Tramadol (Tramal®, Grünenthal) is a centrally acting analgesic drug. It is a racemic mixture of two enantiomers, (+)tramadol and (–)tramadol and it has been found in vivo and in vitro to have a low but preferential activity at mu opioid receptors, and it also inhibits both norepinephrine and 5-hydroxytryptamine neuronal re-uptake [1,2]. Tramadol is used in the treatment of acute and chronic pain. Compared with morphine it is thought to cause less tolerance, addiction and respiratory depression [3,4].
Patients with pain often suffer from sleep disturbances . It may be difficult in these patients to differentiate between sleep disturbance caused by pain, by the underlying disease, or by the analgesic treatment. There are some data on the effects of opioids on sleep structure from experiments in animals and studies in patients undergoing surgery (i.e. acute pain setting). For instance, in cats, microinjection of morphine (an opioid with a high affinity to mu receptors) into the medial pontine reticular formation, a structure playing a key role in regulating paradoxical sleep, significantly inhibited paradoxical sleep and increased the number of apnoeic episodes . These authors postulated that morphine via the mu opioid receptor may inhibit the release of acetylcholine in the reticular formation and that this may inhibit paradoxical sleep [7,8]. Uncontrolled observations in patients undergoing surgery receiving morphine to treat acute postoperative pain indicated that there may be a rebound in paradoxical sleep some days after the end of morphine administration . It has also been suggested that sleep could be disturbed by inhibiting norepinephrine and 5-hydroxytryptamine neuronal re-uptake [10,11]. Norepinephrine and 5-hydroxytryptamine neurons play an important role in inhibiting paradoxical sleep, as was found with antidepressant drugs [12,13].
The effect of tramadol on sleep has never been investigated. The aim of this study was to establish a model to investigate the relative impact of analgesics on sleep structure in man, using standardized and validated methods of polysomnography, and to investigate the effects of tramadol on sleep structure in healthy volunteers without pain. We addressed the following questions. First, what are the short-term effects of tramadol on sleep structure (i.e. is the structure of sleep modified in the night of tramadol application)? Second, what are the medium-term effects of tramadol on sleep structure (i.e. is the structure of sleep modified in the night after tramadol application)?
Materials and methods
The protocol was approved by the local ethical committee and written informed consent was obtained from the subjects. Healthy volunteers, both sexes, aged 20–40 years, all students or nurses of our hospital, were invited to take part in this paid experiment. Candidates were screened by medical evaluation and laboratory testing. We applied the following exclusion criteria: regular medication, allergy to drugs, porphyria, toxicomania, alcoholism, epilepsy, myasthenia gravis or other neurological disorders, sleep disturbances, psychiatric disorders, arterial hypertension, dysrhythmia or other cardiovascular diseases, chronic obstructive pulmonary disease or other pulmonary diseases, hepatic disorders, renal dysfunction, intestinal disorders, status after transplantation, diabetes or other endocrinological problems, pregnancy or potential pregnancy.
The subjects were requested to refrain from napping and extreme or exceptional physical exercise, from drinking alcohol, from using any medication, and to keep their caffeine, tea, cola and tobacco consumption within the habitual range during the entire study period.
The volunteers were studied in two equally structured sessions with a minimum interval of two weeks between the sessions. One session consisted of four nights. During the first night (adaptation night), volunteers did not receive any drug and there was no monitoring. During the second night (predrug placebo-night), volunteers received, single-blinded, two placebo tablets. During the third night (drug-night), volunteers received in a randomized, double-blind, cross-over fashion two 50 mg tramadol tablets or one 50 mg tramadol tablet plus one placebo. During the fourth night (postdrug placebo-night), volunteers received again, single-blinded, two placebo tablets. Polysomnographic monitoring was conducted during nights two to four. Randomization was Carried out with a computed random generator program. Tramadol 50 mg tablets and placebo tablets were of identical shape, colour and taste. Thus, the trial was truly double-blind during the drug-night. Volunteers were unaware of the nature of the tablets at any time. Each night, the volunteers swallowed the tablets at 22:00 hours under supervision, with a small amount of water. Each morning, the volunteers filled out a standard adverse effect checklist.
During recordings, subjects stayed alone in a quiet dark room, lying in a comfortable bed. Sleep structure was recorded using standardized validated polysomnographic methods . Recordings were carried out with four electroencephalographic derivations (F4-Cz, T3-Cz, T4-Pz, O2-Pz) to monitor cerebral activity, one right electrooculogram derivation for eye movement measurement, one submental electromyogram derivation for muscle activity monitoring, a one-lead electrocardiographic derivation to monitor heart activity, and a nasal ventilatory recording for respiratory surveillance. All raw signals were recorded and stored on the hard disk of the sleep laboratory computer. Automatic sleep analysis (Medilog SAC 847, Oxford, UK) with a 30-s epoch was performed between 22:00 and 07:00 hours. Automatic analyses were cross-checked by the investigators and by qualified technicians of our sleep laboratory. The analysis allowed identification of sleep structure according to the sleep parameters described by Rechtshaffen and Kales . A similar model to ours has been used to evaluate the effect of antidepressants on sleep structure [12,13].
