Psychopharmacology(1992) 108:67-71
Psychopharmacology © Springer-Verlag 1992
Rebound insomnia in normals and patients with insomnia after abrupt and tapered discontinuation Timothy Roehrs, Lori Merlotti, Frank Zorick, and Thomas Roth Henry Ford Hospital, 2921 West Grand Boulevard, Sleep Disorders and Research Center, Detroit, MI 48202, USA Received January 12, 1991 / Final version November 29, 1991
Abstract. Rebound insomnia was studied in subjects, aged 2 5 - 5 0 years, with insomnia complaints and normal sleep, insomnia complaints and disturbed sleep, and normal sleep with no complaints (N= 21, n = 7 per group). Standard sleep recordings were collected on a baseline night and after abrupt discontinuation of 6 nights of 0.50 mg triazolam, tapered discontinuation (3 nights of 0.50 rag, 2 nights of 0.25 mg, and 1 night of 0.125 mg triazotam) and 6 nights of placebo. Significantly disturbed sleep on the discontinuation night compared to the baseline night was found. The relative degree of rebound insomnia was greater in the abrupt condition than in either the tapered or placebo conditions. The tapered condition reduced sleep time by half that of the abrupt condition which was twice the reduction found in the placebo condition. An overall (regardless of group or condition) difference in baseline versus discontinuation sleep was found, suggesting that pill discontinuation itself leads to sleep disturbance. Subjects did not differ in rebound insomnia as a function of pre-existing sleep disturbance. Key words: Insomnia - Triazolam - Individuality - Substance withdrawal syndrome - Sleep - Rebound insomnia
Rebound insomnia, the sleep disturbance that follows discontinuation of hypnotic drugs, is characterized by a worsening of sleep beyond basal levels which endures for 1-2 nights. It was initially reported as an observation in patients with insomnia who were participating in efficacy studies of benzodiazepine hypnotics (Bixler et al. 1978; Kales et al. 1978). Studies to determine whether rebound insomnia is primarily a clinical phenomenon or whether it has a pharmacological basis, showed that rebound insomnia does occur in healthy normals and furthermore, that it is associated with dose (Vela-Bueno et al. 1983; Roehrs et al. 1986). Studies comparing multiple Offprint requests to: T. Roehrs
doses have found that the nature of the dose relation may depend upon the comparison used to assess rebound insomnia. In comparison to a parallel placebo control on a "pseudo" discontinuation night, rebound insomnia appears to be linearly related to dose, while in comparison to the subject's baseline sleep it appears dose specific, that is, occurring only with high doses (Roehrs et al. 1990a). Another important determinant of rebound insomnia is drug half-life. Rebound insomnia has been reported following all short- and intermediate-acting compounds, but not long-acting compounds (Kales et al. 1982). It is felt that rebound insomnia occurs because of the rapid elimination of the short- and intermediate-acting compounds which then leaves the CNS acutely deficient in inhibitory neuronal mechanisms (Kales et al. 1983). A recent study using subjective measures of sleep and a basal sleep comparison has shown that a gradual reduction of the dose of a short-acting benzodiazepine (i.e. the equivalent of the slow elimination that occurs with a long-acting drug) avoids the rebound insomnia seen with the abrupt discontinuation of a high dose of that compound (Greenblatt et al. 1987). This study sought to replicate and extend the finding of rebound insomnia with abrupt, but not tapered, discontinuation using objective measures of sleep and using a parallel placebo control. While rebound insomnia clearly has a pharmacological basis, a recent study of healthy normals found that there are also individual differences in rebound insomnia (Merlotti et al. 1991). Some subjects consistently experienced rebound insomnia while others consistently did not. Differences in basal sleep and response to treatment differentiated these two groups of subjects. Compared to subjects without rebound insomnia, subjects with rebound insomnia had poorer basal sleep, albeit still within what is generally considered a normal range (i.e. sleep efficiencies greater than 85%). The question arises as to whether rebound insomnia is more severe in persons with objectively disturbed sleep (i.e. sleep efficiencies of less than 85%) and insomnia complaints. In addition, while most patients with insomnia complaints have ob-
68 jectively d i s t u r b e d sleep, a m i n o r i t y o f i n s o m n i a p a t i e n t s have objectively n o r m a l sleep (Zorick et al. 1984). Also of interest then, is h o w r e b o u n d i n s o m n i a relates to the presence o f a sleep c o m p l a i n t regardless o f the q u a l i t y o f sleep. T h u s a n o t h e r p u r p o s e of the present study was to directly c o m p a r e r e b o u n d i n s o m n i a i n p a t i e n t s with ins o m n i a a n d in h e a l t h y n o r m a l s . T h e d a t a presented in this p a p e r are derived f r o m a larger s t u d y o f benzodiazepine h y p n o t i c self a d m i n i s t r a t i o n . The b e h a v i o r a l self a d m i n i s t r a t i o n d a t a are presented in a c o m p a n i o n p a p e r (Roehrs et al. 1991).
