This material may be protected by Copyright law (Title 17 U.S. Code)
`The Effects and Effectiveness of 'Y-Hydroxy butyrate in Patients
`with Narcolepsy
`MARTIN B. SCHARF, Ph.D., A.C.P., DANIEL BROWN, M.A., MARCIA WOODS, LOUIS BROWN, Ph.D.,
`and JACK HIRSCHOWITZ, M.D.
`
`Thirty patients with polysomnographically con(cid:173)
`firmed narcolepsy were treated with GliB (gamma(cid:173)
`hydroxybutyrate) for up to 30 weeks. The number of
`nightly awakenings significantly decreased, while Stages
`3 and 4 sleep substantially increased. The clinical symp(cid:173)
`toms of cataplexy, sleep paralysis, hypnogogic hallucina(cid:173)
`tions, daily naps, and sleep attacks all showed
`significant improvements. Daytime sleepiness, while not
`completely eliminated, was controlled with lower doses
`of stimulant medication than patients were taking be(cid:173)
`fore the study. No patient developed tolerance to the
`drug, and no serious side effects were noted.
`(J Clin Psychiatry 46:222-225, 1985)
`
`Narcolepsy, a major sleep disorder which is reported to
`affect between 100,000 and 200,000 Americans; has re(cid:173)
`cently attracted increased public attention. The average de(cid:173)
`lay between onset of narcoleptic symptoms and a correct
`l diagnosis has been reported to be approximately 15 years. 2
`I' The objective nature of sleep laboratory evaluations has
`i eliminated much of the ambiguity from the diagnostic pro-
`cedures for this disorder, so thai diagnoses are being made
`more quickly. Improvements in treatment also have re(cid:173)
`cently begun to catch up with advancements in diagnosis.
`Since 1976, Broughton and Mamelak3-6 have been re(cid:173)
`porting on the successful use of 'Y-hydroxybutyrate (GHB)
`in the treatment of Canadian patients with narcolepsy. A
`recent report described results in 29 narcoleptic patients
`who were treated with GHB for 3 months to 6 years. 6 These
`data, while clear-cut, are primarily anecdotal in nature and
`do not provide quantitative measures of changes in specific
`narcoleptic symptoms. The present study reports on the
`first clinical trials of GHB in the United States and was
`conducted to confirm and extend the results of the Canadian
`studies.
`
`METHOD
`
`Thirty patients (17 women, 13 men) between the ages of
`27 and 65 years (mean=45.7±9.6) have completed up to
`
`From the Sleep Disorders Laboratory of The Jewish Hospital, Cincin·
`nati (all authors}, the Depanmenr of Otolaryngology and Maxil/ofocial
`Surgery (Dr. Scharf), and the Depanment of Psychiatry, University of Cin(cid:173)
`cinnati School of Medicine (Drs. Scharf and Hirschowitz).
`Reprint requests to: Manin B. Scharf, Ph.D., Director, Sleep Disor(cid:173)
`ders laboratory of The Jewish Hospital, 515 Melish Avenue, Cincinnati,
`OH 45229.
`
`30 weeks of the study to date. Each patient has a history of
`polysomnographically confirmed narcolepsy, a clinical
`complaint of excessive daytime sleepiness, and at least one
`of the auxiliary symptoms of narcolepsy (cataplexy, sleep
`paralysis, and hypnogogic hallucinations). Before entering
`the study, 17 of the patients had been treated with a tricyclic
`antidepressant (TCA) to control cataplexy and 24 were tak(cid:173)
`ing stimulant medications to control daytime sleepiness; 14
`patients were receiving both a TCA and stimulant medica(cid:173)
`tion. Tricyclic and stimulant drugs and dosages used by the
`patients on baseline and following GHB ~reatm~nt are listed
`in Tables 1 and 2.
