- Email: [email protected]
Can valerian improve the sleep of insomniacs after benzodiazepine withdrawal?
Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539 – 545
Article
Can valerian improve the sleep of insomniacs after benzodiazepine withdrawal? Dalva R. Poyaresa,b, Christian Guilleminaultb,*, Maurice M. Ohayonb, Sergio Tufika a
Sleep Laboratory of the Department of Psychobiology, Universidade Federal de Sa˜o Paulo, Sao Paulo, Brazil b Stanford University Sleep Disorders Center, Suite 3301, 401 Quarry Road, Stanford, CA 94305, USA
Abstract Purpose: The authors studied the sleep of patients with insomnia who complained of poor sleep despite chronic use of benzodiazepines (BZDs). The sample consisted of 19 patients (mean age 43.3 ± 10.6 years) with primary insomnia (DSM-IV), who had taken BZDs nightly, for 7.1 ± 5.4 years. The control group was composed of 18 healthy individuals (mean age 37 ± 8 years). Sleep electroencephalogram (EEG) of the patients was analyzed with period amplitude analysis (PAA) and associated algorithms, during chronic BZD use (Night 1), and after 15 days of a valerian placebo trial (initiated after washout of BZD, Night 2). Sleep of control subjects was monitored in parallel. Results: Valerian subjects reported significantly better subjective sleep quality than placebo ones, after BZD withdrawal, despite the presence of a few side effects. However, some of the differences found in sleep structure between Night 1 and Night 2 in both the valerian and placebo groups may be due to the sleep recovery process after BZD washout. Example of this are: the decrease in Sleep Stage 2 and in sigma count; the increase in slow-wave sleep (SWS), and delta count, which were found to be altered by BZD ingestion. There was a significant decrease in wake time after sleep onset (WASO) in valerian subjects when compared to placebo subjects; results were similar to normal controls. Nonetheless, valerian-treated patients also presented longer sleep latency and increased alpha count in SWS than control subjects. Conclusions: The decrease in WASO associated with the mild anxiolytic effect of valerian appeared to be the major contributor to subjective sleep quality improvement found after 2-week of treatment in insomniacs who had withdrawn from BDZs. Despite subjective improvement, sleep data showed that valerian did not produce faster sleep onset; the increase in alpha count compared with normal controls may point to residual hyperarousabilty, which is known to play a role in insomnia. Nonetheless, we lack data on the extent to which a sedative drug can improve alpha sleep EEG. Thus, the authors suggest that valerian had a positive effect on withdrawal from BDZ use. D 2001 Elsevier Science Inc. All rights reserved. Keywords: Benzodiazepines; Drug withdrawal; Insomnia; Period amplitude analysis; Sleep EEG; Valerian
1. Introduction The chronic use of benzodiazepines (BZDs) in the treatment of insomnia causes several problems such as dependence, tolerance, long-term memory impairment, and altered sleep structure (Griffiths and Weerts, 1997; Murphy and Tyrer, 1991; Gorenstein et al., 1994; Lader and Russel, 1993). It is also associated with increased mortality risk
Abbreviations: Benzodiazepine, BZD; Electrocardiogram, ECG; Electroencephalogram, EEG; Electromyogram, EMG; Electrooculogram, EOG; First-Derivative Analysis, FDA; Period amplitude analysis, PAA; Rapid eye movement, REM; Nonrapid eye movement, NREM; Wake time after sleep onset, WASO * Corresponding author. Tel.: +1-650-723-6601; fax: +1-650-7258910. E-mail address: [email protected] (C. Guilleminault).
(Kripke et al., 1998). In animal models, the chronic use of BZDs results in a decreased number of GABA-A receptors in the brain (Chiu and Rosenberg, 1978; Miller et al., 1988; Sher et al., 1983). However, the causal relation between this finding and tolerance of and dependence on these substances is not completely understood. Because of the problems associated with long-term use of BZDs, efforts have been made to manage successful withdrawal. However, most authors have used substances that act upon the GABAergic system with intermittent reports of success (Ries et al., 1989; Rickels et al., 1990; Schweizer et al., 1991; Pailhous et al., 1994; Lemoine and Ohayon, 1997; Shapiro, 1994; Genton et al., 1994; Smits et al., 1996). Valerian, as a naturally occurring root used in many countries by patients with sleep complaints, may be useful as an alternative treatment in withdrawal from BZDs use. In rats, valerian prevents the development of the withdrawal
0278-5846/01/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 2 7 8 - 5 8 4 6 ( 0 1 ) 0 0 3 0 5 - 0
540
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545
syndrome resulting from the removal of diazepam following prolonged periods of administration (Andreatini and Leite, 1994) and does not induce toxic effects (Tufik et al., 1994). The fact that valerian is able to displace in vitro [3H]muscimol, a GABA agonist (Cavadas et al., 1995), suggests that it may act on the GABA receptor. Further evidence suggests that valerian effects are mediated by GABA (Santos et al., 1994a,b). Since 1977 (Jansen, 1977), several randomized, doubleblind trials have considered the effect of valerian extract on sleep. The evaluations were initially based on symptom scores and questionnaires (Kamm-Kohl et al., 1984; Vorbach et al., 1996), but several investigations used actigraphy or polygraphic recordings (Leathwood et al., 1982a, 1985; Balderer and Borbe´ly, 1985; Gebner and Klasser, 1984; Schulz et al., 1994; Donath et al., 2000). The surveyed populations were often healthy volunteers (young or elderly), poor sleepers, or chronically ill seniors. Results were variable but usually improved sleep latency was reported. Balderer and Borbe´ly (1985) reported shorter sleep latency and wake time after sleep onset. Similarly, Schulz et al. (1994) reported a significant increase in slow-wave sleep (SWS) and K-complex density after 8 days, and a decrease in Stage 1 nonrapid eye movement (NREM) sleep in elderly poor sleepers treated with valerian root. These authors suggested a selective effect of valerian extract on NREM sleep. Overall, the results were usually positive when poor sleepers were the targeted group rather than healthy volunteers. Valerian is considered a mild sedative and hypnotic substance in the presence of a lack of daytime somnolence and toxicity (Tufik et al., 1992) or alterations of sleep architecture in rats and in humans (Leathwood and Chauffard, 1982). Because of the variability of results, however, the United States Pharmacopoeia (USP) (Blumenthal, 1998) did not recommend valerian as a short-term treatment for insomnia. Because of the reported effects of valerian, we hypothesized that it may be helpful after withdrawal from chronic BZDs use. We asked how much improvement in sleep could be obtained after BZD withdrawal through ingestion of valepotriates, the active ingredient of valerian root. We also sought to discover whether valerian treatment after BZD withdrawal could improve sleep in insomnia patients and have it approach the values found in a healthy control group.
2. Methods 2.1. Subjects For 3 months at the sleep clinic at the Federal University of Sao Paulo, Brazil, the authors screened insomnia patients who continued to sleep poorly after long-term use of BZDs. All insomnia and control subjects were submitted to clinical interviews and clinical evaluations including bio-
chemical and hormonal tests to exclude medical disorders. These evaluations consisted of a structured psychiatric interview (SCID) after which subjects were classified according to DSM-IV criteria (American Psychiatry Association, 1994) and those with major psychiatric disorders were excluded from the selection. Sleep disorders other than insomnia in the patient group were excluded in the first recording night at sleep laboratory. Nineteen insomniac subjects (15 women and 4 men, mean age 43.3 ± 10.6 years) met the criteria for primary insomnia (DSM-IV). They had taken BZDs daily for at least 2.5 years and also met the DSM-IV BDZ drug-dependency criteria, registering a positive score and 4/7 of the listed criteria for this disorder. They agreed to undergo the protocol and signed informed consent forms. Eighteen healthy individuals (nine men, nine women; mean age 37 ± 8 years) formed the control group. 2.2. Protocol design Subjects were monitored with polysomnography at baseline while taking BZD (Night 1). Drug use was gradually withdrawn over a 2-week period and at the end of the withdrawal period they were kept without medication for 48 h and their sleep recorded again. After completion of withdrawal, subjects were then randomly assigned to a placebo/valerian trial with 50% of the subjects receiving the active ingredient, a concoction of valepotriates called Valmane, and the other half receiving a placebo in a double-blind administration. A second nocturnal polysomnographic recording was performed after 15 days of treatment with both the active drug and placebo (Night 2). The valerian concoction contained 80% didrovaltrate, 15% valtrate, and 5% acelvaltrate from the Valerian wallichii root and was administered three times daily in 100-mg doses obtained from a commercially available product, Valmane (Whitehall Pharmaceutical). 2.3. Withdrawal procedure All 19 insomniacs were taking BZD medication daily. The most frequently used BZD was lorazepam (3 patients), followed in order of decreasing frequency of intake by bromazepam and flurazepam (10 patients), midazolam (4 patients), clonazepam (3 patients), flunitrazepam (2 patients), and finally alprazolam and estazolam (1 patient each). Five patients took two different BZDs each night. The withdrawal schedule was calculated on the basis of total BZD dosage intake. All subjects received only BZDs and all doses were within the therapeutic range recommended by regulatory agencies. Subjects who received two BZD drugs were never at the highest range of the recommended dosage, even taking their combined BZD dosage. Total dosage was then reduced in two stages to 50% by Day 7, gradually reaching complete discontinuance by Day 14. During withdrawal, all subjects were followed clinically