Materials and methods
Subjects. Twenty-one men and women, 25-50 years of age, were studied. All were in good health (except for an insomnia complaint) as determined by the screening procedures described below. All subjects signed a written, informed consent and received payment for their participation. Screening. Subjects underwent a medical history, physical examination, and laboratory blood and urine tests prior to a screening nocturnal polysomnogram. Subjects were excluded if they had acute or chronic medical conditions that required treatment or were currently taking CNS-acting drugs. Urine screens were used to document the absence of current drug use (licit or illicit) and breathalyzer tests for alcohol use. Subjects with a past history of psychiatric disorders, drug addiction, or alcoholism were excluded. Finally, subjects who had taken benzodiazepine hypnotics within the past year were also excluded. Each subject underwent a sleep disorders evaluation including a sleep history and a nocturnal polysomnogram. Three groups of subjects (n = 7 per group) were selected based on the sleep history and polysomnogram (see Table 1 for group descriptions). Subjects recruited as normals (NOR) reported no history of insomnia, estimated their nightly sleep time as greater than 7 h, and had a sleep efficiency on the nocturnal polysomnogram of 90% or greater. Subjects with insomnia (DIMS) had the complaint for at least 1 year, estimated their nightly sleep time as less than 6.5 h, and had a sleep efficiency of 85% or less on the screening polysomnogram. Subjects qualifying as persons with insomnia complaints, but no objective sleep disturbance (NOF), had a 1 year history of insomnia, estimated their nightly sleep time as less than 6.5 h, and had a sleep efficiency of 90% or more on the polysomnogram while estimating that their sleep time for the screening night was at least 1 h less. All subjects had no evidence of apneas or periodic leg movements on the nocturnal polysomnogram. Procedure. Each subject who passed the screening was exposed to three different treatments organized in a Latin Square design and administered in a double-blind fashion with 7 nights between treatments. Each treatment consisted of 1 adaptation night and 1 baseline night with no pill administration then 6 nights of 0.5 mg
triazolam (abrupt condition), or 3 nights of 0.5 mg, 2 nights of 0.25 mg, and 1 night of 0.125 mg triazolam (taper condition), or 6 nights of placebo. Following each treatment night was 1 night with no pill administration to test for the presence of rebound insomnia. All capsules were matched for appearance and given to the subject 30 min prior to lights out. Subjects arrived at the sleep laboratory about 1.5 h before their usual bedtime which was held constant throughout the study. They were prepared for standard polysomnographic assessment (Rechtschaffen and Kales 1968). A standard all-night polysomnogram was obtained on the 1 adaptation, and the 1 baseline night, the first, third, and last night of each treatment and on the discontinuation night. The polysomnogram included the central and occipital electroencephalograms (EEG), electrooculograms (EOG), and submental electromyogram (EMG) collected from standard placements and monitored continuously for 8 h. On the diagnostic screening night the recording also included nasal and oral respiration measured with thermistors and bilateral leg (anterior tibialis) electromyograms, Subjects were awakened after 8 h in bed (960 pages from lights out). After arising, each subject then completed a sleep questionnaire regarding the quantity and quality of the previous night's sleep. Recordings were made using Grass model 78-D or Nihon