`The protocol and investigative nature of the study were
`carefully explained to each patient and informed consent
`was obtained. A review of their history and a physical ex(cid:173)
`amination were performed, and the following laboratory
`tests were completed: hemogram, liver survey, renal sur(cid:173)
`vey, chest x-ray, EEG, and ECG. Each patient was also
`psychologically evaluated with the Minnesota Multiphasic
`Personality Inventory (MMPI). Baseline polysomnographic
`data were collected for 2 nights and a multiple sleep latency
`test (MSLT) consisting of five nap opportunities was con(cid:173)
`ducted. GHB was administered nightly beginning on the
`third night in two doses of 20-25 cc of orange-flavored fluid
`containing .15 gm/cc GHB. The first dose was adminis(cid:173)
`tered at lights out, and the second was administered 4 hours
`later. Patients were required to stay in bed for 8 hours.
`Some woke up again after the second dose and required a
`third dose (10 cc) to go back to sleep. Total nightly dosage
`ranged from 5 to 7 grams of GHB.
`During the first week of GHB administration, the 17
`patients taking a TCA to control auxiliary narcoleptic
`symptoms took half their usual dosage. Beginning on the
`second week, they were withdrawn from their TCA at the
`rate of one clinical dose per week until total withdrawal
`occurred. All patients were administered stimulant medica(cid:173)
`tion as needed during the day to combat excessive daytime
`sleepiness (Table 1). Dosages were titrated on an individual
`basis. Patients were evaluated in the sleep laboratory on
`Nights 1 and 2 and again at the end of 4 weeks. Throughout
`the study, each patient also completed daily pre- and
`postsleep questionnaires which subjectively surveyed sleep
`latency, the number of nightly awakenings, hours of sleep,
`sleep attacks, cataplectic events, hypnogogic hallucina(cid:173)
`tions, sleep paralysis, and medications taken.
`Data from the daily questionnaires of 29 patients were
`collected and averaged for the baseline period and each suc(cid:173)
`cessive treatment week. Comparisons to baseline were
`
`222
`
`Ranbaxy Ex. 1012
`IPR Petition - USP 8, 772,306
`
`

`
`J CLIN PSYCHIATRY 46:6- JUNE 1985
`
`SCHARF ET AL.
`
`TABLE 1. Stimulant Usage and Dosage Changes From Baseline to GHB Administration Period
`Baseline
`Dosage (mg)
`Mean
`20.00
`52.86
`34.38
`21.25
`20.00
`
`Drug
`Methylphenidate SR
`Methylphenidate
`Pemoline
`Dextroamphetamine
`Methamphetamine
`None
`
`Number
`of Patients
`1
`7
`6
`8
`2
`6
`
`so
`
`40.30
`7.65
`9.16
`14.14
`
`GHB Administration Period
`Dosage (mg)
`Mean
`20.00
`11.25
`
`14.17
`26.67
`
`so
`0.00
`1.77
`
`10.57
`20.21
`
`N
`12
`2
`0
`6
`2
`8
`
`TABLE 2. Tricyclic Usage and Mean Dosage Prior to GHB (Base(cid:173)
`line)
`
`Number
`of Patients
`Drug'
`Imipramine
`8
`Protriptyline
`6
`Amitriptyline
`2
`Chlorimipramine
`1
`14
`None
`'One patient was taking imipramine and amitriptyline concurrently
`prior to entering the study.
`
`Dosage (mg)
`SO
`Mean
`36.70
`51.67
`4.18
`17.50
`35.36
`125.00
`50.00
`
`TABLE 3. Predrug (Baseline) Means Versus Week 4 on GHB in
`Narcoleptic Patients (N =30) Based on Polysomnagraphic Re-
`cordings
`
`Baseline
`so
`Mean
`
`Week4
`so
`Mean
`
`406.23"
`
`TABLE 4. Predrug (Baseline} Means Versus 6 Months on GHB in
`Narcoleptic Patients (N = 12) Based on Polysomnographic Re·
`cordings
`
`Baseline
`so
`Mean
`
`6 Month
`so
`Mean
`
`387.78
`
`35.11
`
`407.78
`
`28.48
`
`90.00
`
`37.83
`
`74.02'
`
`28.71
`
`81.00
`17.08
`
`15.38
`
`51.54
`
`2.25
`
`12.63
`
`7.79
`5.08
`
`5.33
`
`6.00
`
`4.05
`
`5.51
`
`84.58'
`9.27b
`
`14.00
`
`56.38
`
`4.42
`
`12.00
`
`6.06
`3.95
`
`5.07
`
`8.81
`
`6.36
`
`3.39
`
`86.63
`
`87.85
`
`32.36'
`
`39.37
`
`Sleep Variable
`Total sleep time
`(in minutes)
`Total wake time
`(in minutes)
`Sleep efficiency
`(in percents)
`Number of awakenings
`Percent of
`Stage 1
`Percent of
`Stage 2
`Percent of
`Stages 3 and 4
`Percent of
`Stage REM
`Latency to REM
`(in minutes)
`MSLT
`Sleep latency
`(in minutes)
`Latency to REM
`(in minutes)
`'p < .05.
