Pharmacology and Sleep

Transcription

1 Sleep. 16(4): American Sleep Disorders Association and Sleep Research Society Pharmacology and Sleep Sedative Effects and Plasma Concentrations Following Single Doses of Triazolam, Diphenhydramine, Ethanol and Placebo Timothy Roehrs, Ardith Zwyghuizen-Doorenbos and Thomas Roth Sleep Disorders and Research Center. Henry Ford Hospital. Detroit Michigan. U.S.A. Summary: Twelve, healthy, young men received.25 mg triazolam and ethanol placebo, 5 mg diphenhydramine and ethanol placebo,.6 g/kg ethanol and placebo pill and ethanol placebo and placebo pill in a double-blind Latin Square design. Each of the four treatments were administered for 2 days at 9 hours with blood samples drawn on day one at 83, 13, 123, 143 and 163 hours, and sleep latency and performance assessed on day two at 1,12,14 and 16 hours. Significant sedative effects of ethanol, triazolam and diphenhydramine relative to placebo were observed on the sleep latency and performance measures with the effects being detected over the fuji 6.5 hours of assessment. Among the active drugs, triazolam and diphenhydramine had similar sedative effects which differed from that of ethanol. Plasma concentration of each drug declined significantly over the 6.5 hours. Ethanol reached zero, but triazolam and diphenhydramine did not. Continued sedation (sedative effects after plasma concentration reached zero) was observed with ethanol. Key words: Alcohol-Diphenhydramine-Triazolam Hypersomnia - Electroencephalography. A recent Gallup poll indicated that very few individuals who complain of insomnia receive medical attention and benzodiazepine hypnotic prescriptions (the drug class of choice) for their insomnia (1). A percentage of those individuals with insomnia are prescribed one of the HI antihistamines (which cross the blood-brain barrier readily), or they report using overthe-counter (OTC) hypnotics, in which antihistamine typically is the active ingredient. Others with insomnia in the survey reported using ethanol as a hypnotic. No studies have compared the hypnotic effects of these drugs, much less merely assessed their relative sedative effects. Separate daytime studies have documented the sedative effects of triazolam, antihistamines and ethanol in waking, healthy "normals" using the Multiple Sleep Latency Test (MSLT) (2,3,4). But no single study using an objective measure of sedation (i.e. the MSLT) has compared the sedative effects of these drugs at commonly used doses and assessed the time course of Accepted for publication January Address correspondence and reprint requests to Timothy Roehrs, Ph.D., Henry Ford Hospital, 2921 West Grand Blvd., Detroit, Michigan 4822, U.S.A. 31 those effects relative to plasma concentrations of the drugs. The MSLT studies assessing the daytime sedative effects of ethanol have shown that ethanol-induced sedation continues after ethanol is no longer detectable in breath. The studies have found such continued sedation for as long as 4.5 hours (5,6,7). The same phenomenon also has been reported recently by another laboratory studying nighttime administration of ethanol in subjects remaining awake on a simulated night shift schedule (8). The question arises as to whether this continued sedation is unique to ethanol or whether other sedating drugs, such as the benzodiazepines and antihistamines, when administered to waking humans will show comparable continued sedation. No study has related plasma concentration to polysomnographic (PSG) measures of sedation. Studies using self-rated sedation and psychomotor performance tests have found that the sedative effects of benzodiazepines (9,1,11) and diphenhydramine (12,13,14) generally parallel the plasma concentrations. Such results suggest that ethanol is unique in showing continued sedation. This study compared the sedative effects and plasma

2 32 T. ROEHRS ET AL. concentrations over time of triazolam, diphenhydramine and ethanol at commonly used doses. It also attempted to determine whether the continued sedation seen in previous studies with ethanol also could be found with these other sedating drugs, when measured with PSG. Subjects METHODS The subjects were 12 healthy, non-smoking men, aged 21-4 years, who reported drinking an average of 1-14 alcoholic drinks per week. Subjects' good health and non-use oflicit and illicit drugs other than alcohol were confirmed by medical histories, physical examinations, drug-use histories and blood and urine tests. Subjects reported normal sleep habits, with nocturnal sleep latencies