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545
every 2 days and could also contact the supervising physician at any other time if needed. This close follow-up was put in place to provide support during withdrawal and ensure compliance with the prescribed BZD withdrawal schedule. No other instructions, except sleep hygiene (such as avoidance of alcohol and caffeine during the entire course of the study), were given to patients. In order to encourage and facilitate this follow-up, patients were provided with transportation, and in some cases, there was support for other expenses. 2.4. Data collection 2.4.1. Sleep diaries Sleep diaries were given to patients over the course of the study to assess sleep subjective quality, quantity, occurrence of naps, daily pattern of medication intake, and any consumption of caffeine or alcohol. Patients updated their diaries every morning. They also included three consecutive questions to check for BZD relapse, and one to check correct times of intake of valerian and placebo. The sleep diaries contained a visual– analogical scale of sleep quality with values ranging from 0 (the worst sleep) to 10 (the best sleep quality) to be recorded daily by the patients. 2.4.2. Urine BZD screening Urine BZD screening was collected twice: before the beginning of the study, and in the morning before the last night in the sleep laboratory (i.e., 15 days after BZD withdrawal), Night 2. The threshold considered positive for BZD usage was 200 ng or more as specified in the commercial kit. 2.4.3. Sleep recording Sleep was investigated on three different nights during the course of the study. After one night of habituation in the sleep laboratory with full monitoring, subjects were recorded prior to the beginning of drug withdrawal (experimental Night 1), and 15 days into the double-blind valerian/ placebo study (experimental Night 2). Normal controls were submitted to the same recording conditions including habituation and one test recording night (experimental Night 1) in the same laboratory setting. Polygraphic recording monitored the same parameters on experimental Nights 1 and 2, in insomnia patients and on Night 1 in controls. The following channels were systematically recorded: electroencephalogram [EEG] (C3 – A2; C4 – A1), submental and legs electromyogram [EMG], right and left electrooculograms (EOG), electrocardiogram [ECG] (with a modified V2 lead), airflow, chest and abdominal respiratory effort, and pulse oxymetry. 2.5. Data analyses During the habituation night, absence of other sleep disorders (such as sleep apnea syndrome, periodic leg
541
movement syndrome, etc.) was confirmed for both insomnia group and normal controls. Sleep/wake state on Nights 1 and 2 in the insomnia group and Night 1 in controls was scored blindly in relation to the experimental groups. Sleep stages were tabulated for 30-s epochs using the Rechtschaffen and Kales (1968) international scoring manual. 2.5.1. Sleep parameters The following sleep parameters were analyzed: sleep latency, total sleep time, sleep efficiency, number of arousals (16 s or longer), wake time after sleep onset (WASO), rapid eye movement (REM) and NREM stages as percentage of sleep time and REM density, defined as a rate representing the number of eye movements per REM sleep stage. 2.5.2. Sleep EEG The sleep EEG from each recorded night was downloaded on an IBM PC compatible, with a sampling rate of 128 Hz per channel. Sleep analysis of EEG frequencies contained in each band width during total sleep time was assessed using the period amplitude analysis (PAA) software. This technique measures EEG waves that cross the zero-volts baseline. Nonetheless, the isolated use of PAA is limiting because it is not capable of detecting fast- and low-amplitude EEG frequencies, which do not produce zero-crossings (Feinberg et al., 1978). Consequently, the complementary technique, First-Derivative Analysis (FDA), which is capable of detecting faster EEG frequencies, was added to the analysis. FDA is described in detail by Armitage (1995) and Armitage et al. (1995). The authors used a commercially available software program, which includes PAA and FDA. These analyses were performed in 30-s epochs. We considered each NREM sleep stage (and REM sleep) in totality, i.e., independent of the cycle in which the epoch was recorded for the analysis. The data are presented as averaged count (i.e., the averaged number of bursts of the specific EEG frequency band per minute of given sleep stage). The following EEG frequency bands were considered for analysis, because they are the most affected by BZD usage: delta (0.5 – 3.5 Hz), alpha (8 –11.5 Hz), and sigma (12 –16 Hz). 2.6. Statistical analysis 2.6.1. Sleep parameters A MANOVA was performed, with Night (1 and 2) and Treatment (valerian or placebo) as the main factors; the level of significance set was at P < .05. 2.6.2. Sleep EEG analysis A MANOVA was also performed to compare the sleep EEG frequency bands during Nights 1 and 2, using the same above-mentioned design. Student’s t test was performed to compare the values obtained from the sleep quality analogic –visual scale.
542
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545
3. Results 3.1. General results The average duration of the continuous BDZ intake was 7.1 ± 5.4 years, minimum of 2.5, and maximum of 14 years. Sleep diaries showed no reports of regular napping during daytime or intake of caffeine and alcohol over the course of the study. The urine analysis results were negative for BZD usage after withdrawal (i.e., values were < 200 ng). 3.2. Subjective data Analysis of next-morning sleep evaluation showed that valerian subjects reported on average significantly better sleep than placebo ones during the second week of treatment. The sleep quality analogical –visual scale average was 7.4 ± 0.9 for valerian subjects, against 5.4 ± 0.8 for placebo subjects ( P=.000). However, in the first week of treatment, three patients taking valerian complained of gastrointestinal symptoms, such as diarrhea, stomachache, and bitter taste in their mouth. Since these symptoms were limited, there was no need to open the codes, and these patients were allowed to continue in the protocol. 3.3. Analysis of sleep data The subsequent analyses were performed in all 19 insomnia subjects who completed their withdrawal from BZDs. Table 1 presents the results obtained on Night 1 (the baseline recording) on the control subjects when compared to the insomnia individuals after 15 days of treatment. As seen in Table 1, total sleep time and sleep latency were still significantly higher in control subjects, but not WASO, which was similar in both groups. All sleep stage percentage values, as well as REM density values, were also similar to those from normal controls. The sleep EEG data comparison showed a significant increase in alpha count during Sleep Stages 3 and 4 NREM in valerian-treated subjects.
3.3.1. Sleep parameters Table 2 presents the data obtained on experimental Night 1 (i.e., during BZD chronic use) and Night 2 (i.e., after the BZD withdrawal and 15 days treatment with placebo/valerian). These results show that patients experienced improved nocturnal sleep on Night 2, with a significant increase in SWS and deep sleep and a reduction in Stage 2 NREM sleep compared to the baseline (Night 1) in both valerian and placebo groups. REM density was also higher on Night 2 for both groups. Nonetheless, valerian-treated subjects showed a significant decrease in WASO in comparison to placebo. Conversely, the placebo group presented significantly shorter sleep latency. The other sleep parameters, as well as the sleep EEG ones, were similar between the two treatment groups. 3.3.2. Sleep EEG analysis Comparison between Nights 1 and 2 showed that BZD withdrawal was associated with a significant increase in ‘‘delta count’’ (the averaged number of bursts of delta per minute during this sleep stage). There was a significant increase in delta count in Stages 2, 3, and 4 NREM sleep on Night 2 compared to Night 1 for both placebo and valerian subjects. A significant decrease in 12 – 16 Hz, waves (labeled ‘‘sigma’’ and covering spindles and ‘‘BDZ spindles’’) was also observed in both groups during Stages 2, 3 NREM and REM sleep. In short, when compared to Night 1, both groups showed clear changes on Night 2 in Stages 2, 3, and 4 NREM sleep for all frequencies (as reported in Table 3). There was a significant increase in alpha and delta band count in REM sleep (outlined in Table 3).
4. Discussion This clinical study looks at a group of subjects that is rarely studied polysomnographically: chronic users of BZDs with regular drug intake who continue to complain
Table 1 Sleep stages and sleep EEG parameters obtained after BDZ withdrawal and treatment with valerian compared to a normal control group Sleep parameters
Night 1, control (n = 18)
Night 2, placebo group (n = 9)
Night 2, valerian group (n = 10)
F and P values
Sleep latency (min) Total sleep time (min) WASO (%) Sleep efficiency (%) REM latency (min) Stage 2 (%) Stages 3 and 4 (%) REM sleep (%) Alpha count Stage 3 Alpha count Stage 4
8.9 ± 16.2 418.0 ± 36.5 12.3 ± 15.9 89.6 ± 6.8 70.8 ± 33.0 52.5 ± 7.0 19.3 ± 7.9 21.3 ± 4.4 1.5 ± 1.2 0.8 ± 1.0
5.6 ± 5.5* 361.2 ± 59.7* 19.3 ± 6.9 79.2 ± 6.8* 123.6 ± 136.5 54.7 ± 12.3 18.3 ± 10.4 20.7 ± 9.6 3.3 ± 1.6* 3.0 ± 1.7*
30.8 ± 31.8* 356.2 ± 67.7* 12.6 ± 6.5 81.4 ± 8.0* 125.4 ± 74.2 50.7 ± 9.4 23.9 ± 8.4 19.9 ± 5.3 4.5 ± 1.8* 3.8 ± 1.7*
4.83; P =.0142 6.04; P =.006 1.09; n.s. 7.90; P =.002 2.09; n.s. 0.47; n.s. 1.23; n.s. 0.16; n.s. 13.77; P < .0001 17.39; P < .0001
The values are presented as mean ± S.D. * Differs from control group, P < .05; Student’s – Newman – Keuls test.