Pre-drug, drug-, and postdrug sleep structure (described with the predefined sleep parameters) were compared using a Kruskall–Wallis ANOVA non-parametric test. Paired comparisons of sleep parameters (for each dose vs. placebo) were assessed with a Wilcoxon test, if the Kruskall–Wallis ANOVA non-parametric test was statistically significant for a sleep parameter. Because we assumed that there was a physiological dependency between most sleep parameters (for instance, an increase in the duration of sleep stage 2 is always associated with a compensatory decrease in the duration of sleep stage 4 ), we did not correct for multiple testing. A P-value < 0.05 was considered significant.
Twelve volunteers were screened. Four were excluded: one male volunteer had elevated liver enzymes (suspicion of Gilbert's disease), two female volunteers had sleep disturbances during the predrug placebo night, and one female volunteer had anorexia nervosa. The remaining eight volunteers, four men and four women, had a mean age of 24 (SD ±2) years and a mean weight of 64 (SD ±12) kg.
There was no difference between the predrug placebo-nights of the first and second session. All volunteers showed a normal sleep structure with placebo on both occasions. Results were similar to previously reported data in healthy subjects receiving placebo under identical conditions .
Drug-nights and postdrug placebo-nights
Sleep-wake balance, sleep discontinuity and sleep cycles
Compared with placebo, neither dose of tramadol had any statistically significant effect on sleep-wake balance, sleep discontinuity or sleep cycles (Table 1).
Orthodox and paradoxical sleep
No statistically significant time alterations with or without the drug were observed in stage 1 sleep and in stage 3 sleep (Figure 1a and c respectively).
Drug-related changes were observed in stage 2 sleep, in stage 4 sleep, and in paradoxical sleep. Both tramadol doses significantly increased duration of stage 2 sleep compared with both predrug and postdrug placebo-nights (Figure 1b). Duration of stage 2 sleep in the postdrug placebo-night after tramadol 100 mg, but not after 50 mg, was significantly shorter than in the predrug placebo-night.
Both tramadol doses significantly decreased duration of stage 4 sleep compared with both the predrug and postdrug placebo-night (Figure 1d). Duration of stage 4 sleep in the postdrug placebo-night after tramadol 100 mg, but not after 50 mg, was significantly longer than in the predrug night (Figure 1d).
Tramadol 100 mg, but not 50 mg, significantly decreased the duration of paradoxical sleep compared with both predrug and postdrug placebo-nights (Figure 1e). Duration of paradoxical sleep during the postdrug placebo-night was not different from the predrug placebo-night.
Adverse drug reactions
No adverse drug reactions were reported. The modified electroencephalogram monitoring did not detect any signs of epileptic potentials. The nasal ventilatory recordings showed no decreased respiratory rate.
What are the main results of this study?
The analgesic tramadol has been on the market for more than 20 years but has never been, to our knowledge, the subject of a polysomnographic investigation. A single therapeutic dose of tramadol was found to have some effects on sleep stages in healthy volunteers. In the drug-night, both tramadol 50 and 100 mg decreased duration of stage 4 sleep and prolonged duration of stage 2 sleep. This may be interpreted as a compensatory mechanism. In the postdrug placebo night, tramadol 100 mg prolonged duration of stage 4 sleep and shortened duration of stage 2 sleep. This again may be interpreted as a compensatory mechanism. The effect of a single therapeutic dose of tramadol 100 mg on sleep in the postdrug placebo night is remarkable because tramadol and its main metabolite mono-o-desmethyltramadol have relatively short elimination half-times of 6 h and 7.4 h respectively . This most probably represents a rebound phenomenon.
During the drug-night, tramadol 100 mg shortened duration of paradoxical sleep. Rebound of paradoxical sleep did not happen. Thus, tramadol 100 mg seemed to have an effect on paradoxical sleep in the drug-night, but not in the postdrug placebo night.
The supplementary effects of the higher dose of tramadol on paradoxical sleep and on stage 4 during the postdrug night suggested a dose-responsiveness. However, the claim of a dose response may be premature with only two doses tested. We did not test higher doses although they may be used in clinical practice (for instance, 150 mg). Thus, we do not know if higher doses of tramadol would have an even stronger impact on sleep structure.
Our experiment was conducted in healthy volunteers, using standardized and validated methods of polysomnography. Two different therapeutic single doses of tramadol were compared in a randomized, double-blind, cross-over design. Sleep structure before, during and after drug application was recorded. Thus, the model enabled us to evaluate both short-term and medium-term effects of tramadol on sleep structure, and to test dose-responsiveness.