of 3 minutes or less, a lack of habitual daytime napping and no sleep difficulties. Each reported an average of 6-8 hours of sleep per night with little nightly variation in bedtime and risetime. Each signed an informed consent and was paid for participation. Procedure Those subjects who passed the initial screening and the physical examination were asked to come to the laboratory for one night oflaboratory adaptation. Sleep during the laboratory adaptation night was monitored by means of an actigraph, which monitored activity or inactivity during the night. The actigraph has been shown to predict polysomnographically determined sleep or wake with 93% accuracy (15). The Multiple Sleep Latency Test (MSLT) was administered at 1, 12, 14 and 16 hours the next day according to standard procedures (16). Prior to the MSLT, electrodes were placed at standard placements to record the electroencephalogram (central and occipital), electrooculograms (left and right outer canthi) and electromyogram (submental) (17). Subjects were awakened after one minute of unambiguous stage 1 sleep, the first sign of stage 2 or REM sleep or 2 minutes of continuous wake, according to the standards of Recht schaff en and Kales (17). Those subjects with a nighttime sleep efficiency of 85% or greater and mean sleep latencies on the MSLT of greater than 1 minutes were admitted into the study. The study was conducted in a complete, repeated measures design and consisted of four 2-day blocks. The 2-day blocks were separated by no less than 2 days and no more than 2 weeks. On each study night, subjects reported to the laboratory at 23 hours and went to bed at 233 hours. Their sleep was monitored by actigraph until they arose at 73 hours the next morn- Sleep, Vol. 16, No.4, 1993 ing. They were given breakfast consisting of a danish and a non-caffeinated beverage. On the first day of each 2-day block, an intravenous catheter (IV) was inserted into the lower right or left forearm soon after risetime. A heparin lock was attached to the catheter, and heparin was inserted to prevent blood clots between blood draws. Blood samples were taken from the IV site every 2 hours beginning at 83 hours and continued until 163 hours. After collection, samples were centrifuged and plasma layers were frozen. Gas chromatographic analyses were used to assess plasma concentrations of triazolam and diphenhydramine (18,19). On the second study day of each 2-day block, subjects received the same drug treatment and assessment of sleepiness and performance, but plasma sampling was not conducted. At 9 hours on every study day, subjects received either triazolam.25 mg and ethanol placebo, diphenhydramine 5 mg and ethanol placebo, ethanol.6 glkg and placebo pill or ethanol placebo and placebo pill. Placebo, diphenhydramine and triazolam were all prepared in identical capsules. The ethanol was 8 proof vodka, mixed in a 1:4 ratio with tonic water. Ethanol placebo was tonic water with three drops of 8 proof vodka floating on the surface for taste and odor. Both ethanol and placebo were flavored with lemon or lime juice. Subjects paced consumption of beverages over 3 minutes such that the entire drink was consumed steadily over the time allotted. Drinks were consumed in a nonsocial environment. These four drug conditions were presented to each subject in a sequence according to a Latin Square design. On each day, subjects were tested for sleep latency at 1, 12, 14 and 16 hours by means of the MSLT as described above. Immediately prior to each latency test, breath ethanol concentration was measured using an Alcotest 71 National Draeger Breathalyzer. Subjects also completed a short performance battery after each latency test. The battery included the Digit Symbol Substitution Test (DSST) and the Symbol Copying Test (SCT) with number of correct copies and correct substitutions as dependent measures. These were followed by a IS-minute, divided attention task. This task required tracking a moving target across a video display using a joystick, while responding (by pressing a key) to the appearance of a stimulus on the periphery or the center of the video display. Tracking deviations and reaction times (RT) to the peripheral and central stimuli were recorded. Subjects were allowed to leave the laboratory after the last performance test, and were required to return to the laboratory one hour before bedtime on the next study day. During the course of the study, subjects refrained from napping, caffeine and alcohol and drug use other than that required in the study.