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545
543
Table 2 Parameters of sleep structure and percentage of sleep stages on two PSG recording nights Placebo (n = 9)
Valerian (n = 10)
Sleep parameter
Night 1 (baseline)
Night 2 (after treatment)
Night 1 (baseline)
Night 2 (after treatment)
Sleep latency (min) Sleep efficiency Total sleep time (min) WASO (%) REM latency (min) REM density Stage 2 (%) Stage 3 (%) Stage 4 (%) REM sleep (%)
22.6 82.1 383.7 14.6 140.1 2.6 60.4 6.8 5.5 19.6
5.5 79.1 361.1 19.3 123.6 2.5 54.7 10.2 8.0 20.7
25.2 81.4 377.9 16.2 137.2 1.9 64.3 8.4 3.3 16.8
30.7 81.4 356.1 12.5 125.4 3.2 50.6 14.3 9.5 19.8
( ± 15.1) ( ± 12.9) ( ± 50.0) ( ± 10.4) ( ± 69.4) ( ± 1.8) ( ± 8.7) ( ± 6.0) ( ± 5.6) ( ± 5.7)
( ± 5.4) ( ± 6.8) ( ± 59.7) ( ± 6.9) ( ± 136.4) ( ± 1.3)* ( ± 12.2)* ( ± 5.5)* ( ± 7.2)* ( ± 9.5)
( ± 26.6) ( ± 7.8) ( ± 34.1) ( ± 7.8) ( ± 52.9) ( ± 1.4) ( ± 9.3) ( ± 5.2) ( ± 4.8) ( ± 4.6)
( ± 31.8)# ( ± 7.9) ( ± 67.6) ( ± 6.4)# ( ± 74.1) ( ± 1.9)* ( ± 9.4)* ( ± 7.0)* ( ± 7.2)* ( ± 5.2)
F(1,17)
F(1,17)
F(1,17)
A
B
C
n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.
n.s. n.s. n.s. n.s. n.s. 6.4; P = .02 12.5; P = .002 6.8; P = .02 11.7; P = .003 n.s.
10.8; P = .004 n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.
Values are presented as mean ± S.D. MANOVA: A = group effect; B = night effect; C = interaction effect; N.S. = nonsignificant. * Differs from Night 1; P < 0.05, MANOVA. # Differs from placebo.
of insomnia. Chronic use was defined here as nightly usage for over 1 year, but our population had a much longer regular intake with a mean of 7 years and a minimum of 2.5 years. There is a lack of recognized routines and procedures for withdrawal of BZD use in insomnia patients after chronic usage. In addition, as mentioned in the Introduction, several drugs have been tested to help withdrawal of BZDs, but in general, they may not present clear advantages compared to BZDs. We acknowledge that some chronic insomnia patients show a tendency toward chronic use of psychoactive substances (Ohayon et al., 1998), despite the progress achieved with nonpharmacological treatments of insomnia. Herbs and natural substances have been also consumed and well accepted by these types of patients. We chose valerian (i.e., valepotriates) from these natural substances, because of its convenient pharmacological properties as a mild anxiolytic and hypnotic with minor side effects and apparent absence of toxicity, as mentioned in the Introduction.
4.1. Is valerian effective in BZD withdrawal? Most of the differences found in sleep structure between Nights 1 and 2 in both groups, valerian and placebo, may be due to the sleep recovery process after BZD washout. Examples of these are: the decrease in Sleep Stage 2 and in sigma count; the increase in SWS, delta count, and REM density which were found to be altered by BZD ingestion in some studies (Gaillard and Blois, 1989; Borbe´ly and Achermann, 1991; Mitler et al., 1979), reporting on longterm (54 nights) use of temazepam in insomniacs. Although REM latency remained above 120 min in both groups of treatment, this was not significantly different from control value. Therefore, it is reasonable to suppose that some of these improvements in sleep could result solely from BZD removal. The significant decrease in WASO in valerian subjects is believed to account for valerian subjects reporting more satisfaction with their sleep than placebo ones. Curiously, valerian subjects presented higher sleep latency,
Table 3 Parameters of sleep EEG analysis F Placebo (n = 9)
EEG frequency band count Stage 2 Stage 3 Stage 4 REM sleep
Delta Sigma Delta Sigma Delta Alpha Delta Sigma
Valerian (n = 10)
Night 1
Night 2
Night 1
Night 2
Group effect, F(1,17)
Night effect, F(1,17)
8.3 ± 2.3 7.2 ± 2.1 24.4 ± 4.2 5.9 ± 4.3 22.4 ± 22.0 2.2 ± 1.2 1.5 ± 0.8 0.5 ± 0.3
12.0 ± 4.4* 4.0 ± 1.7* 32.4 ± 2.3* 4.2 ± 2.4* 42.8 ± 19.4* 2.2 ± 1.8 6.1 ± 9.8* 0.2 ± 0.1*
8.7 ± 4.5 8.5 ± 2.4 20.9 ± 4.4 9.5 ± 4.7# 18.3 ± 20.0 1.6 ± 0.8 1.2 ± 0.9 0.4 ± 0.3
14.6 ± 7.0* 4.1 ± 1.8* 35.9 ± 3.0* 5.4 ±2.7* 46.0 ± 17.9* 2.8 ± 1.6* 3.7 ± 3.4* 0.3 ±0.1*
n.s. n.s. n.s. 5.07; P = .04 n.s. n.s. n.s. n.s.
11.4; P = .004 44.5; P = .000 14.0; P = .002 4.6; P = .05 16.4; P = .001 4.6; P = .04 4.4; P = .05 4.8; P = .04
Values are expressed as mean ± S.D. MANOVA: F = group and night effects; n.s. = nonsignificant. There was no interaction between the two treatment groups. * Differs from respective Night 1; P < .05. # Differs from Night 2; P < .05.
544
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545
which may have resulted either from individual differences in insomnia markers severity, or from different responses to BZD withdrawal (i.e., some valerian treated subjects may have presented a longer sleep latency than their placebo counterparts before the treatments started, soon after BZD washout). Therefore, these results may not suggest that valerian failed to promote faster sleep onset. When compared to normal controls, valerian subjects also presented increased sleep latency, which accounted for the decrease in TST, but when compared to placebo, valerian was able to improve WASO to the level found in healthy control subjects. 4.2. What are the meanings of EEG findings in valerian group? The sleep EEG findings, such as the increase in alpha count in SWS in valerian group when compared to normal controls, as well as the increase in delta count for both groups on Night 2 when compared to Night 1 (during BDZ use), are more difficult to interpret. Increased alpha power during SWS is a finding already related to nonrestorative sleep, found, for example, in fibromyalgic and anxious patients (Perlis et al., 1997; Drewes et al., 1995). However, in this case, this higher alpha count may be due to residual hyperarousability, which is characteristic of insomnia patients (Bonnet and Arand, 1997), compared to normal controls. This raises a question: To what extent can sedative – hypnotics improve high alpha EEG in insomnia patients? The increase in delta frequency associated with augmented Stages 3 and 4 NREM sleep percentage may be considered as evidence of better consolidation of sleep, in these patients, after drug removal (Brown et al., 1978; Borbely and Archermann, 1992). We believe it may indicate a delta and SWS recovery process following BZD withdrawal, despite the fact that SWS and delta activity were not expected to present values close to the control group’s levels, since a decrease in SWS and delta activity has been reported in chronic insomnia patients (Merica et al., 1998). However, there is very little knowledge on sleep recovery after very chronic BZD usage and valerian treatment. Finally, previous studies with valerian were usually positive when poor sleepers were targeted and compared to healthy volunteers. This may be the case for this study, since we were able to demonstrate a decrease in WASO in insomnia BZD chronic users. As mentioned above, in the Introduction, valerian is more likely to shorten sleep latency than to improve sleep maintenance, but a population of BZD chronic users has never been the target of a valerian trial. Polysomnographic findings may not completely explain subjective reports from insomnia patients. Subjective changes in sleep perception do not always correlate with objective sleep data. Nowell et al. (1999) found that paroxetine showed a significant subjective improvement of sleep in insomniacs, but there was no improvement in PSG
data. In this study, however, the data suggest that the positive subjective effect of valepotriates compared to placebos after BZD washout could be due to the decrease in WASO, but particularly because of the anxiolytic effect that may have played an important role in the subjective perception of improved sleep quality.
5. Conclusions In the light of the present results, and those of other studies (Vorbach et al., 1996; Leathwood et al., 1985; Schulz et al., 1994), all of which showed positive findings with valerian intake, there is a need for more research into valepotriates that should focus on a larger but homogenous group of patients with insomnia complaints. Moreover, this is supported by data in rats showing that valerian was able to taper BDZ withdrawal signs (Andreatini and Leite, 1994). Due to (a) objective sleep data in chronic BDZ users remaining different to normal controls after withdrawal; (b) sleep data being slightly improved after a 2-week treatment with valerian; (c) the persistence of increased alpha during REM sleep; and (d) the lack of data on predictors of BDZ relapse in insomniacs after chronic use, we recommend that concurrent behavioral techniques to improve the possible hyperarousal condition in these insomniacs should be considered (Morin et al., 1994), even though patients report subjective improvement in sleep.
Acknowledgments The authors would like to thank Dr. Deborah Suchecki for her comments on the manuscript. This work was supported by Associac¸a˜o Fundo de Incentivo a` Psicofarmacologia (AFIP). Dalva Poyares was the recipient of a fellowship from Coordenac¸a˜o de Aperfeic˛oamento de Pessoal de Ensino Superior (CAPES).
References American Psychiatry Association, 1994. Diagnostic and Statistical Manual of Mental Disorders (DSM-IV), fourth ed. American Psychiatry Press, Washington. Andreatini, R., Leite, J.R., 1994. Effect of valepotriates on the behaviour of rats on the elevated plus maze during benzodiazepine withdrawal. Eur. J. Pharmacol. 260, 233 – 235. Armitage, R., 1995a. The distribution of EEG frequencies in REM and NREM sleep stages in healthy young adults. Sleep 18, 334 – 341. Armitage, R., Hoffmann, R., Fitch, T., Morel, C., Bonato, R., 1995b. A comparison of period amplitude and power spectral analysis of sleep EEG in normal adults and depressed out patients. Psychiatry Res. 56, 245 – 256. Balderer, G., Borbe´ly, A.A., 1985. Effect of valerian on human sleep. Psychopharmacology 87, 406 – 409. Blumenthal, M., 1998. USP publishes information on ginger and valerian. Herbalgram 43, 30, 57, 71.