Kohden (models 4312 and 4212) polygraphs. The Grass polygraph was calibrated with a pen deflection of 50 gV = 7.5 mm for the EEG and EOG, and 50 pV= 10.0 mm for the EMG. The 0.5 amp low frequency filter was set at 0,3 with a sensitivity at 5 for the EEG and EOG and 10 with a sensitivity of 1 or the EMG. The 0.5 amp frequency high filter was set at 90. The Nihon Kohden machines were calibrated at 50 ~tV= 10.0 mm for the EEG and EOG and 50 gV = 16.5 mm for the EMG. The 0.5 amp low frequency filters were set at 0.3 with the sensitivity at 5 for the EEG and EOG and 0.003 with a sensitivity of 1 for the EMG. The 0.5 amp frequency high filters were set at 70. All electrode impedances were less than 10000 ohms and paper speed was 10 mm/s. The following study restrictions were adhered to by all subjects: 1) no alcoholic or caffeinated beverages after 4:00 p.m. on study nights, 2) no napping during the study, 3) no changes in bed or wake times during the study, 4) 8 h in bed each night during the study and 5) no medications without the approval of the investigator. Each polysomnographic recording was scored manually in 30-s epochs according to the standards of Rechtschaffen and Kales (1968), maintaining an inter-rater reliability of 90% or better. Records were coded so that scorers were unaware of the treatment. Mixed design MANOVAs (SAS Institute) were conducted on the various polysomnographic parameters with group as the between subject factor and treatment condition and night as the within subject factors. Conservative probability levels corrected by Greenhouse-Geisser procedure were used for all within subject comparisons. Two sets of MANOVA analyses were conducted. The first was conducted on the results of the 3 pill administration nights to establish the pharmacological activity of the drug in our methodology. The second set of analyses comparing baseline nights to discontinuation nights assessed the presence of rebound insomnia.
Results Table 1. Characteristics of study groups Group
NOR
NOF
DIMS
Age Sex females males Insomnia Hx SubjTST CPSG SE
25-50 4 3 Absent > 7h >90%
25-50 3 4 _> 1 yr < 6.5h >90%
25-50 4 3 _> 1 yr < 6.5h
Psychopharmacology © Springer-Verlag 1992
Rebound insomnia in normals and patients with insomnia after abrupt and tapered discontinuation Timothy Roehrs, Lori Merlotti, Frank Zorick, and Thomas Roth Henry Ford Hospital, 2921 West Grand Boulevard, Sleep Disorders and Research Center, Detroit, MI 48202, USA Received January 12, 1991 / Final version November 29, 1991
Abstract. Rebound insomnia was studied in subjects, aged 2 5 - 5 0 years, with insomnia complaints and normal sleep, insomnia complaints and disturbed sleep, and normal sleep with no complaints (N= 21, n = 7 per group). Standard sleep recordings were collected on a baseline night and after abrupt discontinuation of 6 nights of 0.50 mg triazolam, tapered discontinuation (3 nights of 0.50 rag, 2 nights of 0.25 mg, and 1 night of 0.125 mg triazotam) and 6 nights of placebo. Significantly disturbed sleep on the discontinuation night compared to the baseline night was found. The relative degree of rebound insomnia was greater in the abrupt condition than in either the tapered or placebo conditions. The tapered condition reduced sleep time by half that of the abrupt condition which was twice the reduction found in the placebo condition. An overall (regardless of group or condition) difference in baseline versus discontinuation sleep was found, suggesting that pill discontinuation itself leads to sleep disturbance. Subjects did not differ in rebound insomnia as a function of pre-existing sleep disturbance. Key words: Insomnia - Triazolam - Individuality - Substance withdrawal syndrome - Sleep - Rebound insomnia