3 SEDATION AND PLASMA CONCENTRATION E 15 E > o ~ 1 -c -' a.., ~ 5 III o Ethanol 6. Diphenhydramine o Trlazolam 'V Placebo OL-----~------~----~------~----~ Hour FIG. 1. Sleep latency (minutes) after placebo, triazolam (.2S mg), diphenhydramine (SO mg) and ethanol (.6 glkg). The MSLT was scored for sleep latency according to the standards of Rechtschaffen and Kales (17) by scorers unaware of the treatment condition. Sleep latency was defined as time from lights out to the first epoch of sleep. Sleep latency and performance measures on day two of each 2-day drug treatment were analyzed for sedative effects relative to placebo and time effects to drug administration. Day one sleep latency and performance data were not used because of the disruptive effects of the plasma sampling on these measurements. The day two data were analyzed for the two factors (drug condition and time since consumption) using the general linear models multivariate analysis of variance (SAS Institute), followed by post hoc contrasts where indicated. Conservative p-ievels ofp <.5, corrected by the Huynh-Feldt procedures, were used to determine significance. RESULTS Sleep deficiency (inactive time relative to bedtime) on the nights prior to each study day did not differ among conditions and ranged from %. Consequently, differences in prior amount of sleep do not account for the observed differences among drug conditions in sleep latency and performance. Sleep latency on each test (1, 12, 14 and 16 hours) during the second day of each drug treatment is presented in Fig. I. The omnibus analysis for the four drug conditions and the four tests (time since consumption) revealed a main effect of drug condition (F = 2.27, p <.1). In separate analyses of each drug compared to placebo, a significant reduction in sleep latency on the four tests was found with triazolam (F= 4.37, p <.1), diphenhydramine (F= 72.51, p <.1) and ethanol (F = 18.91, p <.1). In the overall analysis and in the separate analyses of each drug compared to placebo, there were no drug-by-test (time since consumption) interactions. The separate analyses comparing sleep latency among active drug pairs showed that triazolam (F = 5.1, p <.5) and diphenhydramine (F = 4.94, p <.5) both differed from ethanol, whereas triazolam and diphenhydramine did not differ. A drug-by-test time interaction was found in the triazolam vs. ethanol analysis (F = 4.5, p <.2) with the drugs differing on the first two tests, but not the last two tests. On the omnibus analysis there also was a main effect of test time (F= 7.89, p <.1). Mean sleep latency over all drug conditions was 8.72 ± 4.27 at 1 hours, 6.29 ± 3.26 at 12 hours, 5.6 ± 2.28 at 14 hours and 7.57 ± 3.1 at 16 hours. Post hoc contrasts showed sleep latency across all drug conditions was longest in the morning and lowest over the midday, which is the typical pattern of sleep latency reported in healthy adults. Sleep latency at test one (1 hours) was longer than sleep latency at each of the other three tests (F = 13.32, p <.5; F = 1.85, p <.9; F = 6.23, p <.3 for tests two, three and four). Sleep latency at test three (14 hours) was shorter than latency at test four (16 hours). The results of the digit symbol substitution test performance are presented in Table I. Digit symbol substitution test performance in an omnibus analysis comparing the four conditions and four test times was altered by drug condition (F = 9.22, p <.2). Performance was reduced by triazolam (F = 25.19, p <.7), diphenhydramine (F = 21.42, p <.9) and ethanol (F = 11.8, p <.8) as compared to placebo in separate analyses of each drug. In the analyses of active drug pairs, no differences among active drugs were observed. There were no interactions of drug condition-by-test time on this measure in the omnibus or the separate analyses. Divided attention performance is presented in Table 2. Drug effects on the divided attention performance TABLE 1. Digit symbol substitution performance Placebo Triazolam Diphenhydramine Ethanol Data are means ± standard deviations. 1 hours 76.8 ± ± ± S.9 ± hours 73. ± 14.S 6S.S ± ± IS ± IS.8 14 hours 77.1 ± ± ± ± hours 7S.4 ± ± ± ± 11.6 Sleep. Vol. 16. No