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545 Bonnet, M.H., Arand, D.L., 1997. Hyperarousal and insomnia. Sleep Med. Rev. 1, 97 – 108. Borbe´ly, A.A., Achermann, P., 1991. Ultradian dynamics of sleep after a single dose of benzodiazepine hypnotics. Eur. J. Pharmacol. 195, 11 – 18. Borbely, A.A., Achermann, P., 1992. Concepts and models of sleep regulation: an overview. J. Sleep Res. 1, 63 – 79. Brown, G.M., Stancer, H.C., Moldofsky, H., Harman, J., Murphy, J.T., Gupta, R.N., 1978. Withdrawal from long-term high-dose desipramine therapy. Clinical and biological changes. Arch. Gen. Psychiatry 35, 1261 – 1264. Cavadas, C., Arau´jo, I., Cotrim, M.D., Amaral, T., Cunha, A.P., Macedo, T., Ribeiro, F.C., 1995. In vitro study on the interaction of Valeriana officinallis L. extracts and their aminoacids on GABAa receptor in rat brain. Arzneim.-Forsch. (Drug Res.) 45, 753 – 755. Chiu, T.H., Rosenberg, H.C., 1978. Reduced diazepam binding following chronic benzodiazepine treatment. Life Sci. 23, 1153 – 1158. Donath, F., Quispe, S., Diefenbach, K., Maurer, A., Fietze, I., Roots, I., 2000. Critical evaluation of the effect of valerian extract on sleep structure and sleep quality. Pharmacopsychiatry 33, 47 – 53. Drewes, A.M., Gade, K., Nielsen, K.D., Bjerregard, K., Taagholt, S.J., Svendsen, L., 1995. Clustering of sleep electroencephalographic patterns in patients with the fibromyalgia syndrome. Br. J. Rheumatol. 134, 1151 – 1156. Feinberg, I., March, J.D., Fein, G., Floyd, T.C., Walker, J.M., Price, L., 1978. Period amplitude analysis of 0.5 – 3 c/sec activity in non-REM sleep of young adults. Electroencephalogr. Clin. Neurophysiol. 44, 202 – 213. Gaillard, J.M., Blois, R., 1989. Differential effects of flunitrazepam on human sleep in combination with flumazenil. Sleep 12, 120 – 132. Gebner, B., Klasser, M., 1984. Untersuchung der wirkung von Harmonicum Much auf den schlaf mit hilfe polygraphischer EEG — aufzeichnungen — 2. EEG – EMG 15, 45 – 51. Genton, P., Blin, O., Borderies, P., 1994. Substitution of chronic flunitrazepam by placebo or zolpidem: a multicentric, double-blind, parallel group study. J. Sleep Res. 3 (Suppl. 1), 172. Gorenstein, C., Bernik, M.A., Pompeia, S., 1994. Differential acute psychomotor and cognitive effects of diazepam on long-term benzodiazepine users. Int. Clin. Psychopharmacol. 9, 145 – 153. Griffiths, R.R., Weerts, E.M., 1997. Benzodiazepine self-administration in humans and laboratory animals implications for problems of long-term use and abuse. Psychopharmacolgy 134, 137. Jansen, W., 1977. Doppelblinstudie mit Baldrisedon. Therapiewoche 27, 2779 – 2786. Kamm-Kohl, A.V., Jansen, W., Brockmann, P., 1984. Moderne baldriantherapie gegen vervose storungen im. Serrium Medwelt. 35, 1450 – 1454. Kripke, D.F., Klauber, M.R., Wingard, D.L., Fell, R.L., Assmus, J.D., Garfinkel, L., 1998. Mortality hazard associated with prescription of hypnotics. Biol. Psychiatry 43, 687 – 693. Lader, M., Russel, J., 1993. Guidelines for the prevention and treatment of benzodiazepine dependence: summary of a report from the Mental Health Foundation. Addiction 88, 1707 – 1708. Leathwood, P.D., Chauffard, F., 1982. Quantifying the effects of mild sedatives. J. Psychiatr. Res. 17, 115 – 122. Leathwood, P.D., Chauffard, F., Heck, E., Munoz-Box, R., 1982. Aqueous extract of valerian root (Valeriana officinalis l.) improves sleep quality in man. Pharmacol., Biochem. Behav. 17, 65 – 71. Leathwood, P.D., Chauffard, F., Heck, E., Munoz-Box, R., 1985. Aqueous extract of valerian reduces latency to fall asleep. J. Psychiatr. Res. 17, 115 – 122. Lemoine, P., Ohayon, M.M., 1997. Is hypnotics withdrawal facilitated by the transitory use of a substitute drug? Prog. Neuro-Psychopharmacol. Biol. Psychiatry 21, 111 – 124. Merica, H., Blois, R., Gaillard, J.M., 1998. Spectral characteristics of sleep in chronic insomnia. Eur. T. Neurosci. 10, 1826 – 1839. Miller, L.G., Greenblatt, D.J., Roy, R.B., Summer, W.R., Shader, R.I., 1988. Chronic benzodiazepine administration: 11. Discontinuation syn-
545
drome is associated with upregulation of gamma-aminobutyric acid A receptor complex binding and function. J. Pharmacol. Exp. Ther. 246, 177 – 182. Mitler, M.M., Carskadon, M.A., Phillip, R.L., Sterling, W.R., Zarcone Jr., V.P., Spiegel, R., Guilleminault, C., Demnet, W.C., 1979. Hypnotic efficacy of temazepam: a long-term sleep laboratory evaluation. Br. J. Clin. Pharmacol. 8, 63S – 68S. Morin, C.M., Culbert, J.P., Schwartz, S.M., 1994. Nonpharmacological interventions for insomnia: a meta-analysis of treatment efficacy. Am. J. Psychiatry 151, 1172 – 1180. Murphy, S.M., Tyrer, P.J., 1991. A double-blind comparison of the effects of gradual withdrawal of lorazepam, diazepam and bromazepam in benzodiazepine dependence. Br. J. Psychiatry 158, 511 – 516. Nowell, P.D., Reynolds 3rd, C.F., Buysse, D.J., Dew, M.A., Kupfer, D.J. 1999. Paroxetine in the treatment of primary insomnia: preliminary clinical and electroencephalogram sleep data. J. Clin. Psychiatry 60, 89 – 95. Ohayon, M.M., Caulet, M., Priest, R.G., Guilleminault, C., 1998. Psychotropic medication consumption patterns in the UK general population. J. Clin. Epidemiol. 51, 273 – 283. Pailhous, E., Monfort, J.C., Goldenberg, F., Benoit, O., 1994. Switching off hypnotics under valproic acid in dependents insomniacs. J. Sleep Res. 3 (Suppl. 1), 379. Perlis, M.L., Giles, D.E., Bootzin, R.R., Dikman, Z.V., Fleming, G.M., Drummond, S.P., Rose, M.W., 1997. Alpha sleep and information processing, perception of sleep, pain, and arousability in fibromyalgia. Int. J. Neurosci. 89, 265 – 280. Rechtschaffen, A., Kales, A., 1968. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Los Angeles, Brain Information/Brain Research Institute, UCLA. Rickels, K., Case, W.G., Warren, G.C., Schweizer, E., Garcia-Espana, F., Fridman, R., 1990. Benzodiazepine dependence: management of discontinuation. Psychopharmacol. Bull. 26, 63 – 68. Ries, R.K., Roy-Byrne, P.P., Ward, N.G., Neppe, P., Cullison, S., 1989. Carbamazepine treatment for benzodiazepine withdrawal. Am. J. Psychiatry 143, 536 – 537. Santos, M.S., Ferreira, F., Cunha, A.P., Carvalho, A.P., Macedo, T., 1994a. An aqueous extract of valerian influences the transport of GABA in synaptosomes. Planta Med. 60, 278 – 279. Santos, M.S., Ferreira, F., Cunha, A.P., Carvalho, A.P., Ribeiro, C.F., Macedo, T., 1994b. Synaptosomal GABA release as by valerian root extract — involvement of the GABA carrier. Arch. Int. Pharmacodyn. Ther. 327, 220 – 231. Schulz, H., Stolz, C., Muller, J., 1994. The effect of valerian extract on sleep polygraphy in poor sleepers: a pilot study. Psychopharmacopsychiatry 27, 146 – 151. Schweizer, E., Rickels, K., Case, W.G., Greenblatt, D.J., 1991. Carbamazepine treatment in patients discontinuing long-term benzodiazepine therapy. Arch. Gen. Psychiatry 48, 448 – 452. Shapiro, C.M., 1994. Long term follow-up of hypnotic dependent subjects after successful withdrawal. J. Sleep Res. 3 (Suppl. 1), 469. Sher, P.K., Study, R.E., Mazzetta, J., Barker, J.L., Nelson, P.G., 1983. Depression of benzodiazepine binding and diazepam potentiation of GABA-mediated inhibition after chronic exposure of spinal cord cultures to diazepam. Brain Res. 268, 171 – 176. Smits, M.G., Declerck, A.C., Verbeek, I., 1996. Zolpidem improves sleep quality and architecture after abrupt discontinuation of chronic use of benzodiazepine treatment. J. Sleep Res. 5, 217. Tufik, S., Lobo, L.L., Medeiros, R., Neumann, B.G., Suchecki, D., 1992. Efeitos de um concentrado de valepotriatos sobre a sonoleˆncia diurna e paraˆmetros de sono noturno em volunta´rios sadios. F. Med. 104, 43 – 46. Tufik, S., Fujita, K., Seabra, M.L.V., Lobo, L.L., 1994. Effects of a prolonged administration of valepotriates in rats on the mother and their offspring. J. Ethnopharmacol. 41, 39 – 44. Vorbach, E.U., Gortelmeyer, R., Brunine, J., 1996. Therapie von insomnien: wirksamkeit und vertratich keit eines Baldrianpraparts. Psychopharmakotherapie 3, 109 – 115.