`"p < .005.
`'p< .025.
`
`4.55
`
`8.37
`
`3.71
`
`4.99
`
`5.79
`
`13.36'
`
`5.14
`
`5.55
`
`RESULTS
`
`Sleep Architecture
`Table 3 shows polysomnographic data of 30 patients for
`selected sleep parameters at baseline and at the end of 4
`consecutive weeks on GHB. GHB administration resulted
`in a moderate increase in total sleep time and a moderate
`qecrease in total wake time, which was significant at
`p < .025 and p < .05, respectively.
`The number of awakenings decreased from 14.57 to
`9.48 (p < .005), and the percent of Stages 3 and 4 sleep
`increased from 1.48 at baseline to 4. 53 at Week 4
`(p < . 005). The percent of REM sleep increased slightly
`(N.S.), and REM latency decreased from 102.88 minute~ at
`baseline to 48.0 after 4 weeks of GHB (p < .005). During
`the MSLT, however, REM latency increased significantly,
`from 9.2 to 14.4 minutes (p < .005), while sleep latencY
`increased from 3.7 to 5.2 minutes (N.S.).
`Table 4 shows polysomnographic data for 12 patients
`who had been on GHB for at least 6 months. The results are
`
`Sleep Variable
`Total sleep time
`(in minutes)
`Total wake time
`(in minutes)
`Sleep efficiency
`(in percents)
`Number of awakenings
`Percent of
`Stage 1
`Percent of
`Stage 2
`Percent of
`Stages 3·4
`Percent of
`Stage REM
`Latency to REM
`(in minutes)
`MSLT
`Sleep latency
`(in minutes)
`Latency to REM
`(in minutes)
`'p< .025.
`•p< .05.
`'p< .005.
`
`377.06
`
`70.18
`
`30.87
`
`94.35
`
`46.40
`
`74.77b
`
`31.59
`
`81.18
`14.57
`
`16.50
`
`51.38
`
`1.48
`
`8.17
`4.66
`
`5.60
`
`8.33
`
`2.88
`
`B4.42b
`9.48'
`
`14.50
`
`54.30
`
`4.53'
`
`11.92
`
`5.17
`
`12.77
`
`6.54
`3.48
`
`6.61
`
`8.57
`
`4.95
`
`5.00
`
`102.88
`
`80.88
`
`48.04<
`
`68.76
`
`3.70
`
`9.22
`
`2.91
`
`5.22
`
`4.42
`
`14.43'
`
`4.24
`
`5.58
`
`made for means ofWeeks 1, 3, 6, 9, 12, 15, 18, 21, 27, and
`30 for those patients reaching the respective treatment
`weeks. Thus, comparisons to baseline were made for 29
`patients for treatment Weeks 1 and 3, but only for 3 patients
`for Week 30. One patient did not complete any subjective
`reports throughout the study. Statistical significance was
`tested with the Wilcoxon signed-ranks test; data from each
`treatment time- point were compared to each patient's own
`baseline.
