CLINICAL TRIALS AND THERAPEUTICS
Rebound insomnia after only brief and intermittent use of rapidly eliminated benzodiazepines In three parallel groups, brief and intermittent administration and withdrawal of triazolam, 0.5 mg, temazepam, 30 mg, and placebo were assessed in a 12-night sleep laboratory study of 18 subjects with insomnia With this intermittent schedule both drugs improved sleep, with about a one-third reduction in total wake time; this reduction was significant for temazepam but not for triazolam. Even though the periods of drug administration were quite brief, withdrawal of triazolam consistently produced rebound insomnia, with increases in total wake time above baseline of 61% and 51%, respectively, for the first night of each withdrawal period. With temazepam this effect was more variable, with total wake time increased only with the second withdrawal period (39%). Thus these findings indicate that even under conditions of brief, intermittent use and withdrawal, triazolam and, to a lesser degree, temazepam produce rebound insomnia after abrupt withdrawal, thereby predisposing to drug-taking behavior and increasing the potential for drug dependence. (CLIN PHARAIACOL THER 1991;49:468-76.)
Anthony Kales, MD, Rocco L. Manfredi, MD, Alexandros N. Vgontzas, MD, Edward 0. Bixler, PhD, Antonio Vela-Bueno, MD, and Eric C. Fee, MD Hershey, Pa.
Withdrawal difficulties can occur after abrupt termination of benzodiazepines used either in high doses or for long periods." More recently, sleep laboratory studies have clearly established rebound insomnia after the abrupt withdrawal of rapidly eliminated benzodiazepine hypnotics or anxiolytics administered for only relatively short periods in therapeutic doses.5-16 Furthermore, patients receiving these rapidly eliminated benzodiazepines develop tolerance to their efficacy rather rapidly (i.e., within only or 2 weeks of 1
From the Sleep Research and Treatment Center and the Department of Psychiatry, Pennsylvania State University College of Medicine. Supported in part by a grant from Sandoz Pharmaceuticals, East Hanover, N.J. Received for publication Sept. 25, 1990; accepted Dec. 26, 1990. Reprint requests: Anthony Kales, MD, Department of Psychiatry, Pennsylvania State University College of Medicine, 500 University Dr., Hershey, PA 17033. 13/1/27610
468
consecutive nightly use).813.14,17-19 Because of the rapid development of tolerance during administration and significant sleep disturbances after withdrawal, the consecutive nightly use of rapidly eliminated benzodiazepines for even relatively short-term periods places patients at high risk for drug dependence. Given these limitations, it would be important to determine if these difficulties with rapidly eliminated benzodiazepines could be avoided under conditions of both brief and intermittent nightly use. Such information is particularly important because these hypnotics are frequently recommended for short-term or asneeded use, such as brief hospital admission, travel including jet lag, shift work, and acute emotional stress 2024 Accordingly, there were three primary objectives in this study of temazepam, 30 mg, triazolam, 0.5 mg, and placebo. These objectives were to evaluate the effects of short-term use of each drug and placebo on sleep induction and maintenance and sleep architecture, the effects
VOLUME 49 NUMBER 4
Rebound insomnia after benzodiazepines 469
Table I. Study design Study night
Baseline Group
2
3
Experimental
4
5
6
7
8
10
9
11
12
Temazepam
n= n
=
3 3
Triazolam
n= n=
3 3
Placebo n
=
n=
3 3
P, Placebo; D, drug (temazepam, 30 mg, or triazolam, 0.5 mg).
of initially withdrawing each drug and placebo, and their effects when readministered and withdrawn.