4 34 T. ROEHRS ET AL. TABLE 2. Divided attention performance Placebo Triazolam Dipenhydramine Ethanol Data are means ± standard deviations. were not as robust, with only triazolam disrupting divided attention performance relative to placebo. Both central (F = 11.5, p <.8) and peripheral (F = 5.91, p <.4) reaction times were slowed by triazolam. In comparisons among active drugs, no differences were seen. There were interactions of drug-bytest time on these measures (F = 4.17, p <. I and F = 9.26, p <.1) with triazolam and ethanol differing on the first two tests, but not the last two. Triazolam also differed from diphenhydramine in central reaction time on the first two tests, but not the last two (F = 9.26, p <.1). The plasma concentrations for each drug are presented in Fig. 2. The plasma concentration data were submitted to the same omnibus analysis as the sleep latency and performance data, with drug (the three active drugs) and time of test as factors. There was a main effect of drug (F = 17, p <.1) which merely indicates that each drug, as might be expected, had a different range of plasma concentrations. A main effect of time was also found (F = 1.83, p <.3), reflecting an overall change in concentration across time. Finally, there was a drug-by-time interaction (F = 1.26, p <.3), indicating differing patterns of plasma concentration change among the drugs. DISCUSSION This study showed sedative effects relative to placebo for triazolam, diphenhydramine and ethanol, as measured by the MSLT and performance tasks (2,3). The study showed comparable levels of sedation with triazolam and diphenhydramine which differed from that of ethanol. Further, as in previous studies, the sedative effects of ethanol continued after blood concentrations of ethanol had declined to zero (5,6,7,2). The secondary question of this study, whether similar continued sedation is found with triazolam and diphenhydramine, cannot be determined conclusively. 1 hours 12 hours 14 hours 16 hours 28. ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± 9.2 In this study, plasma concentrations oftriazolam and diphenhydramine declined, but never reached zero. Thus, whether the pattern of results in this study represents a continued sedation for triazolam and diphenhydramine, as previously shown for ethanol, is not clear. These would be the first data suggesting that the sedating effects of these two drugs are not parallel to their decline in plasma concentration. This study does differ from previous ones in that objective, rather than subjective, measures of sedation were used. Some studies have shown that PSG measures of sedation detect effects which cannot be found on subjective assessment and performance measures (21). Several interesting findings from the comparisons among active drugs should be discussed. The first point is the relative sedative effects of this ethanol dose compared to triazolam (.25 mg) and diphenhydramine (5 mg). The.6 glkg dose for this group of subjects is a relatively large dose. It is approximately 3-4 drinks with a drink defined as 1 oz of ethanol. The subjects recruited for this study were moderate drinkers, de c:: c::.75., u c u.5 E...25!!. D... 8 o Ethanol 6 Diphenhydramine ~ "'''''.m ~ ~~ o I Rx Hour FIG. 2. Plasma concentration of triazolam (nglml), diphenhydramine (mglml x 1) and ethanol (mglml x 1). Arrow indicates time of drug administration. 18 Sleep. Vol. 16. No

5 SEDATION AND PLASMA CONCENTRATION 35 fined as 2-14 drinks per week. This dose of ethanol did not sedate to the extent of that of triazolam and diphenhydramine. Often the point is made that ethanol is not an effective hypnotic because it does not adequately maintain sleep. These data would suggest that ethanol is a relatively poor sedative as well. The second point of interest to be made is the absence of consistent differences between triazolam and diphenhydramine at the doses of this study. Both represent the high clinical doses. Triazolam is available in.125 mg and.25 mg doses, whereas the 5 mg diphenhydramine dose is a prescription dose and 25 mg is the typical OTe dose. The two drugs did not differ appreciably in their sedative effects. Although the power of this study to detect differences may be questioned, the point should be made that differences with ethanol were detected in this study. The