`
`223
`
`

`
`r J CLIN PSYCHIATRY 46:6- JUNE 1985
`
`'Y-HYDROXYBUTYRATE IN NARCOLEPSY
`
`TABLE 5. Long-Term Effects of GHB on the Narcoleptic Tetrad Changes From Baseline Daily Means
`Daytime
`Hypnogogic
`Sleep Attacks'
`Hallucinations'
`so
`so
`N
`Mean
`Mean
`week
`29
`3.22
`2.59
`0.61
`0.76
`Baseline
`1.15
`29
`0.96
`0.12
`0.29
`Week 1
`29
`1.10
`0.95
`0.10
`0.27
`week 3
`23
`0.96
`0.84
`0.08
`0.18
`week 6
`21
`0.78
`1.01
`0.10
`0.27
`week 9
`19
`0.80
`0.74
`0.04
`0.11
`week 12
`15
`0.79
`0.76
`0.07
`0.25
`week 15
`14
`0.86
`0.66
`0.02
`0.08
`Week 18
`0.81
`13
`0.64
`0.00
`Week 21
`11
`0.74
`0.75
`0.00
`week 24
`'All differences from baseline significant at p < .005.
`
`Cataplexy'
`so
`Mean
`3.91
`6.75
`1.66
`2.50
`1.29
`1.80
`0.87
`0.58
`0.59
`0.79
`0.37
`0.74
`0.43
`1.01
`0.24
`0.16
`0.21
`0.34
`0.12
`0.19
`
`Sleep
`Paralysis'
`Mean
`0.47
`0.09
`0.04
`0.04
`0.10
`0.04
`0.00
`0.00
`0.00
`0.00
`
`so
`0.76
`0.22
`0.12
`0.13
`0.25
`0.11
`
`FIGURE 1. Effects of GHB on the Auxiliary Symptoms of Narco(cid:173)
`lepsy•
`"
`
`'"
`
`"'
`
`1>.\SU..IN~
`n-2'1-
`
`liU.K I
`orr=2.')
`
`\>"ED( l
`
`WE:O\ 6
`n=~J
`
`WEr.R l2
`r=l-9
`
`1i1:D1 l'i WED< 111
`n=l5
`..-14
`
`*Shown as percent improvement from baseline.
`
`similar to those at the 4-week time point for the whole
`group.
`
`Narcoleptic Symptoms
`The data show a statistically significant reduction in all
`symptoms, both NREM (sleep attacks and naps) and REM(cid:173)
`related symptoms (cataplexy, sleep paralysis, and hypno(cid:173)
`gogic hallucinations). The number of daytime naps
`decreased an average of 48% by Week 3 from 1. 89 ± 1. 68
`per day to 0.98±0.68 (p < .005). This was despite the fact
`that we requested patients to take at least one nap every day.
`The number of sleep attacks decreased by 64% (p < .005) in
`the first week of treatment (Figure 1 and Table 5). The av(cid:173)
`erage for sleep attacks remained relatively constant and by
`Week 9 was sti1169% below the baseline (p< .005).
`As can be seen in Figure 1, cataplectic events decreased
`an average of 58% in the first week (p < .005). By Week 9
`the total average improvement from baseline was 85%
`(p < .005). Cataplexy decreased steadily from week to
`week regardless of whether the patients had or had not been
`on tricyclic medication previously.
`The symptoms of sleep paralysis and hypnogogic hallu(cid:173)
`cinations also showed marked improvement (see Figure 1).
`Sleep paralysis decreased an average of 81% from baseline
`by the first week (p < .005); 91% by Week 3 (p < .005); and
`was completely absent in all subjects by Week 15
`
`(p < .005). There was an 80% decrease in the number of
`hypnogogic hallucinations during the first week of the study
`(p < .005); 84% by Week 3 (p < .005); 83% by Week 12
`(p < .005); and 97% by Week 18 (p < .005).
`Side effects noted during the course of the study in(cid:173)
`cluded a single episode of protracted sleep paralysis which
`occurred in 3 patients. These were experienced by the pa(cid:173)
`tients as extremely frightening. Each episode took place
`shortly after the initial nightly dose of GHB. All 3 episodes
`occurred within the first 2 weeks of the study. 1\vo of the
`episodes occurred within the laboratory and were witnessed
`by the staff. In these cases, the patients experienced inter(cid:173)
`mittent cataplexy, i.e., sleep paralysis, which lasted almost
`1 hour.
`Other side effects included one instance of drug-related
`enuresis and one complaint of increased transient sexual
`drive. Some patients had difficulty staying asleep after the
`second nightly dose of GHB. This problem was resolved by
`splitting the dosage for these patients into thirds and admin(cid:173)
`istering it three times instead of twice. There were no sub(cid:173)
`sequent difficulties with these patients' ability to sleep.