METHODS Subjects. Eighteen subjects with insomnia (12 women and 6 men) aged 20 to 61 years (mean age, 36.6 3.5 years) were recruited to participate in this study, which also assessed the effects of administration of triazolam, temazepam, and placebo on memory. The results of the memory portion of this study are reported elsewhere.25 Subjects were in good physical health, not using any medication for the past month, and without evidence of any mental disorder that would require psychopharmacologic treatment or preclude the subject from adhering to the study guidelines and requirements. Each subject had a history of difficulty falling asleep, staying asleep, or both. To qualify for the study, subjects had to report a history of taking an estimated 45 minutes to fall asleep or obtaining fewer than 61/2 hours of total sleep time. Throughout the study, subjects were instructed not to nap, not to alter their level of daily activity significantly, and not to use any medication. Each subject was fully informed of the nature of the experiment, gave his or her written consent, and was paid for participation in the study. Study design. Six subjects were randomly assigned to each of the three parallel groups of this study: one group (mean age, 37.3 -± 6.9 years) received triazolam, 0.5 mg, on certain nights and matching placebo on other nights; one group (37.0 ± 6.7 years) received temazepam, 30 mg, on certain nights and placebo on the other nights; and the third group (35.2 ± 5.5 years) received only placebo. Matching capsules -2L-
containing triazolam, temazepam, or placebo were administered each night at lights out. Subjects, as well as research staff, were blind to drug group assignment and the nightly schedule for placebo and drug administration. In addition to an initial placebo-baseline period for each of the two drug groups, further control for the study was provided by the parallel placebo group. Thus double-blind placebo control was established both within and between subjects. To establish baseline measures, all subjects received placebo on nights through 4 (Table I). Night was considered an adaptation night and was excluded from the baseline values. After the baseline period, within each of the drug groups (temazepam and triazolam) there were two different schedules for administration of placebo and active drug. On 5 of the subsequent 8 nights, half of the subjects with insomnia assigned to each of the two drug groups received active drug on nights 5 through 8 and 11, whereas the other half received drug on nights 5 through 7 and 10 and II. On each of the 12 consecutive laboratory nights all subjects were monitored continuously for 8 hours by electroencephalography, electromyography, and electro-oculography.26 Parameters evaluated in the study included sleep-efficiency parameters such as sleep induction (sleep latency) and sleep maintenance (wake time after sleep onset, total wake time, and number of awakenings). An awakening was defined as an EEG arousal lasting 2 minutes or longer. Also assessed were sleep-stage parameters such as rapid eye movement (REM) sleep, 1, 2, 3, and 4, number of REM periods, and interval from sleep onset to the first REM period (REM latency). Daily all subjects were asked to arise and dress, af1
1
CLIN PHARMACOL THER APRIL 1991
470 Kales et al. Table II. Effectiveness of triazolam, 0.5 mg, for sleep induction and maintenance
Sleep latency (min) Wake time after sleep onset (min) Total wake time (min) Total No. of awakenings Sleep time (%) Wake time by thirds of night (%) First Second Third
Baseline (nights 2-4)
Drug (nights 5-7)
Withdrawal (nights 8-9 or 9-10)
Drug (nights 10 or 11)
Withdrawal (night 12)
37.9 30.7 68.6 6.4 85.7
33.1 15.1
42.2 48.2 90.4 6.8 81.2
23.8
51.9 52.0
92.4
7.8 78.4
6.0 5.6 8.9
2.4 2.6 5.1
13.7 12.1
1.9 1.9
5.3 20.3*
7.6
4.5
12.4
48.2 3.1*
90.0
12.7 36.5 2.0*
103.9
All numbers represent mean values for the six subjects. *p < 0.05 compared with baseline.
ter which they completed a set of subjective sleep and mood questionnaires. This included completing postsleep questionnaires to estimate the previous night's sleep latency, total sleep time, quality of sleep, and degree of sleepiness and emotional state in the morning (e.g., tension/anxiety). In the evening before lights out and before drug or placebo was administered, the subjects completed a presleep questionnaire in which they were instructed to record any daytime experiences that might affect their sleep, report any side effects that occurred, and assess how they felt during the day and at bedtime. Data analysis. All sleep records were scored according to standardized criteria,26 independent of any knowledge of the experimental condition. After all the sleep recordings were scored, the drug code was broken and the data were analyzed by drug group. The Dunn multiple-comparison, two-tailed t test was used to contrast the drug and withdrawal conditions with the baseline condition for each of the sleep efficiency and sleep stage variables for each of the two drug groups.27 Comparisons with baseline (nights 2 through 4) were made across all conditions between sets of nights: all 5 drug nights combined, initial short-term drug (nights 5 through 7), second drug period (night 10 or 11), first withdrawal period (nights 8 and 9 or 9 and 10), and second withdrawal period (night 12). To compare the effectiveness of temazepam and triazolam, an additional analysis of the data was conducted. For each subject, difference scores for each parameter were obtained by subtracting the values for each drug or withdrawal night from the mean value of the subject's baseline. Then the Dunn multiplecomparison t test was used to compare by condition each of the two drug groups with each other and with the placebo group. Specifically, mean difference scores obtained for sleep latency, wake time after
sleep onset, total wake time, and number of awakenings were compared.