Article
Can valerian improve the sleep of insomniacs after benzodiazepine withdrawal? Dalva R. Poyaresa,b, Christian Guilleminaultb,*, Maurice M. Ohayonb, Sergio Tufika a
Sleep Laboratory of the Department of Psychobiology, Universidade Federal de Sa˜o Paulo, Sao Paulo, Brazil b Stanford University Sleep Disorders Center, Suite 3301, 401 Quarry Road, Stanford, CA 94305, USA
Abstract Purpose: The authors studied the sleep of patients with insomnia who complained of poor sleep despite chronic use of benzodiazepines (BZDs). The sample consisted of 19 patients (mean age 43.3 ± 10.6 years) with primary insomnia (DSM-IV), who had taken BZDs nightly, for 7.1 ± 5.4 years. The control group was composed of 18 healthy individuals (mean age 37 ± 8 years). Sleep electroencephalogram (EEG) of the patients was analyzed with period amplitude analysis (PAA) and associated algorithms, during chronic BZD use (Night 1), and after 15 days of a valerian placebo trial (initiated after washout of BZD, Night 2). Sleep of control subjects was monitored in parallel. Results: Valerian subjects reported significantly better subjective sleep quality than placebo ones, after BZD withdrawal, despite the presence of a few side effects. However, some of the differences found in sleep structure between Night 1 and Night 2 in both the valerian and placebo groups may be due to the sleep recovery process after BZD washout. Example of this are: the decrease in Sleep Stage 2 and in sigma count; the increase in slow-wave sleep (SWS), and delta count, which were found to be altered by BZD ingestion. There was a significant decrease in wake time after sleep onset (WASO) in valerian subjects when compared to placebo subjects; results were similar to normal controls. Nonetheless, valerian-treated patients also presented longer sleep latency and increased alpha count in SWS than control subjects. Conclusions: The decrease in WASO associated with the mild anxiolytic effect of valerian appeared to be the major contributor to subjective sleep quality improvement found after 2-week of treatment in insomniacs who had withdrawn from BDZs. Despite subjective improvement, sleep data showed that valerian did not produce faster sleep onset; the increase in alpha count compared with normal controls may point to residual hyperarousabilty, which is known to play a role in insomnia. Nonetheless, we lack data on the extent to which a sedative drug can improve alpha sleep EEG. Thus, the authors suggest that valerian had a positive effect on withdrawal from BDZ use. D 2001 Elsevier Science Inc. All rights reserved. Keywords: Benzodiazepines; Drug withdrawal; Insomnia; Period amplitude analysis; Sleep EEG; Valerian
1. Introduction The chronic use of benzodiazepines (BZDs) in the treatment of insomnia causes several problems such as dependence, tolerance, long-term memory impairment, and altered sleep structure (Griffiths and Weerts, 1997; Murphy and Tyrer, 1991; Gorenstein et al., 1994; Lader and Russel, 1993). It is also associated with increased mortality risk
Abbreviations: Benzodiazepine, BZD; Electrocardiogram, ECG; Electroencephalogram, EEG; Electromyogram, EMG; Electrooculogram, EOG; First-Derivative Analysis, FDA; Period amplitude analysis, PAA; Rapid eye movement, REM; Nonrapid eye movement, NREM; Wake time after sleep onset, WASO * Corresponding author. Tel.: +1-650-723-6601; fax: +1-650-7258910. E-mail address: [email protected] (C. Guilleminault).
(Kripke et al., 1998). In animal models, the chronic use of BZDs results in a decreased number of GABA-A receptors in the brain (Chiu and Rosenberg, 1978; Miller et al., 1988; Sher et al., 1983). However, the causal relation between this finding and tolerance of and dependence on these substances is not completely understood. Because of the problems associated with long-term use of BZDs, efforts have been made to manage successful withdrawal. However, most authors have used substances that act upon the GABAergic system with intermittent reports of success (Ries et al., 1989; Rickels et al., 1990; Schweizer et al., 1991; Pailhous et al., 1994; Lemoine and Ohayon, 1997; Shapiro, 1994; Genton et al., 1994; Smits et al., 1996). Valerian, as a naturally occurring root used in many countries by patients with sleep complaints, may be useful as an alternative treatment in withdrawal from BZDs use. In rats, valerian prevents the development of the withdrawal
0278-5846/01/$ – see front matter D 2001 Elsevier Science Inc. All rights reserved. PII: S 0 2 7 8 - 5 8 4 6 ( 0 1 ) 0 0 3 0 5 - 0
540
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545
syndrome resulting from the removal of diazepam following prolonged periods of administration (Andreatini and Leite, 1994) and does not induce toxic effects (Tufik et al., 1994). The fact that valerian is able to displace in vitro [3H]muscimol, a GABA agonist (Cavadas et al., 1995), suggests that it may act on the GABA receptor. Further evidence suggests that valerian effects are mediated by GABA (Santos et al., 1994a,b). Since 1977 (Jansen, 1977), several randomized, doubleblind trials have considered the effect of valerian extract on sleep. The evaluations were initially based on symptom scores and questionnaires (Kamm-Kohl et al., 1984; Vorbach et al., 1996), but several investigations used actigraphy or polygraphic recordings (Leathwood et al., 1982a, 1985; Balderer and Borbe´ly, 1985; Gebner and Klasser, 1984; Schulz et al., 1994; Donath et al., 2000). The surveyed populations were often healthy volunteers (young or elderly), poor sleepers, or chronically ill seniors. Results were variable but usually improved sleep latency was reported. Balderer and Borbe´ly (1985) reported shorter sleep latency and wake time after sleep onset. Similarly, Schulz et al. (1994) reported a significant increase in slow-wave sleep (SWS) and K-complex density after 8 days, and a decrease in Stage 1 nonrapid eye movement (NREM) sleep in elderly poor sleepers treated with valerian root. These authors suggested a selective effect of valerian extract on NREM sleep. Overall, the results were usually positive when poor sleepers were the targeted group rather than healthy volunteers. Valerian is considered a mild sedative and hypnotic substance in the presence of a lack of daytime somnolence and toxicity (Tufik et al., 1992) or alterations of sleep architecture in rats and in humans (Leathwood and Chauffard, 1982). Because of the variability of results, however, the United States Pharmacopoeia (USP) (Blumenthal, 1998) did not recommend valerian as a short-term treatment for insomnia. Because of the reported effects of valerian, we hypothesized that it may be helpful after withdrawal from chronic BZDs use. We asked how much improvement in sleep could be obtained after BZD withdrawal through ingestion of valepotriates, the active ingredient of valerian root. We also sought to discover whether valerian treatment after BZD withdrawal could improve sleep in insomnia patients and have it approach the values found in a healthy control group.
2. Methods 2.1. Subjects For 3 months at the sleep clinic at the Federal University of Sao Paulo, Brazil, the authors screened insomnia patients who continued to sleep poorly after long-term use of BZDs. All insomnia and control subjects were submitted to clinical interviews and clinical evaluations including bio-
chemical and hormonal tests to exclude medical disorders. These evaluations consisted of a structured psychiatric interview (SCID) after which subjects were classified according to DSM-IV criteria (American Psychiatry Association, 1994) and those with major psychiatric disorders were excluded from the selection. Sleep disorders other than insomnia in the patient group were excluded in the first recording night at sleep laboratory. Nineteen insomniac subjects (15 women and 4 men, mean age 43.3 ± 10.6 years) met the criteria for primary insomnia (DSM-IV). They had taken BZDs daily for at least 2.5 years and also met the DSM-IV BDZ drug-dependency criteria, registering a positive score and 4/7 of the listed criteria for this disorder. They agreed to undergo the protocol and signed informed consent forms. Eighteen healthy individuals (nine men, nine women; mean age 37 ± 8 years) formed the control group. 2.2. Protocol design Subjects were monitored with polysomnography at baseline while taking BZD (Night 1). Drug use was gradually withdrawn over a 2-week period and at the end of the withdrawal period they were kept without medication for 48 h and their sleep recorded again. After completion of withdrawal, subjects were then randomly assigned to a placebo/valerian trial with 50% of the subjects receiving the active ingredient, a concoction of valepotriates called Valmane, and the other half receiving a placebo in a double-blind administration. A second nocturnal polysomnographic recording was performed after 15 days of treatment with both the active drug and placebo (Night 2). The valerian concoction contained 80% didrovaltrate, 15% valtrate, and 5% acelvaltrate from the Valerian wallichii root and was administered three times daily in 100-mg doses obtained from a commercially available product, Valmane (Whitehall Pharmaceutical). 2.3. Withdrawal procedure All 19 insomniacs were taking BZD medication daily. The most frequently used BZD was lorazepam (3 patients), followed in order of decreasing frequency of intake by bromazepam and flurazepam (10 patients), midazolam (4 patients), clonazepam (3 patients), flunitrazepam (2 patients), and finally alprazolam and estazolam (1 patient each). Five patients took two different BZDs each night. The withdrawal schedule was calculated on the basis of total BZD dosage intake. All subjects received only BZDs and all doses were within the therapeutic range recommended by regulatory agencies. Subjects who received two BZD drugs were never at the highest range of the recommended dosage, even taking their combined BZD dosage. Total dosage was then reduced in two stages to 50% by Day 7, gradually reaching complete discontinuance by Day 14. During withdrawal, all subjects were followed clinically
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545
