective sleep and change in pain, using Spearman’s correlation coefficient.

General characteristics of patients at baseline

Of 320 patients who were screened, 209 underwent PSG and 195 were randomized. Of 125 patients who failed screening, the most common reasons for failure were thyroid stimulating hormone test results being out of range (24), withdrawal of consent (18), sleep apnea or Apnea Hypopnea Index out of range (16), inability to discontinue current compromising drugs (12), did not meet PVAS criteria or ACR criteria for FM (12), lost to followup (7), and other pain disorders (6). No other screening failure reason included more than 4 subjects. Of note, in the population of 209 FM patients screened using PSG, patients had an elevated incidence of maximum alpha EEG-intrusion > 24 min/hour of sleep (66%), periodic limb movements of sleep (20.1% ≥ 5/hour), and moderate to severe obstructive sleep apnea disorder (15.3%; Apnea-Hypopnea Index ≥ 15/hour).

Of the 192 treated patients, a higher percentage of patients completed the study in the placebo (82%) and SXB 4.5 g/night (82%) groups than in the SXB 6 g/night group (69%; Figure 1). More patients discontinued due to adverse events in the SXB 6 g/night group (n = 14) than in the SXB 4.5 g/night group (n = 6) or the placebo group (n = 3).

Demographic and baseline clinical characteristics were similar among the treatment groups for the randomized population. Baseline characteristics were also similar for the randomized patients and for the population that completed the study and indicated poor sleep quality (Tables 1 and 2) as well as high levels of pain, as reported by Russell, et al⁴². For the randomized population, the mean (SD) age was 46.5 (11.3) years, 94% were female, and the mean body mass index was 30.2 (7.5). A majority (74%) had experienced FM symptoms for > 5 years.

Subjective sleep, fatigue, and vitality at baseline

The self-reported sleep history at baseline was characterized by a high prevalence of sleep disturbances (Table 1), including short duration of sleep (≤ 6 hours/night; reported by 74%), awakening ≥ 3 times/night (66%), and light or very light sleep (78%).

Patients described their sleep as unrefreshing, with a mean VAS score (SD) for this question of 77.0 (19.9), suggesting poor sleep quality. Patients also reported high levels of morning and evening sleepiness, indicated by mean sleepiness VAS scores < 50, where lower numbers correspond to greater sleepiness. The mean ESS score (SD) was 11.2 (5.7), suggesting a predisposition to daytime sleepiness. Sleep disturbance was paralleled by a baseline JSS mean score (SD) of 16.6 (3.4), a FOSQ mean score of 12.0 (3.5), and an SF-36 Vitality domain mean score of 28.7 (6.0). High levels of fatigue were reported, with mean baseline FVAS scores of 69.7 (16.2) on a 0–100 scale where higher scores indicate worse fatigue.

Polysomnography

Baseline PSG data were available for 181 of the randomized patients meeting ACR criteria for FM, with CAP data analyzed in 88 patients (Table 2). The sleep architecture was consistent with the self-reported sleep history in showing delayed onset to sleep and impaired sleep efficiency. A total CAP rate of 62.7% was recorded in these patients.

Efficacy analysis

Of 320 patients screened, 195 were randomized and 192 were treated. Adverse events led to early study termination in 3 (4.5%), 6 (9.7%), and 14 (20.9%) patients from the placebo, SXB 4.5 g/night, and SXB 6 g/night groups, respectively. Reasons for termination from the study were summarized according to prespecified categories.

Efficacy

Subjective sleep and daytime fatigue and vitality

As shown in Figure 2, SXB resulted in a dose-dependent decrease in daytime fatigue measured by FVAS, with significant reductions compared with placebo noted for morning, afternoon, evening, and overall with SXB 6 g and for morning only with SXB 4.5 g.

• Download figure
• Open in new tab
• Download powerpoint

Figure 2.

Mean (± SD) change from baseline in daytime fatigue visual analog scale (VAS) in patients who completed the study. P values indicate pairwise comparison of given SXB dose and placebo.

As shown in Figure 3, both SXB 4.5 g and SXB 6 g significantly improved JSS and ESS. Dose-dependent reductions with SXB 4.5 g and SXB 6 g versus placebo were noted for both JSS and ESS. On the JSS, there were statistically significantly greater improvements with SXB 4.5 g and SXB 6 g, of −7.1 ± 5.9 (p = 0.002) and −8.4 ± 5.1 (p < 0.001), respectively, than with placebo (−3.8 ± 5.2). On the ESS, there were statistically significantly greater improvements with SXB 4.5 g and SXB 6 g, −3.1 ± 4.8 (p = 0.039) and −4.7 ± 5.1 (p < 0.001), respectively, than with placebo (−1.0 ± 5.2).

• Download figure
• Open in new tab
• Download powerpoint

Figure 3.

Mean (± SD) change from baseline on 2 subjective sleep questionnaires in patients who completed the study. P values indicate pairwise comparison of given SXB dose and placebo.

Polysomnography

PSG data were available in 138 of 151 completed patients, with CAP data analyzed in 47 patients. As shown in Table 4, compared to placebo, SXB 6 g significantly increased non-REM stage 2 sleep, SWS, and total non-REM sleep. Both SXB 4.5 g and SXB 6 g significantly reduced REM sleep time compared with placebo. The change from baseline in total sleep time and sleep efficiency was increased with SXB 6 g compared with placebo with near-statistical significance. A statistically significant reduction in WASO was noted for SXB 6 g, and an increase was noted with near-statistical significance for SXB 4.5 g. The phase A1 CAP rate showed an increase for both doses of SXB versus placebo; however, this increase was not statistically significant. For the phase A2/A3 CAP rate, both SXB doses showed a greater decrease than placebo and this change was statistically significant for the SXB 6 g group.

Correlations

Weak correlations (absolute value ≤ 0.38) were seen between most measures of change in PSG and measures of change in subjective sleep. However, for the SXB 6 g group, moderate correlations (0.56 to 0.61) were demonstrated between reduction in the phase A1 CAP rate and reductions in the JSS total score, the JSS item on trouble falling asleep, and the JSS item regarding waking up after usual amount of sleep feeling tired and worn out (n = 13).

Moderate correlations (absolute value ≥ 0.40) were demonstrated between several measures of improvement in subjective sleep (JSS total, JSS items, and the ESS total) and reductions in pain using several measures (PVAS, FIQ Pain, and the SF-36 Bodily Pain domain). For each treatment group and overall, the strongest correlations were seen between reduction in the Jenkins item regarding waking up after usual amount of sleep feeling tired and worn out and improvement on 2 pain items: FIQ Pain (≥ 0.58) and SF-36 Bodily Pain (≤ −0.53).

Tolerability and safety

Safety data for all treated patients are shown in Table 5, as reported⁴². Treatment-emergent adverse events were reported in the majority of patients receiving placebo, SXB 4.5 g, and SXB 6 g (60.0%, 68.3%, and 77.6%, respectively). The difference in these proportions was not statistically significant (p = 0.09). Two patients experienced serious adverse events: one patient experienced respiratory tract infection and exacerbation of asthma; a second patient experienced tachycardia, hypertension, and bipolar disorder. Although none were considered related to study drug or procedures by the investigators, the sponsor could not rule out a relationship to study drug for the bipolar disorder. Adverse events occurring at > 2% and differing significantly between active and placebo groups included nausea, pain in extremity, nervous system disorders, dizziness, restlessness, renal and urinary disorders, and urinary incontinence. Most of the observed adverse events were mild or moderate in severity. No clinically important changes in vital signs, laboratory measures, general examination findings, neurological examination findings, or electrocardiograms were observed.