RESULTS In comparing the three groups at baseline, the total wake time for the temazepam group was found to be significantly greater than that of the triazolam group. Efficacy of triazolam. The mean overall results for sleep induction and sleep maintenance with triazolam, 0.5 mg, are reported in Table II. On baseline nights (2 through 4) total wake time was 68.6 minutes, consisting of a sleep latency period of 37.9 minutes and wake time after sleep onset of 30.7 minutes. With administration of triazolam (all 5 drug nights combined), total wake time decreased by 32.9% to 46.0 minutes (difference not significant), with a sleep latency period of 31.0 minutes (difference not significant) and wake time after sleep onset of 15.0 minutes (difference not significant). With initial short-term administration of triazolam (nights 5 through 7), total wake time decreased by 29.7% to 48.2 minutes, compared with 68.6 minutes at baseline (difference not significant). This short-term drug administration value consisted of a sleep latency period of 33.1 minutes (difference not significant) and wake time after sleep onset of 15.1 minutes (difference not significant). Night-by-night evaluation of short-term administration showed a consistent but nonsignificant decrease in total wake time across all 3 drug nights (45.4, 49.7, and 49.6 minutes for nights 5, 6, and 7, respectively). Withdrawal effects of triazolam. After withdrawal of triazolam (all 3 nights of placebo withdrawal combined), total wake time was 38.3% above baseline (94.9 minutes compared with 68.6 minutes at baseline; p < 0.05). This worsening of sleep was the re-
VOLUME 49 NUMBER 4
Rebound insomnia after benzodiazepines
471
Table III. Effectiveness of temazepam, 30 mg, for sleep induction and maintenance
Sleep latency (min) Wake time after sleep onset (min) Total wake time (min) Total No. of awakenings Sleep time (%) Wake time by thirds of night (%) First Second Third
Baseline (nights 2-4)
Drug (nights 5-7)
Withdrawal (nights 8-9 or 9-10)
57.8 34.5 92.3 4.7 80.8
33.7* 18.0
45.7 42.4
51.71'
88.1
3.8
6.9 81.6
89.2t
6.6 6.7 10.9
1.3
4.7 6.0
6.1 13.2
9.4
Drug (night 10 or 11)
45.5 24.2 69.7 4.0 85.5 2.5 7.4 6.5
Withdrawal (night 12)
79.0 49.6 128.6* 3.2 73.2* 17.2 12.3
5.0
All numbers represent mean values for the six subjects. *p < 0.05 compared with baseline. tp < 0.01 compared with baseline.
suit of an increase in both sleep latency (45.5 minutes versus 37.9 minutes; difference not significant) and wake time after sleep onset (49.4 minutes versus 30.7 minutes; difference not significant). The first withdrawal period (nights 8 and 9 or 9 and 10) resulted in an increase in total wake time to 90.4 minutes from 68.6 minutes at baseline (difference not significant). Night-by-night analysis showed a significant increase (60.6%) on the first withdrawal night to 110.2 minutes of total wake time versus 68.6 minutes (p < 0.05) but not on the second withdrawal night (total wake time of 70.6 minutes). The final withdrawal period (night 12) showed an increase in total wake time 51.4% above baseline (103.9 minutes versus 68.6 minutes; difference not significant). Efficacy of temazepam. The mean overall results for sleep induction and sleep maintenance with temazepam, 30 mg, are reported in Table III. On baseline nights (2 through 4) total wake time was 92.3 minutes, consisting of a sleep latency period of 57.8 minutes and wake time after sleep onset of 34.5 minutes. With administration of temazepam (all 5 drug nights combined), total wake time decreased by 35.3% to 59.7 minutes (p < 0.01), with a sleep latency of 39.4 minutes (p < 0.01) and wake time after sleep onset of 20.3 minutes (difference not significant). With initial short-term administration of temazepam (nights 5 through 7), total wake time decreased by 44.0% to 51.7 minutes compared with 92.3 minutes at baseline (p < 0.01), with a sleep latency period of 33.7 minutes (p < 0.05) and wake time after sleep onset of 18.0 minutes (difference not significant). Nightby-night evaluation of the short-term drug period (nights 5 through 7) showed that temazepam, 30 mg, was consistently effective across all 3 drug nights, with a total wake time of 54.6 (p < 0.05), 52.2 (p
Rebound insomnia after only brief and intermittent use of rapidly eliminated benzodiazepines In three parallel groups, brief and intermittent administration and withdrawal of triazolam, 0.5 mg, temazepam, 30 mg, and placebo were assessed in a 12-night sleep laboratory study of 18 subjects with insomnia With this intermittent schedule both drugs improved sleep, with about a one-third reduction in total wake time; this reduction was significant for temazepam but not for triazolam. Even though the periods of drug administration were quite brief, withdrawal of triazolam consistently produced rebound insomnia, with increases in total wake time above baseline of 61% and 51%, respectively, for the first night of each withdrawal period. With temazepam this effect was more variable, with total wake time increased only with the second withdrawal period (39%). Thus these findings indicate that even under conditions of brief, intermittent use and withdrawal, triazolam and, to a lesser degree, temazepam produce rebound insomnia after abrupt withdrawal, thereby predisposing to drug-taking behavior and increasing the potential for drug dependence. (CLIN PHARAIACOL THER 1991;49:468-76.)