`
`DISCUSSION
`
`The usual treatment for narcolepsy includes sympto(cid:173)
`matic treatment of daytime drowsiness and sleep attacks
`with stinmlants such as dextroamphetamines, methylpheni(cid:173)
`date, or pemoline, while the auxiliary symptoms (cataplexy,
`sleep paralysis, and hypnogogic hallucinations) are treated
`with tricyclic antidepressants such as imipramine or pro(cid:173)
`triptyline. These treatment modalities are often unsatisfac(cid:173)
`tory for a number of reasons. Amphetamines and other
`stimulants can cause undesirable side effects, including in(cid:173)
`somnia, hypertension, palpitations, and, at higher doses,
`may mimic symptoms of schizophrenia. Since tolerance
`frequently develops, the dosages must be increased. This,
`in tum, often leads to an increase in frequency and severity
`of side effects. Tricyclic antidepressants also can cause un(cid:173)
`desirable side effects, including dry mouth, impotence, loss
`of libido, tachycardia, and others. In addition, they are
`somewhat cardiotoxic and can exacerbate or cause conduc(cid:173)
`tion disturbance, heart block, or bundle branch block. Fi-
`
`224
`
`

`
`J CLIN PSYCHIATRY 46:6- JUNE 1985
`
`nally, their concurrent use with stimulants may increase risk
`excessively in patients with hypertension.
`Our results confirm those of previous clinical studies'·"
`of GHB, namely, that it is a safe, nontoxic substance that is
`effective in the treatment of narcolepsy. GHB administra(cid:173)
`tion results in an increase in slow-wave sleep and does not
`suppress REM sleep. In addition, there was no evidence of
`drug tolerance after 24 weeks of treatment. The improve(cid:173)
`ments we found in sleep attacks, daytime drowsiness, cata(cid:173)
`plexy, hypnogogic hallucinations, and sleep paralysis are
`consistent with findings of Broughton and Mamelak4 in
`their 12-month study of 16 narcoleptic patients.
`Although GHB is not purported to be a primary treat(cid:173)
`ment for daytime sleepiness, our data showed a reduction in
`both the number of naps taken and the number of sleep
`attacks. These changes occurred despite the fact that pa(cid:173)
`tients were taking less stimulant medication than before and
`were encouraged to take at least one nap every day. Narco(cid:173)
`leptics often have frequent disruptions in nocturnal sleep.
`The improvement in nocturnal sleep consolidation as seen
`in the decreased number of awakenings may have contrib(cid:173)
`uted to decreased daytime sleepiness. Despite the decreases
`in naps and sleep attacks, daytime drowsiness persisted. It
`was, however, milder than pretreatment levels. None of the
`patients were able to withdraw completely from stimulant
`medication during the drug period. Twenty-two patients re(cid:173)
`quired stimulants daily, but at lowered dosages. The re(cid:173)
`maining 8 patients were able to function well without stimu(cid:173)
`lants part of the time, but needed mild doses at other times.
`The side .effects experienced by our patients were pre(cid:173)
`dictable and self-limiting. They were consistent with the
`hypothesized explanation oftbe drug's action7 and had been
`previously reported by Broughton and Mamelak. 4 In the
`few cases where side effects occurred, they were rated by
`the patients as less bothersome than the side effects of pre(cid:173)
`vious narcolepsy medications. No patient expressed a de(cid:173)
`sire to discontinue treatment with GHB.
`Broughton and Mamelak5 performed continuous 48-
`hour polysomnographic recordings of sleep/waking pat(cid:173)
`terns on 14 narcoleptic patients before and after 7-10 days
`of GHB administration. They found that GHB improved the
`quality of night sleep by increasing the amount of Stages 3
`and 4 sleep, reducing Stage 1 sleep, increasing sleep effi(cid:173)
`ciency, and reducing the number of short sleep periods (less
`than 15 minutes). They also found that nighttime REM
`sleep was reduced in latency and became less fragmented.
`All of these changes were statistically significant. Our re(cid:173)
`sults were consistent with those of Broughton and Mamelak.