Rebound insomnia, the sleep disturbance that follows discontinuation of hypnotic drugs, is characterized by a worsening of sleep beyond basal levels which endures for 1-2 nights. It was initially reported as an observation in patients with insomnia who were participating in efficacy studies of benzodiazepine hypnotics (Bixler et al. 1978; Kales et al. 1978). Studies to determine whether rebound insomnia is primarily a clinical phenomenon or whether it has a pharmacological basis, showed that rebound insomnia does occur in healthy normals and furthermore, that it is associated with dose (Vela-Bueno et al. 1983; Roehrs et al. 1986). Studies comparing multiple Offprint requests to: T. Roehrs
doses have found that the nature of the dose relation may depend upon the comparison used to assess rebound insomnia. In comparison to a parallel placebo control on a "pseudo" discontinuation night, rebound insomnia appears to be linearly related to dose, while in comparison to the subject's baseline sleep it appears dose specific, that is, occurring only with high doses (Roehrs et al. 1990a). Another important determinant of rebound insomnia is drug half-life. Rebound insomnia has been reported following all short- and intermediate-acting compounds, but not long-acting compounds (Kales et al. 1982). It is felt that rebound insomnia occurs because of the rapid elimination of the short- and intermediate-acting compounds which then leaves the CNS acutely deficient in inhibitory neuronal mechanisms (Kales et al. 1983). A recent study using subjective measures of sleep and a basal sleep comparison has shown that a gradual reduction of the dose of a short-acting benzodiazepine (i.e. the equivalent of the slow elimination that occurs with a long-acting drug) avoids the rebound insomnia seen with the abrupt discontinuation of a high dose of that compound (Greenblatt et al. 1987). This study sought to replicate and extend the finding of rebound insomnia with abrupt, but not tapered, discontinuation using objective measures of sleep and using a parallel placebo control. While rebound insomnia clearly has a pharmacological basis, a recent study of healthy normals found that there are also individual differences in rebound insomnia (Merlotti et al. 1991). Some subjects consistently experienced rebound insomnia while others consistently did not. Differences in basal sleep and response to treatment differentiated these two groups of subjects. Compared to subjects without rebound insomnia, subjects with rebound insomnia had poorer basal sleep, albeit still within what is generally considered a normal range (i.e. sleep efficiencies greater than 85%). The question arises as to whether rebound insomnia is more severe in persons with objectively disturbed sleep (i.e. sleep efficiencies of less than 85%) and insomnia complaints. In addition, while most patients with insomnia complaints have ob-
68 jectively d i s t u r b e d sleep, a m i n o r i t y o f i n s o m n i a p a t i e n t s have objectively n o r m a l sleep (Zorick et al. 1984). Also of interest then, is h o w r e b o u n d i n s o m n i a relates to the presence o f a sleep c o m p l a i n t regardless o f the q u a l i t y o f sleep. T h u s a n o t h e r p u r p o s e of the present study was to directly c o m p a r e r e b o u n d i n s o m n i a i n p a t i e n t s with ins o m n i a a n d in h e a l t h y n o r m a l s . T h e d a t a presented in this p a p e r are derived f r o m a larger s t u d y o f benzodiazepine h y p n o t i c self a d m i n i s t r a t i o n . The b e h a v i o r a l self a d m i n i s t r a t i o n d a t a are presented in a c o m p a n i o n p a p e r (Roehrs et al. 1991).