Anthony Kales, MD, Rocco L. Manfredi, MD, Alexandros N. Vgontzas, MD, Edward 0. Bixler, PhD, Antonio Vela-Bueno, MD, and Eric C. Fee, MD Hershey, Pa.
Withdrawal difficulties can occur after abrupt termination of benzodiazepines used either in high doses or for long periods." More recently, sleep laboratory studies have clearly established rebound insomnia after the abrupt withdrawal of rapidly eliminated benzodiazepine hypnotics or anxiolytics administered for only relatively short periods in therapeutic doses.5-16 Furthermore, patients receiving these rapidly eliminated benzodiazepines develop tolerance to their efficacy rather rapidly (i.e., within only or 2 weeks of 1
From the Sleep Research and Treatment Center and the Department of Psychiatry, Pennsylvania State University College of Medicine. Supported in part by a grant from Sandoz Pharmaceuticals, East Hanover, N.J. Received for publication Sept. 25, 1990; accepted Dec. 26, 1990. Reprint requests: Anthony Kales, MD, Department of Psychiatry, Pennsylvania State University College of Medicine, 500 University Dr., Hershey, PA 17033. 13/1/27610
468
consecutive nightly use).813.14,17-19 Because of the rapid development of tolerance during administration and significant sleep disturbances after withdrawal, the consecutive nightly use of rapidly eliminated benzodiazepines for even relatively short-term periods places patients at high risk for drug dependence. Given these limitations, it would be important to determine if these difficulties with rapidly eliminated benzodiazepines could be avoided under conditions of both brief and intermittent nightly use. Such information is particularly important because these hypnotics are frequently recommended for short-term or asneeded use, such as brief hospital admission, travel including jet lag, shift work, and acute emotional stress 2024 Accordingly, there were three primary objectives in this study of temazepam, 30 mg, triazolam, 0.5 mg, and placebo. These objectives were to evaluate the effects of short-term use of each drug and placebo on sleep induction and maintenance and sleep architecture, the effects
VOLUME 49 NUMBER 4
Rebound insomnia after benzodiazepines 469
Table I. Study design Study night
Baseline Group
2
3
Experimental
4
5
6
7
8
10
9
11
12
Temazepam
n= n
=
3 3
Triazolam
n= n=
3 3
Placebo n
=
n=
3 3
P, Placebo; D, drug (temazepam, 30 mg, or triazolam, 0.5 mg).
of initially withdrawing each drug and placebo, and their effects when readministered and withdrawn.