clinician should then appreciate the extent of the sedative effects of diphenhydramine when used in waking individuals for its antihistaminic effects. On the other hand, these data do not necessarily indicate that antihistamines should be considered effective hypnotics. That determination requires further and more extensive assessment. Finally, comment can be made regarding the relative sensitivity of the MSLT and the performance tests to the sedative effects of this study. Differences between drugs and placebo and differences among drugs were seen with the MSLT, whereas the performance tests of this study did not show such consistent sedative effects. This illustrates the importance of using both performance and PSG methods of assessing the sedative potential of a given drug and dose. Acknowledgements: This work was supported by National Institutes of Health grant No. RO I-AA714 7 awarded to Dr. T. Roehrs. We thank Dr. David J. Greenblatt who did the diphenhydramine and triazolam plasma concentration assays. REFERENCES 1. The Gallup Organization. Sleep in America: a national survey of U.s. adults. Princeton, New Jersey, Roehrs T, Tietz E, Zorick F, Roth T. Daytime sleepiness and antihistamines. Sleep 1984;7: Roehrs T, Zwyghuizen-Doorenbos A, Smith D, Zorick F, Roth T. Reversal by caffeine of triazo1am-induced impairment of waking function. In: Hindmarch I, Ott H, eds. Benzodiazepine receptor ligands, memory, and information processing. Berlin: Springer-Verlag, 1988: Gengo F, Gabos C, Miller K. The pharmacodynamics of diphenhydramine-induced drowsiness and changes in mental performance. Clin Pharmacol Ther 1989;45: Zwyghuizen-Doorenbos A, Roehrs T, Lamphere J, Zorick F, Roth T. Increased daytime sleepiness enhances ethanol's sedative effects. Neuropsychopharmacology 1988; 1 : Roehrs T, Zwyghuizen-Doorenbos A, Zwyghuizen H, Roth T. Sedating effects of ethanol after a nap. Alcohol, Drugs, & Driving 199;6: Zwyghuizen-Doorenbos A, Roehrs T, Timms V, Roth T. Individual differences in the sedating effects of ethanol. Alcohol Clin Exp Res 199; 14: Walsh JK, Humm T, Muehlbach MJ, Sugarman JL, Schweitzer PK. Sleepiness/alertness at night following ethanol consumption. Sleep Res 1989;18: Greenblatt DJ, Harmatz JS, Dorsey S, Shader RI. Comparative single-dose kinetics and dynamics of lorazepam, alprazo1am, prazepam, and placebo. Clin Pharmacol Ther 1988;44: Greenblatt DJ, Harmatz JS, Engelhardt N, Shader RI. Pharmacokinetic determinants of dynamic differences among three benzodiazepine hypnotics. Arch Gen Psychiatry 1989;46: Greenblatt DJ, Harmatz JS, Shapiro L, Engelhardt N, Gouthro TT, Shader RI. Sensitivity to triazolam in the elderly. New Eng J Med 1991;324: Carruthers SG, Shoeman DW, Hignite CE, AzamoffDL. Correlation between plasma diphenhydramine level and sedative and antihistamine effects. Clin Pharmacol Ther 1978;23: Berlinger WG, Goldberg MJ, Spector R, Ciang C, Ghoneim MM. Diphenhydramine: kinetics and psychomotor effects in elderly women. Clin Pharmacol Ther 1982;32: Cohen AF, Posner J, Ashby L, Smith R, Peck AW. A comparison of methods for assessing the sedative effects of diphenhydramine on skills related to car driving. Eur J Clin Pharmacol 1984;27: Levine B, Roehrs T, Zorick F, Roth T. Actigraphic monitoring and polygraphic recording in determination of sleep and wake. Sleep Res 1986;15: Carskadon MA, Dement WC, MitIer M, Roth T, Westbrook P, Keenan S. Guidelines for the Multiple Sleep Latency Test (MSLT): a standard measure of sleepiness. Sleep 1986;9: Rechtschaffen A, Kales A. A manual of standardized terminology, techniques, and scoring system for sleep stages of human subjects. Los Angeles: BIS/BRI, UCLA, B1yen GT, Greenblatt DJ, Scavone JM, Shader RI. Pharmacokinetics of diphenhydramine and a demethylated metabolite following intravenous and oral administration. J Clin Pharmacol 1986;26: Greenblatt DJ, Divoll M, Moschitto U, Shader RI. Electroncapture gas chromatographic analysis of the triazolobenzodiazepinesalprazolam and triazolam. JChromat 1981;225: Roth T, Roehrs T, Koshorek G, Sicklesteel J, Zorick F. Sedative effects of antihistamines. J Allerg Clin Immun 1987;8: Roth T, Roehrs T, Zorick F. Sleepiness: its measurement and determinants. Sleep 1982;5:S Sleep, Vol. 16, No.4, 1993