every 2 days and could also contact the supervising physician at any other time if needed. This close follow-up was put in place to provide support during withdrawal and ensure compliance with the prescribed BZD withdrawal schedule. No other instructions, except sleep hygiene (such as avoidance of alcohol and caffeine during the entire course of the study), were given to patients. In order to encourage and facilitate this follow-up, patients were provided with transportation, and in some cases, there was support for other expenses. 2.4. Data collection 2.4.1. Sleep diaries Sleep diaries were given to patients over the course of the study to assess sleep subjective quality, quantity, occurrence of naps, daily pattern of medication intake, and any consumption of caffeine or alcohol. Patients updated their diaries every morning. They also included three consecutive questions to check for BZD relapse, and one to check correct times of intake of valerian and placebo. The sleep diaries contained a visual– analogical scale of sleep quality with values ranging from 0 (the worst sleep) to 10 (the best sleep quality) to be recorded daily by the patients. 2.4.2. Urine BZD screening Urine BZD screening was collected twice: before the beginning of the study, and in the morning before the last night in the sleep laboratory (i.e., 15 days after BZD withdrawal), Night 2. The threshold considered positive for BZD usage was 200 ng or more as specified in the commercial kit. 2.4.3. Sleep recording Sleep was investigated on three different nights during the course of the study. After one night of habituation in the sleep laboratory with full monitoring, subjects were recorded prior to the beginning of drug withdrawal (experimental Night 1), and 15 days into the double-blind valerian/ placebo study (experimental Night 2). Normal controls were submitted to the same recording conditions including habituation and one test recording night (experimental Night 1) in the same laboratory setting. Polygraphic recording monitored the same parameters on experimental Nights 1 and 2, in insomnia patients and on Night 1 in controls. The following channels were systematically recorded: electroencephalogram [EEG] (C3 – A2; C4 – A1), submental and legs electromyogram [EMG], right and left electrooculograms (EOG), electrocardiogram [ECG] (with a modified V2 lead), airflow, chest and abdominal respiratory effort, and pulse oxymetry. 2.5. Data analyses During the habituation night, absence of other sleep disorders (such as sleep apnea syndrome, periodic leg
541
movement syndrome, etc.) was confirmed for both insomnia group and normal controls. Sleep/wake state on Nights 1 and 2 in the insomnia group and Night 1 in controls was scored blindly in relation to the experimental groups. Sleep stages were tabulated for 30-s epochs using the Rechtschaffen and Kales (1968) international scoring manual. 2.5.1. Sleep parameters The following sleep parameters were analyzed: sleep latency, total sleep time, sleep efficiency, number of arousals (16 s or longer), wake time after sleep onset (WASO), rapid eye movement (REM) and NREM stages as percentage of sleep time and REM density, defined as a rate representing the number of eye movements per REM sleep stage. 2.5.2. Sleep EEG The sleep EEG from each recorded night was downloaded on an IBM PC compatible, with a sampling rate of 128 Hz per channel. Sleep analysis of EEG frequencies contained in each band width during total sleep time was assessed using the period amplitude analysis (PAA) software. This technique measures EEG waves that cross the zero-volts baseline. Nonetheless, the isolated use of PAA is limiting because it is not capable of detecting fast- and low-amplitude EEG frequencies, which do not produce zero-crossings (Feinberg et al., 1978). Consequently, the complementary technique, First-Derivative Analysis (FDA), which is capable of detecting faster EEG frequencies, was added to the analysis. FDA is described in detail by Armitage (1995) and Armitage et al. (1995). The authors used a commercially available software program, which includes PAA and FDA. These analyses were performed in 30-s epochs. We considered each NREM sleep stage (and REM sleep) in totality, i.e., independent of the cycle in which the epoch was recorded for the analysis. The data are presented as averaged count (i.e., the averaged number of bursts of the specific EEG frequency band per minute of given sleep stage). The following EEG frequency bands were considered for analysis, because they are the most affected by BZD usage: delta (0.5 – 3.5 Hz), alpha (8 –11.5 Hz), and sigma (12 –16 Hz). 2.6. Statistical analysis 2.6.1. Sleep parameters A MANOVA was performed, with Night (1 and 2) and Treatment (valerian or placebo) as the main factors; the level of significance set was at P < .05. 2.6.2. Sleep EEG analysis A MANOVA was also performed to compare the sleep EEG frequency bands during Nights 1 and 2, using the same above-mentioned design. Student’s t test was performed to compare the values obtained from the sleep quality analogic –visual scale.
542
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545
3. Results 3.1. General results The average duration of the continuous BDZ intake was 7.1 ± 5.4 years, minimum of 2.5, and maximum of 14 years. Sleep diaries showed no reports of regular napping during daytime or intake of caffeine and alcohol over the course of the study. The urine analysis results were negative for BZD usage after withdrawal (i.e., values were < 200 ng). 3.2. Subjective data Analysis of next-morning sleep evaluation showed that valerian subjects reported on average significantly better sleep than placebo ones during the second week of treatment. The sleep quality analogical –visual scale average was 7.4 ± 0.9 for valerian subjects, against 5.4 ± 0.8 for placebo subjects ( P=.000). However, in the first week of treatment, three patients taking valerian complained of gastrointestinal symptoms, such as diarrhea, stomachache, and bitter taste in their mouth. Since these symptoms were limited, there was no need to open the codes, and these patients were allowed to continue in the protocol. 3.3. Analysis of sleep data The subsequent analyses were performed in all 19 insomnia subjects who completed their withdrawal from BZDs. Table 1 presents the results obtained on Night 1 (the baseline recording) on the control subjects when compared to the insomnia individuals after 15 days of treatment. As seen in Table 1, total sleep time and sleep latency were still significantly higher in control subjects, but not WASO, which was similar in both groups. All sleep stage percentage values, as well as REM density values, were also similar to those from normal controls. The sleep EEG data comparison showed a significant increase in alpha count during Sleep Stages 3 and 4 NREM in valerian-treated subjects.
3.3.1. Sleep parameters Table 2 presents the data obtained on experimental Night 1 (i.e., during BZD chronic use) and Night 2 (i.e., after the BZD withdrawal and 15 days treatment with placebo/valerian). These results show that patients experienced improved nocturnal sleep on Night 2, with a significant increase in SWS and deep sleep and a reduction in Stage 2 NREM sleep compared to the baseline (Night 1) in both valerian and placebo groups. REM density was also higher on Night 2 for both groups. Nonetheless, valerian-treated subjects showed a significant decrease in WASO in comparison to placebo. Conversely, the placebo group presented significantly shorter sleep latency. The other sleep parameters, as well as the sleep EEG ones, were similar between the two treatment groups. 3.3.2. Sleep EEG analysis Comparison between Nights 1 and 2 showed that BZD withdrawal was associated with a significant increase in ‘‘delta count’’ (the averaged number of bursts of delta per minute during this sleep stage). There was a significant increase in delta count in Stages 2, 3, and 4 NREM sleep on Night 2 compared to Night 1 for both placebo and valerian subjects. A significant decrease in 12 – 16 Hz, waves (labeled ‘‘sigma’’ and covering spindles and ‘‘BDZ spindles’’) was also observed in both groups during Stages 2, 3 NREM and REM sleep. In short, when compared to Night 1, both groups showed clear changes on Night 2 in Stages 2, 3, and 4 NREM sleep for all frequencies (as reported in Table 3). There was a significant increase in alpha and delta band count in REM sleep (outlined in Table 3).
4. Discussion This clinical study looks at a group of subjects that is rarely studied polysomnographically: chronic users of BZDs with regular drug intake who continue to complain
Table 1 Sleep stages and sleep EEG parameters obtained after BDZ withdrawal and treatment with valerian compared to a normal control group Sleep parameters
Night 1, control (n = 18)
Night 2, placebo group (n = 9)
Night 2, valerian group (n = 10)
F and P values
Sleep latency (min) Total sleep time (min) WASO (%) Sleep efficiency (%) REM latency (min) Stage 2 (%) Stages 3 and 4 (%) REM sleep (%) Alpha count Stage 3 Alpha count Stage 4
8.9 ± 16.2 418.0 ± 36.5 12.3 ± 15.9 89.6 ± 6.8 70.8 ± 33.0 52.5 ± 7.0 19.3 ± 7.9 21.3 ± 4.4 1.5 ± 1.2 0.8 ± 1.0
5.6 ± 5.5* 361.2 ± 59.7* 19.3 ± 6.9 79.2 ± 6.8* 123.6 ± 136.5 54.7 ± 12.3 18.3 ± 10.4 20.7 ± 9.6 3.3 ± 1.6* 3.0 ± 1.7*
30.8 ± 31.8* 356.2 ± 67.7* 12.6 ± 6.5 81.4 ± 8.0* 125.4 ± 74.2 50.7 ± 9.4 23.9 ± 8.4 19.9 ± 5.3 4.5 ± 1.8* 3.8 ± 1.7*
4.83; P =.0142 6.04; P =.006 1.09; n.s. 7.90; P =.002 2.09; n.s. 0.47; n.s. 1.23; n.s. 0.16; n.s. 13.77; P < .0001 17.39; P < .0001
The values are presented as mean ± S.D. * Differs from control group, P < .05; Student’s – Newman – Keuls test.
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545
543
Table 2 Parameters of sleep structure and percentage of sleep stages on two PSG recording nights Placebo (n = 9)
Valerian (n = 10)
Sleep parameter
Night 1 (baseline)
Night 2 (after treatment)
Night 1 (baseline)
Night 2 (after treatment)
Sleep latency (min) Sleep efficiency Total sleep time (min) WASO (%) REM latency (min) REM density Stage 2 (%) Stage 3 (%) Stage 4 (%) REM sleep (%)
22.6 82.1 383.7 14.6 140.1 2.6 60.4 6.8 5.5 19.6
5.5 79.1 361.1 19.3 123.6 2.5 54.7 10.2 8.0 20.7
25.2 81.4 377.9 16.2 137.2 1.9 64.3 8.4 3.3 16.8
30.7 81.4 356.1 12.5 125.4 3.2 50.6 14.3 9.5 19.8
( ± 15.1) ( ± 12.9) ( ± 50.0) ( ± 10.4) ( ± 69.4) ( ± 1.8) ( ± 8.7) ( ± 6.0) ( ± 5.6) ( ± 5.7)
( ± 5.4) ( ± 6.8) ( ± 59.7) ( ± 6.9) ( ± 136.4) ( ± 1.3)* ( ± 12.2)* ( ± 5.5)* ( ± 7.2)* ( ± 9.5)
( ± 26.6) ( ± 7.8) ( ± 34.1) ( ± 7.8) ( ± 52.9) ( ± 1.4) ( ± 9.3) ( ± 5.2) ( ± 4.8) ( ± 4.6)
( ± 31.8)# ( ± 7.9) ( ± 67.6) ( ± 6.4)# ( ± 74.1) ( ± 1.9)* ( ± 9.4)* ( ± 7.0)* ( ± 7.2)* ( ± 5.2)
F(1,17)
F(1,17)
F(1,17)
A
B
C
n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.
n.s. n.s. n.s. n.s. n.s. 6.4; P = .02 12.5; P = .002 6.8; P = .02 11.7; P = .003 n.s.