METHODS Subjects. Eighteen subjects with insomnia (12 women and 6 men) aged 20 to 61 years (mean age, 36.6 3.5 years) were recruited to participate in this study, which also assessed the effects of administration of triazolam, temazepam, and placebo on memory. The results of the memory portion of this study are reported elsewhere.25 Subjects were in good physical health, not using any medication for the past month, and without evidence of any mental disorder that would require psychopharmacologic treatment or preclude the subject from adhering to the study guidelines and requirements. Each subject had a history of difficulty falling asleep, staying asleep, or both. To qualify for the study, subjects had to report a history of taking an estimated 45 minutes to fall asleep or obtaining fewer than 61/2 hours of total sleep time. Throughout the study, subjects were instructed not to nap, not to alter their level of daily activity significantly, and not to use any medication. Each subject was fully informed of the nature of the experiment, gave his or her written consent, and was paid for participation in the study. Study design. Six subjects were randomly assigned to each of the three parallel groups of this study: one group (mean age, 37.3 -± 6.9 years) received triazolam, 0.5 mg, on certain nights and matching placebo on other nights; one group (37.0 ± 6.7 years) received temazepam, 30 mg, on certain nights and placebo on the other nights; and the third group (35.2 ± 5.5 years) received only placebo. Matching capsules -2L-
containing triazolam, temazepam, or placebo were administered each night at lights out. Subjects, as well as research staff, were blind to drug group assignment and the nightly schedule for placebo and drug administration. In addition to an initial placebo-baseline period for each of the two drug groups, further control for the study was provided by the parallel placebo group. Thus double-blind placebo control was established both within and between subjects. To establish baseline measures, all subjects received placebo on nights through 4 (Table I). Night was considered an adaptation night and was excluded from the baseline values. After the baseline period, within each of the drug groups (temazepam and triazolam) there were two different schedules for administration of placebo and active drug. On 5 of the subsequent 8 nights, half of the subjects with insomnia assigned to each of the two drug groups received active drug on nights 5 through 8 and 11, whereas the other half received drug on nights 5 through 7 and 10 and II. On each of the 12 consecutive laboratory nights all subjects were monitored continuously for 8 hours by electroencephalography, electromyography, and electro-oculography.26 Parameters evaluated in the study included sleep-efficiency parameters such as sleep induction (sleep latency) and sleep maintenance (wake time after sleep onset, total wake time, and number of awakenings). An awakening was defined as an EEG arousal lasting 2 minutes or longer. Also assessed were sleep-stage parameters such as rapid eye movement (REM) sleep, 1, 2, 3, and 4, number of REM periods, and interval from sleep onset to the first REM period (REM latency). Daily all subjects were asked to arise and dress, af1
1
CLIN PHARMACOL THER APRIL 1991
470 Kales et al. Table II. Effectiveness of triazolam, 0.5 mg, for sleep induction and maintenance
Sleep latency (min) Wake time after sleep onset (min) Total wake time (min) Total No. of awakenings Sleep time (%) Wake time by thirds of night (%) First Second Third
Baseline (nights 2-4)
Drug (nights 5-7)
Withdrawal (nights 8-9 or 9-10)
Drug (nights 10 or 11)
Withdrawal (night 12)
37.9 30.7 68.6 6.4 85.7
33.1 15.1
42.2 48.2 90.4 6.8 81.2
23.8
51.9 52.0
92.4
7.8 78.4
6.0 5.6 8.9
2.4 2.6 5.1
13.7 12.1
1.9 1.9
5.3 20.3*
7.6
4.5
12.4
48.2 3.1*
90.0
12.7 36.5 2.0*
103.9
All numbers represent mean values for the six subjects. *p < 0.05 compared with baseline.
ter which they completed a set of subjective sleep and mood questionnaires. This included completing postsleep questionnaires to estimate the previous night's sleep latency, total sleep time, quality of sleep, and degree of sleepiness and emotional state in the morning (e.g., tension/anxiety). In the evening before lights out and before drug or placebo was administered, the subjects completed a presleep questionnaire in which they were instructed to record any daytime experiences that might affect their sleep, report any side effects that occurred, and assess how they felt during the day and at bedtime. Data analysis. All sleep records were scored according to standardized criteria,26 independent of any knowledge of the experimental condition. After all the sleep recordings were scored, the drug code was broken and the data were analyzed by drug group. The Dunn multiple-comparison, two-tailed t test was used to contrast the drug and withdrawal conditions with the baseline condition for each of the sleep efficiency and sleep stage variables for each of the two drug groups.27 Comparisons with baseline (nights 2 through 4) were made across all conditions between sets of nights: all 5 drug nights combined, initial short-term drug (nights 5 through 7), second drug period (night 10 or 11), first withdrawal period (nights 8 and 9 or 9 and 10), and second withdrawal period (night 12). To compare the effectiveness of temazepam and triazolam, an additional analysis of the data was conducted. For each subject, difference scores for each parameter were obtained by subtracting the values for each drug or withdrawal night from the mean value of the subject's baseline. Then the Dunn multiplecomparison t test was used to compare by condition each of the two drug groups with each other and with the placebo group. Specifically, mean difference scores obtained for sleep latency, wake time after
sleep onset, total wake time, and number of awakenings were compared.