Materials and methods
Subjects. Twenty-one men and women, 25-50 years of age, were studied. All were in good health (except for an insomnia complaint) as determined by the screening procedures described below. All subjects signed a written, informed consent and received payment for their participation. Screening. Subjects underwent a medical history, physical examination, and laboratory blood and urine tests prior to a screening nocturnal polysomnogram. Subjects were excluded if they had acute or chronic medical conditions that required treatment or were currently taking CNS-acting drugs. Urine screens were used to document the absence of current drug use (licit or illicit) and breathalyzer tests for alcohol use. Subjects with a past history of psychiatric disorders, drug addiction, or alcoholism were excluded. Finally, subjects who had taken benzodiazepine hypnotics within the past year were also excluded. Each subject underwent a sleep disorders evaluation including a sleep history and a nocturnal polysomnogram. Three groups of subjects (n = 7 per group) were selected based on the sleep history and polysomnogram (see Table 1 for group descriptions). Subjects recruited as normals (NOR) reported no history of insomnia, estimated their nightly sleep time as greater than 7 h, and had a sleep efficiency on the nocturnal polysomnogram of 90% or greater. Subjects with insomnia (DIMS) had the complaint for at least 1 year, estimated their nightly sleep time as less than 6.5 h, and had a sleep efficiency of 85% or less on the screening polysomnogram. Subjects qualifying as persons with insomnia complaints, but no objective sleep disturbance (NOF), had a 1 year history of insomnia, estimated their nightly sleep time as less than 6.5 h, and had a sleep efficiency of 90% or more on the polysomnogram while estimating that their sleep time for the screening night was at least 1 h less. All subjects had no evidence of apneas or periodic leg movements on the nocturnal polysomnogram. Procedure. Each subject who passed the screening was exposed to three different treatments organized in a Latin Square design and administered in a double-blind fashion with 7 nights between treatments. Each treatment consisted of 1 adaptation night and 1 baseline night with no pill administration then 6 nights of 0.5 mg
triazolam (abrupt condition), or 3 nights of 0.5 mg, 2 nights of 0.25 mg, and 1 night of 0.125 mg triazolam (taper condition), or 6 nights of placebo. Following each treatment night was 1 night with no pill administration to test for the presence of rebound insomnia. All capsules were matched for appearance and given to the subject 30 min prior to lights out. Subjects arrived at the sleep laboratory about 1.5 h before their usual bedtime which was held constant throughout the study. They were prepared for standard polysomnographic assessment (Rechtschaffen and Kales 1968). A standard all-night polysomnogram was obtained on the 1 adaptation, and the 1 baseline night, the first, third, and last night of each treatment and on the discontinuation night. The polysomnogram included the central and occipital electroencephalograms (EEG), electrooculograms (EOG), and submental electromyogram (EMG) collected from standard placements and monitored continuously for 8 h. On the diagnostic screening night the recording also included nasal and oral respiration measured with thermistors and bilateral leg (anterior tibialis) electromyograms, Subjects were awakened after 8 h in bed (960 pages from lights out). After arising, each subject then completed a sleep questionnaire regarding the quantity and quality of the previous night's sleep. Recordings were made using Grass model 78-D or Nihon Kohden (models 4312 and 4212) polygraphs. The Grass polygraph was calibrated with a pen deflection of 50 gV = 7.5 mm for the EEG and EOG, and 50 pV= 10.0 mm for the EMG. The 0.5 amp low frequency filter was set at 0,3 with a sensitivity at 5 for the EEG and EOG and 10 with a sensitivity of 1 or the EMG. The 0.5 amp frequency high filter was set at 90. The Nihon Kohden machines were calibrated at 50 ~tV= 10.0 mm for the EEG and EOG and 50 gV = 16.5 mm for the EMG. The 0.5 amp low frequency filters were set at 0.3 with the sensitivity at 5 for the EEG and EOG and 0.003 with a sensitivity of 1 for the EMG. The 0.5 amp frequency high filters were set at 70. All electrode impedances were less than 10000 ohms and paper speed was 10 mm/s. The following study restrictions were adhered to by all subjects: 1) no alcoholic or caffeinated beverages after 4:00 p.m. on study nights, 2) no napping during the study, 3) no changes in bed or wake times during the study, 4) 8 h in bed each night during the study and 5) no medications without the approval of the investigator. Each polysomnographic recording was scored manually in 30-s epochs according to the standards of Rechtschaffen and Kales (1968), maintaining an inter-rater reliability of 90% or better. Records were coded so that scorers were unaware of the treatment. Mixed design MANOVAs (SAS Institute) were conducted on the various polysomnographic parameters with group as the between subject factor and treatment condition and night as the within subject factors. Conservative probability levels corrected by Greenhouse-Geisser procedure were used for all within subject comparisons. Two sets of MANOVA analyses were conducted. The first was conducted on the results of the 3 pill administration nights to establish the pharmacological activity of the drug in our methodology. The second set of analyses comparing baseline nights to discontinuation nights assessed the presence of rebound insomnia.
Results Table 1. Characteristics of study groups Group
NOR
NOF
DIMS
Age Sex females males Insomnia Hx SubjTST CPSG SE
25-50 4 3 Absent > 7h >90%
25-50 3 4 _> 1 yr < 6.5h >90%
25-50 4 3 _> 1 yr < 6.5h