Dose-responsiveness is regarded as a possibility for testing the validity of an experimental model. For instance, single doses of tramadol 50 mg and 100 mg have been shown to be significantly more efficacious than placebo, in relieving acute postoperative pain . In our study, compared with placebo, statistically significant changes in stage 2 and 4 were observed after single doses of tramadol 50 mg and 100 mg; thus, our model which describes sleep structure electro-physiologically showed evidence of dose-responsiveness. There was some further evidence of dose-responsiveness, in that tramadol 100 mg but not 50 mg had an effect on paradoxical sleep and on orthodox sleep in the postdrug night. Therefore, the findings of this study were consistent with the data from the acute pain setting . In addition, sleep structure with placebo in this study was similar to historical controls (i.e. healthy volunteers receiving placebo under identical conditions) .
Limitations of the study
There were three main limitations. First, the study was performed in healthy volunteers without pain. Only one dose of tramadol was administered during one night, and no concomitant medication was given. Thus, our study population did not represent typical pain patients. The aim of our study, however, was to establish a model which allowed estimation of drug-related effects on physiological sleep, without confounding factors such as concomitant drugs or sleep disturbance due to the underlying disease or pain.
Second, we measured electro-physiological parameters of sleep only. The clinical relevance of those, however, needs to be clarified. Sleep structure has been studied for more than three decades . Most interest has focused on the clinical relevance of disturbances in sleep–wake balance, deep orthodox sleep, and paradoxical sleep. It has been shown, for instance, that chronic reduction or fragmentation of total sleep time can produce mood alterations . It was suggested that stage 3 and 4 sleep could renew energy stock induced by brain and body cooling . In animals, deprivation of paradoxical sleep was shown to lead to neural excitability . In humans, however, shortening of paradoxical sleep per se does not seem to be harmful . Indeed, there is evidence that deprivation of paradoxical sleep is one of the underlying mechanisms of the therapeutic effect of antidepressants . Increased duration of paradoxical sleep, on the other hand, may be associated with increased sympathetic activation, and, possibly, with an increased risk of subsequent ischaemic events . This may happen as a rebound after deprived paradoxical sleep . These data suggest that changes in sleep structure may have both beneficial and harmful effects. This underlines the need for more insight into the effect of drugs on both electro-physiological parameters of sleep and on long-term outcome.
Finally, our single-dose study did not allow us to investigate true long-term effects of tramadol. Blockade of amine uptake leads to a series of events some of which are probably to be observed only after days or weeks of treatment. Thus, to investigate true long-term effects on sleep pattern, tramadol would have to be given for several days or weeks. This also applies to the possibility of delayed withdrawal phenomena. Investigations of long-term effects are of interest because chronic pain patients are likely to take tramadol over a prolonged time period.
Comparisons with other relevant drugs in pain treatment
Tramadol has pharmacological effects on several central transmitter systems. Therefore, comparisons with other mu-receptor agonists like morphine, or drugs which modulate norepinephrine and serotonin re-uptake like antidepressants are of interest. Data on the effects of other drugs which are frequently used in pain treatment on sleep structure are rare. The only available clinical data on morphine are from an abstract reporting on a non-randomized, non-blinded study in six volunteers . The volunteers received a placebo during the second night and morphine during the third night. The effects were dose dependent, morphine was given in two doses (0.1 mg kg−1 and 0.2 mg kg−1 intramuscularly). Morphine 0.1 mg kg−1 exhibited only little effects while morphine 0.2 mg kg−1 decreased the total sleeping time, the duration of stage 3 and 4 sleep was decreased by half, and the duration of paradoxical sleep was decreased by a factor of four. No medium-term effects were reported. Comparisons of these data with our results on tramadol are difficult. Intramuscular injections, for instance, as used in the morphine study, are likely to disturb sleep structure.
The effects of tricyclics and serotonin re-uptake inhibitors on sleep structure were investigated in healthy volunteers using similar methods of polysomnography as in our study [12,13]. A typical effect of these antidepressants is inhibition of paradoxical sleep [12,13]. Rebound of paradoxical sleep was not reported with paroxetine 20, 30 and 40 mg , and amitriptyline 75 mg  but with both antidepressants duration of paradoxical sleep did not return to normal in the night after drug application. Thus, tramadol 100 mg, paroxetine 20, 30 and 40 mg, and amitriptyline 75 mg seem to have similar shortening effects on paradoxical sleep during the night of drug application, and none of these drugs induce a rebound of paradoxical sleep in the subsequent night. However, unlike with tramadol in our study, deprivation of paradoxical sleep with paroxetine or with amitriptyline was maintained even in the postdrug night [12,13]. This prolonged effect on paradoxical sleep with the antidepressants compared with tramadol may be related to the longer elimination half-time of about 16 h with both paroxetine and amitriptyline. Contrary to tramadol, both paroxetine and amitriptyline did not reduce duration of stage 4 sleep during the drug-night [12,13]. Accordingly, a rebound of stage 4 sleep in the postdrug night, as seen with tramadol, was not observed with the antidepressants.