`Polysomnographic data have also been reported by Ma(cid:173)
`melak et al. 8 for a study of GHB in the treatment of insom(cid:173)
`niac patients. Again, it was found that Stages 3 and 4 sleep
`were significantly increased and REM latency significantly
`decreased.
`The patients in this stttdY had serious cases of narco(cid:173)
`lepsy and were experiencing several attacks of cataplexy
`per day prior to the study despite the use of a TCA in 17
`
`225
`
`SCHARF ET AL.
`
`cases. Thus, the effects of GHB in this study are contrasted
`to a "treated" baseline condition. When REM suppressing
`agents are taken for protracted periods, the withdrawal
`REM rebound is usually prolonged, lasting several weeks!
`This can be debilitating in narcoleptic patients because it is
`often accompanied by increased cataplexy, hypnogogic hal(cid:173)
`lucinations, and sleep paralysis. Broughton and Mamelak:·'
`in their studies of 16 narcoleptics, withdrew patients from
`all previous drug treatment for at least 14 days before initi(cid:173)
`ating GHB administration. The treatment strategy for this
`study was to overlap GHB administration with TCA with(cid:173)
`drawal to minimize the marked REM rebound that typically
`occurs. As a result of the TCA withdrawal, the nature of
`cataplectic attacks changed somewhat. During baseline ob(cid:173)
`servations attacks were predictable, usually occurring in
`conjunction with emotional stimuli such as anger or laugh(cid:173)
`ter. While the TCA patients were in the rebound phase,
`some cataplectic attacks appeared spontaneously without a
`precipitating emotional arousal. The number of attacks
`gradually diminished, presumably as the REM rebound ef(cid:173)
`fects dissipated.
`Behavioral changes were noted in several patients dur(cid:173)
`ing the course of the study. During the baseline period they
`carefully avoided situations that might induce cataplexy.
`This seemed to be a well developed self-protective mechan-
`. ism. During treatment with GHB, however, patients began
`testing the drug's limits by exercising less emotional re(cid:173)
`straint and in .some cases seeking out situations that nor(cid:173)
`mally induced cataplexy. This behavioral change did not
`lead to an increase in cataplexy. In fact, some patients found
`they could not purposely "induce" cataplexy unless they
`were fatigued.
`Our results to date clearly support the efficacy and su(cid:173)
`periority of GHB compared to previous treatments for the
`treatment of narcolepsy. Unequivocal efficacy, however,
`can only be established by double-blind placebo studies,
`which are currently underway in our laboratory and will be
`reported at a later date.
`
`REFERENCES
`1. Dement WC, Carskadon M, Ley R: The prevalence of narcolepsy. ll.
`Sleep Res 2:147, 1973
`2. Ferriss GS: Narcolepsy. Continuing Education. May 1982, pp 41-48
`3. Broughton R, Mame1ak M: Gamma-hydroxybutyrate in the treatment
`uf <.:ompound narcolepsy: A preliminary report. In Guilleminault C,
`Dement WC, Passouant P (eds): Narcolepsy. New York, Spectrum,
`1976
`4. Broughton R, Mamelak M: The treatment of narcolepsy-cataplexy with
`nocturnal gamma-hydroxy butyrate. Can J Neural Sci 6:1-6, 1979
`5. Broughton R, Mamelak M: Effects of nu<.:tumal gamma-hydroxybu(cid:173)
`tyrate on sleepfwaking patterns in narcolepsy-cataplexy. Can J Neural
`Sci 7:23-30, 1980
`6. Mamelak M: The treatment of narcolepsy with gamma-hydroxybu(cid:173)
`tyrate. Sleep Res (in press)
`7 •. Vickers MD: Gamma-hydroxybutyric acid. Int Anesthesia! Clin 7:75-
`89, 1969
`8. Mamelak M, Escriu JM, Stokan 0: The effects of gammahydrm<.ybu-
`.
`tyrate on sleep. Biol Psychiatry 12:273-288, 1977
`9. Dunleavy DLF, Brezinova V, Oswald I, et a!: Changes during weeks ID
`effects of tricyclic drugs on the human sleeping brain. Br J Psycbiatr)'
`120:663-672, 1972