10.8; P = .004 n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s. n.s.
Values are presented as mean ± S.D. MANOVA: A = group effect; B = night effect; C = interaction effect; N.S. = nonsignificant. * Differs from Night 1; P < 0.05, MANOVA. # Differs from placebo.
of insomnia. Chronic use was defined here as nightly usage for over 1 year, but our population had a much longer regular intake with a mean of 7 years and a minimum of 2.5 years. There is a lack of recognized routines and procedures for withdrawal of BZD use in insomnia patients after chronic usage. In addition, as mentioned in the Introduction, several drugs have been tested to help withdrawal of BZDs, but in general, they may not present clear advantages compared to BZDs. We acknowledge that some chronic insomnia patients show a tendency toward chronic use of psychoactive substances (Ohayon et al., 1998), despite the progress achieved with nonpharmacological treatments of insomnia. Herbs and natural substances have been also consumed and well accepted by these types of patients. We chose valerian (i.e., valepotriates) from these natural substances, because of its convenient pharmacological properties as a mild anxiolytic and hypnotic with minor side effects and apparent absence of toxicity, as mentioned in the Introduction.
4.1. Is valerian effective in BZD withdrawal? Most of the differences found in sleep structure between Nights 1 and 2 in both groups, valerian and placebo, may be due to the sleep recovery process after BZD washout. Examples of these are: the decrease in Sleep Stage 2 and in sigma count; the increase in SWS, delta count, and REM density which were found to be altered by BZD ingestion in some studies (Gaillard and Blois, 1989; Borbe´ly and Achermann, 1991; Mitler et al., 1979), reporting on longterm (54 nights) use of temazepam in insomniacs. Although REM latency remained above 120 min in both groups of treatment, this was not significantly different from control value. Therefore, it is reasonable to suppose that some of these improvements in sleep could result solely from BZD removal. The significant decrease in WASO in valerian subjects is believed to account for valerian subjects reporting more satisfaction with their sleep than placebo ones. Curiously, valerian subjects presented higher sleep latency,
Table 3 Parameters of sleep EEG analysis F Placebo (n = 9)
EEG frequency band count Stage 2 Stage 3 Stage 4 REM sleep
Delta Sigma Delta Sigma Delta Alpha Delta Sigma
Valerian (n = 10)
Night 1
Night 2
Night 1
Night 2
Group effect, F(1,17)
Night effect, F(1,17)
8.3 ± 2.3 7.2 ± 2.1 24.4 ± 4.2 5.9 ± 4.3 22.4 ± 22.0 2.2 ± 1.2 1.5 ± 0.8 0.5 ± 0.3
12.0 ± 4.4* 4.0 ± 1.7* 32.4 ± 2.3* 4.2 ± 2.4* 42.8 ± 19.4* 2.2 ± 1.8 6.1 ± 9.8* 0.2 ± 0.1*
8.7 ± 4.5 8.5 ± 2.4 20.9 ± 4.4 9.5 ± 4.7# 18.3 ± 20.0 1.6 ± 0.8 1.2 ± 0.9 0.4 ± 0.3
14.6 ± 7.0* 4.1 ± 1.8* 35.9 ± 3.0* 5.4 ±2.7* 46.0 ± 17.9* 2.8 ± 1.6* 3.7 ± 3.4* 0.3 ±0.1*
n.s. n.s. n.s. 5.07; P = .04 n.s. n.s. n.s. n.s.
11.4; P = .004 44.5; P = .000 14.0; P = .002 4.6; P = .05 16.4; P = .001 4.6; P = .04 4.4; P = .05 4.8; P = .04
Values are expressed as mean ± S.D. MANOVA: F = group and night effects; n.s. = nonsignificant. There was no interaction between the two treatment groups. * Differs from respective Night 1; P < .05. # Differs from Night 2; P < .05.
544
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545
which may have resulted either from individual differences in insomnia markers severity, or from different responses to BZD withdrawal (i.e., some valerian treated subjects may have presented a longer sleep latency than their placebo counterparts before the treatments started, soon after BZD washout). Therefore, these results may not suggest that valerian failed to promote faster sleep onset. When compared to normal controls, valerian subjects also presented increased sleep latency, which accounted for the decrease in TST, but when compared to placebo, valerian was able to improve WASO to the level found in healthy control subjects. 4.2. What are the meanings of EEG findings in valerian group? The sleep EEG findings, such as the increase in alpha count in SWS in valerian group when compared to normal controls, as well as the increase in delta count for both groups on Night 2 when compared to Night 1 (during BDZ use), are more difficult to interpret. Increased alpha power during SWS is a finding already related to nonrestorative sleep, found, for example, in fibromyalgic and anxious patients (Perlis et al., 1997; Drewes et al., 1995). However, in this case, this higher alpha count may be due to residual hyperarousability, which is characteristic of insomnia patients (Bonnet and Arand, 1997), compared to normal controls. This raises a question: To what extent can sedative – hypnotics improve high alpha EEG in insomnia patients? The increase in delta frequency associated with augmented Stages 3 and 4 NREM sleep percentage may be considered as evidence of better consolidation of sleep, in these patients, after drug removal (Brown et al., 1978; Borbely and Archermann, 1992). We believe it may indicate a delta and SWS recovery process following BZD withdrawal, despite the fact that SWS and delta activity were not expected to present values close to the control group’s levels, since a decrease in SWS and delta activity has been reported in chronic insomnia patients (Merica et al., 1998). However, there is very little knowledge on sleep recovery after very chronic BZD usage and valerian treatment. Finally, previous studies with valerian were usually positive when poor sleepers were targeted and compared to healthy volunteers. This may be the case for this study, since we were able to demonstrate a decrease in WASO in insomnia BZD chronic users. As mentioned above, in the Introduction, valerian is more likely to shorten sleep latency than to improve sleep maintenance, but a population of BZD chronic users has never been the target of a valerian trial. Polysomnographic findings may not completely explain subjective reports from insomnia patients. Subjective changes in sleep perception do not always correlate with objective sleep data. Nowell et al. (1999) found that paroxetine showed a significant subjective improvement of sleep in insomniacs, but there was no improvement in PSG
data. In this study, however, the data suggest that the positive subjective effect of valepotriates compared to placebos after BZD washout could be due to the decrease in WASO, but particularly because of the anxiolytic effect that may have played an important role in the subjective perception of improved sleep quality.
5. Conclusions In the light of the present results, and those of other studies (Vorbach et al., 1996; Leathwood et al., 1985; Schulz et al., 1994), all of which showed positive findings with valerian intake, there is a need for more research into valepotriates that should focus on a larger but homogenous group of patients with insomnia complaints. Moreover, this is supported by data in rats showing that valerian was able to taper BDZ withdrawal signs (Andreatini and Leite, 1994). Due to (a) objective sleep data in chronic BDZ users remaining different to normal controls after withdrawal; (b) sleep data being slightly improved after a 2-week treatment with valerian; (c) the persistence of increased alpha during REM sleep; and (d) the lack of data on predictors of BDZ relapse in insomniacs after chronic use, we recommend that concurrent behavioral techniques to improve the possible hyperarousal condition in these insomniacs should be considered (Morin et al., 1994), even though patients report subjective improvement in sleep.
Acknowledgments The authors would like to thank Dr. Deborah Suchecki for her comments on the manuscript. This work was supported by Associac¸a˜o Fundo de Incentivo a` Psicofarmacologia (AFIP). Dalva Poyares was the recipient of a fellowship from Coordenac¸a˜o de Aperfeic˛oamento de Pessoal de Ensino Superior (CAPES).
References American Psychiatry Association, 1994. Diagnostic and Statistical Manual of Mental Disorders (DSM-IV), fourth ed. American Psychiatry Press, Washington. Andreatini, R., Leite, J.R., 1994. Effect of valepotriates on the behaviour of rats on the elevated plus maze during benzodiazepine withdrawal. Eur. J. Pharmacol. 260, 233 – 235. Armitage, R., 1995a. The distribution of EEG frequencies in REM and NREM sleep stages in healthy young adults. Sleep 18, 334 – 341. Armitage, R., Hoffmann, R., Fitch, T., Morel, C., Bonato, R., 1995b. A comparison of period amplitude and power spectral analysis of sleep EEG in normal adults and depressed out patients. Psychiatry Res. 56, 245 – 256. Balderer, G., Borbe´ly, A.A., 1985. Effect of valerian on human sleep. Psychopharmacology 87, 406 – 409. Blumenthal, M., 1998. USP publishes information on ginger and valerian. Herbalgram 43, 30, 57, 71.