RESULTS In comparing the three groups at baseline, the total wake time for the temazepam group was found to be significantly greater than that of the triazolam group. Efficacy of triazolam. The mean overall results for sleep induction and sleep maintenance with triazolam, 0.5 mg, are reported in Table II. On baseline nights (2 through 4) total wake time was 68.6 minutes, consisting of a sleep latency period of 37.9 minutes and wake time after sleep onset of 30.7 minutes. With administration of triazolam (all 5 drug nights combined), total wake time decreased by 32.9% to 46.0 minutes (difference not significant), with a sleep latency period of 31.0 minutes (difference not significant) and wake time after sleep onset of 15.0 minutes (difference not significant). With initial short-term administration of triazolam (nights 5 through 7), total wake time decreased by 29.7% to 48.2 minutes, compared with 68.6 minutes at baseline (difference not significant). This short-term drug administration value consisted of a sleep latency period of 33.1 minutes (difference not significant) and wake time after sleep onset of 15.1 minutes (difference not significant). Night-by-night evaluation of short-term administration showed a consistent but nonsignificant decrease in total wake time across all 3 drug nights (45.4, 49.7, and 49.6 minutes for nights 5, 6, and 7, respectively). Withdrawal effects of triazolam. After withdrawal of triazolam (all 3 nights of placebo withdrawal combined), total wake time was 38.3% above baseline (94.9 minutes compared with 68.6 minutes at baseline; p < 0.05). This worsening of sleep was the re-
VOLUME 49 NUMBER 4
Rebound insomnia after benzodiazepines
471
Table III. Effectiveness of temazepam, 30 mg, for sleep induction and maintenance
Sleep latency (min) Wake time after sleep onset (min) Total wake time (min) Total No. of awakenings Sleep time (%) Wake time by thirds of night (%) First Second Third
Baseline (nights 2-4)
Drug (nights 5-7)
Withdrawal (nights 8-9 or 9-10)
57.8 34.5 92.3 4.7 80.8
33.7* 18.0
45.7 42.4
51.71'
88.1
3.8
6.9 81.6
89.2t
6.6 6.7 10.9
1.3
4.7 6.0
6.1 13.2
9.4
Drug (night 10 or 11)
45.5 24.2 69.7 4.0 85.5 2.5 7.4 6.5
Withdrawal (night 12)
79.0 49.6 128.6* 3.2 73.2* 17.2 12.3
5.0
All numbers represent mean values for the six subjects. *p < 0.05 compared with baseline. tp < 0.01 compared with baseline.
suit of an increase in both sleep latency (45.5 minutes versus 37.9 minutes; difference not significant) and wake time after sleep onset (49.4 minutes versus 30.7 minutes; difference not significant). The first withdrawal period (nights 8 and 9 or 9 and 10) resulted in an increase in total wake time to 90.4 minutes from 68.6 minutes at baseline (difference not significant). Night-by-night analysis showed a significant increase (60.6%) on the first withdrawal night to 110.2 minutes of total wake time versus 68.6 minutes (p < 0.05) but not on the second withdrawal night (total wake time of 70.6 minutes). The final withdrawal period (night 12) showed an increase in total wake time 51.4% above baseline (103.9 minutes versus 68.6 minutes; difference not significant). Efficacy of temazepam. The mean overall results for sleep induction and sleep maintenance with temazepam, 30 mg, are reported in Table III. On baseline nights (2 through 4) total wake time was 92.3 minutes, consisting of a sleep latency period of 57.8 minutes and wake time after sleep onset of 34.5 minutes. With administration of temazepam (all 5 drug nights combined), total wake time decreased by 35.3% to 59.7 minutes (p < 0.01), with a sleep latency of 39.4 minutes (p < 0.01) and wake time after sleep onset of 20.3 minutes (difference not significant). With initial short-term administration of temazepam (nights 5 through 7), total wake time decreased by 44.0% to 51.7 minutes compared with 92.3 minutes at baseline (p < 0.01), with a sleep latency period of 33.7 minutes (p < 0.05) and wake time after sleep onset of 18.0 minutes (difference not significant). Nightby-night evaluation of the short-term drug period (nights 5 through 7) showed that temazepam, 30 mg, was consistently effective across all 3 drug nights, with a total wake time of 54.6 (p < 0.05), 52.2 (p