D.R. Poyares et al. / Progress in Neuro-Psychopharmacology & Biological Psychiatry 26 (2002) 539–545 Bonnet, M.H., Arand, D.L., 1997. Hyperarousal and insomnia. Sleep Med. Rev. 1, 97 – 108. Borbe´ly, A.A., Achermann, P., 1991. Ultradian dynamics of sleep after a single dose of benzodiazepine hypnotics. Eur. J. Pharmacol. 195, 11 – 18. Borbely, A.A., Achermann, P., 1992. Concepts and models of sleep regulation: an overview. J. Sleep Res. 1, 63 – 79. Brown, G.M., Stancer, H.C., Moldofsky, H., Harman, J., Murphy, J.T., Gupta, R.N., 1978. Withdrawal from long-term high-dose desipramine therapy. Clinical and biological changes. Arch. Gen. Psychiatry 35, 1261 – 1264. Cavadas, C., Arau´jo, I., Cotrim, M.D., Amaral, T., Cunha, A.P., Macedo, T., Ribeiro, F.C., 1995. In vitro study on the interaction of Valeriana officinallis L. extracts and their aminoacids on GABAa receptor in rat brain. Arzneim.-Forsch. (Drug Res.) 45, 753 – 755. Chiu, T.H., Rosenberg, H.C., 1978. Reduced diazepam binding following chronic benzodiazepine treatment. Life Sci. 23, 1153 – 1158. Donath, F., Quispe, S., Diefenbach, K., Maurer, A., Fietze, I., Roots, I., 2000. Critical evaluation of the effect of valerian extract on sleep structure and sleep quality. Pharmacopsychiatry 33, 47 – 53. Drewes, A.M., Gade, K., Nielsen, K.D., Bjerregard, K., Taagholt, S.J., Svendsen, L., 1995. Clustering of sleep electroencephalographic patterns in patients with the fibromyalgia syndrome. Br. J. Rheumatol. 134, 1151 – 1156. Feinberg, I., March, J.D., Fein, G., Floyd, T.C., Walker, J.M., Price, L., 1978. Period amplitude analysis of 0.5 – 3 c/sec activity in non-REM sleep of young adults. Electroencephalogr. Clin. Neurophysiol. 44, 202 – 213. Gaillard, J.M., Blois, R., 1989. Differential effects of flunitrazepam on human sleep in combination with flumazenil. Sleep 12, 120 – 132. Gebner, B., Klasser, M., 1984. Untersuchung der wirkung von Harmonicum Much auf den schlaf mit hilfe polygraphischer EEG — aufzeichnungen — 2. EEG – EMG 15, 45 – 51. Genton, P., Blin, O., Borderies, P., 1994. Substitution of chronic flunitrazepam by placebo or zolpidem: a multicentric, double-blind, parallel group study. J. Sleep Res. 3 (Suppl. 1), 172. Gorenstein, C., Bernik, M.A., Pompeia, S., 1994. Differential acute psychomotor and cognitive effects of diazepam on long-term benzodiazepine users. Int. Clin. Psychopharmacol. 9, 145 – 153. Griffiths, R.R., Weerts, E.M., 1997. Benzodiazepine self-administration in humans and laboratory animals implications for problems of long-term use and abuse. Psychopharmacolgy 134, 137. Jansen, W., 1977. Doppelblinstudie mit Baldrisedon. Therapiewoche 27, 2779 – 2786. Kamm-Kohl, A.V., Jansen, W., Brockmann, P., 1984. Moderne baldriantherapie gegen vervose storungen im. Serrium Medwelt. 35, 1450 – 1454. Kripke, D.F., Klauber, M.R., Wingard, D.L., Fell, R.L., Assmus, J.D., Garfinkel, L., 1998. Mortality hazard associated with prescription of hypnotics. Biol. Psychiatry 43, 687 – 693. Lader, M., Russel, J., 1993. Guidelines for the prevention and treatment of benzodiazepine dependence: summary of a report from the Mental Health Foundation. Addiction 88, 1707 – 1708. Leathwood, P.D., Chauffard, F., 1982. Quantifying the effects of mild sedatives. J. Psychiatr. Res. 17, 115 – 122. Leathwood, P.D., Chauffard, F., Heck, E., Munoz-Box, R., 1982. Aqueous extract of valerian root (Valeriana officinalis l.) improves sleep quality in man. Pharmacol., Biochem. Behav. 17, 65 – 71. Leathwood, P.D., Chauffard, F., Heck, E., Munoz-Box, R., 1985. Aqueous extract of valerian reduces latency to fall asleep. J. Psychiatr. Res. 17, 115 – 122. Lemoine, P., Ohayon, M.M., 1997. Is hypnotics withdrawal facilitated by the transitory use of a substitute drug? Prog. Neuro-Psychopharmacol. Biol. Psychiatry 21, 111 – 124. Merica, H., Blois, R., Gaillard, J.M., 1998. Spectral characteristics of sleep in chronic insomnia. Eur. T. Neurosci. 10, 1826 – 1839. Miller, L.G., Greenblatt, D.J., Roy, R.B., Summer, W.R., Shader, R.I., 1988. Chronic benzodiazepine administration: 11. Discontinuation syn-
545
drome is associated with upregulation of gamma-aminobutyric acid A receptor complex binding and function. J. Pharmacol. Exp. Ther. 246, 177 – 182. Mitler, M.M., Carskadon, M.A., Phillip, R.L., Sterling, W.R., Zarcone Jr., V.P., Spiegel, R., Guilleminault, C., Demnet, W.C., 1979. Hypnotic efficacy of temazepam: a long-term sleep laboratory evaluation. Br. J. Clin. Pharmacol. 8, 63S – 68S. Morin, C.M., Culbert, J.P., Schwartz, S.M., 1994. Nonpharmacological interventions for insomnia: a meta-analysis of treatment efficacy. Am. J. Psychiatry 151, 1172 – 1180. Murphy, S.M., Tyrer, P.J., 1991. A double-blind comparison of the effects of gradual withdrawal of lorazepam, diazepam and bromazepam in benzodiazepine dependence. Br. J. Psychiatry 158, 511 – 516. Nowell, P.D., Reynolds 3rd, C.F., Buysse, D.J., Dew, M.A., Kupfer, D.J. 1999. Paroxetine in the treatment of primary insomnia: preliminary clinical and electroencephalogram sleep data. J. Clin. Psychiatry 60, 89 – 95. Ohayon, M.M., Caulet, M., Priest, R.G., Guilleminault, C., 1998. Psychotropic medication consumption patterns in the UK general population. J. Clin. Epidemiol. 51, 273 – 283. Pailhous, E., Monfort, J.C., Goldenberg, F., Benoit, O., 1994. Switching off hypnotics under valproic acid in dependents insomniacs. J. Sleep Res. 3 (Suppl. 1), 379. Perlis, M.L., Giles, D.E., Bootzin, R.R., Dikman, Z.V., Fleming, G.M., Drummond, S.P., Rose, M.W., 1997. Alpha sleep and information processing, perception of sleep, pain, and arousability in fibromyalgia. Int. J. Neurosci. 89, 265 – 280. Rechtschaffen, A., Kales, A., 1968. A Manual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Los Angeles, Brain Information/Brain Research Institute, UCLA. Rickels, K., Case, W.G., Warren, G.C., Schweizer, E., Garcia-Espana, F., Fridman, R., 1990. Benzodiazepine dependence: management of discontinuation. Psychopharmacol. Bull. 26, 63 – 68. Ries, R.K., Roy-Byrne, P.P., Ward, N.G., Neppe, P., Cullison, S., 1989. Carbamazepine treatment for benzodiazepine withdrawal. Am. J. Psychiatry 143, 536 – 537. Santos, M.S., Ferreira, F., Cunha, A.P., Carvalho, A.P., Macedo, T., 1994a. An aqueous extract of valerian influences the transport of GABA in synaptosomes. Planta Med. 60, 278 – 279. Santos, M.S., Ferreira, F., Cunha, A.P., Carvalho, A.P., Ribeiro, C.F., Macedo, T., 1994b. Synaptosomal GABA release as by valerian root extract — involvement of the GABA carrier. Arch. Int. Pharmacodyn. Ther. 327, 220 – 231. Schulz, H., Stolz, C., Muller, J., 1994. The effect of valerian extract on sleep polygraphy in poor sleepers: a pilot study. Psychopharmacopsychiatry 27, 146 – 151. Schweizer, E., Rickels, K., Case, W.G., Greenblatt, D.J., 1991. Carbamazepine treatment in patients discontinuing long-term benzodiazepine therapy. Arch. Gen. Psychiatry 48, 448 – 452. Shapiro, C.M., 1994. Long term follow-up of hypnotic dependent subjects after successful withdrawal. J. Sleep Res. 3 (Suppl. 1), 469. Sher, P.K., Study, R.E., Mazzetta, J., Barker, J.L., Nelson, P.G., 1983. Depression of benzodiazepine binding and diazepam potentiation of GABA-mediated inhibition after chronic exposure of spinal cord cultures to diazepam. Brain Res. 268, 171 – 176. Smits, M.G., Declerck, A.C., Verbeek, I., 1996. Zolpidem improves sleep quality and architecture after abrupt discontinuation of chronic use of benzodiazepine treatment. J. Sleep Res. 5, 217. Tufik, S., Lobo, L.L., Medeiros, R., Neumann, B.G., Suchecki, D., 1992. Efeitos de um concentrado de valepotriatos sobre a sonoleˆncia diurna e paraˆmetros de sono noturno em volunta´rios sadios. F. Med. 104, 43 – 46. Tufik, S., Fujita, K., Seabra, M.L.V., Lobo, L.L., 1994. Effects of a prolonged administration of valepotriates in rats on the mother and their offspring. J. Ethnopharmacol. 41, 39 – 44. Vorbach, E.U., Gortelmeyer, R., Brunine, J., 1996. Therapie von insomnien: wirksamkeit und vertratich keit eines Baldrianpraparts. Psychopharmakotherapie 3, 109 – 115.