Conclusion and research agenda
In healthy volunteers, a single dose of tramadol 50 mg disturbs sleep in the night of drug application. With 100 mg, sleep is disturbed in both the night of drug application and in the subsequent night. Analgesic-induced changes of sleep structure may have implications for the management of patients with acute and chronic pain. More relevant and valid data are needed on the effects of drugs that are currently used in pain treatment on sleep structure. Our model may be used to achieve this.
We thank Iris Henzi, MD, for technical assistance.
1 Desmeules JA, Piguet V, Collart L, Dayer P. Contribution of monoaminergic modulation to the analgesic effect of tramadol. Br J Clin Pharmacol
1996; 41: 7–12.
2 Raffa RB, Friderichs E, Reimann W et al. Opioid
and nonopioid components independently contribute to the mechanism of action of tramadol, an atypical opioid
analgesic J Pharmaco Exp Ther
1992; 260: 274–284.
3 Houmes RJM, Voets MA, Verkaaik A et al.
Efficacy and safety of tramadol versus morphine for moderate and severe postoperative pain
with special regard to respiratory depression Anesth Analg
1992; 74: 1–5.
4 Preston KL, Jasinski DR, Testa M. Abuse potential and pharmacological comparison of tramadol and morphine. Drug Alcohol Depen
1991; 27: 7–17.
5 Lehtinen I, Kirjavainem T, Hurme M et al.
Sleep-related disorders in carpal tunnel syndrome. Acta Neurol Scand
1996; 93: 360–365.
6 Keifer JC, Baghdoyan HA, Lydic R. Sleep disruption and increased apneas after pontine microinjection of morphine. Anesthesiology
1992; 77: 973–982.
7 Lydic R, Keifer JC, Baghdoyan HA, Becker L. Microdialysis of the pontine reticular formation reveals inhibition of acetylcholine release by morphine. Anesthesiology
1993; 79: 1003–1012.
8 Cronin A, Keifer JC, Baghdoyan HA, Lydic R. Opioid
inhibition of rapid eye movement sleep by a specific mu receptor agonist. Br J Anaesth
1995; 74: 188–192.
9 Knill RL, Moote CA, Skinner MI, Rose EA. Anesthesia with abdominal surgery leads to intense REM sleep during the first postoperative week. Anesthesiology
1990; 73: 52–61.
10 Jones BE. Paradoxical sleep and ist chemical/structural substrates in the brain. Neuroscience
1991; 40: 637–656.
11 Vogel GW. Evidence for REM sleep deprivation as the mechanism of action of antidepressant drugs. Prog Neuro-Psychoph
1983; 7: 343–349.
12 Riemann D, Velthaus S, Laubenthal S et al.
REM-suppressing effects of amitriptyline and amitriptyline-N-oxide after acute medication in healthy volunteers: results of two uncontrolled pilot trials. Pharmacopsychiatry
1990; 23: 253–258.
13 Saletu B, Frey R, Krupka M et al.
Sleep laboratory studies on the single-dose effects of sertonin reuptake inhibitors paroxetine and fluoxetine on human sleep and awakening qualities. Sleep
1991; 14: 439–447.
14 Blois R, Gaillard JM, Attali P, Coquelin JP. Effect of zolpidem on sleep in healthy subjects: a placebo-controlled trial with polysomnographic recordings. Clin Ther
1993; 15: 797–809.
15 Rechtschaffen A, Kales A. A Manual of Standardized Terminology, Techniques and Scoring Sytem for Sleep Stages of Human Subjects
. Los Angeles: UCLA Brain Information Service, Brain Research Institute, 1968.
16 Merica H, Gaillard JM. Statistical description and evaluation of the interrelationships of standard sleep variables for normal subjects. Sleep
1985; 8: 261–273.
17 Gibson TP. Pharmacokinetics, efficacy, and safety of analgesia
with a focus on tramadol HCl. Am J Med
1996; 101: 47S–53S.
18 Moore RA, McQuay HJ. Single-patient data meta-analysis of 3453 postoperative patients: oral tramadol versus placebo, codeine and combination analgesics. Pain
1997; 69: 287–294.
19 Gaillard JM. Eléments de physiologie. In: Gaillard JM, ed. Le Sommeil Ses Mécanismes et Ses Troubles
. Lausanne: Payot E, 1990: 65–92.
20 McGinty D, Szymusiak R. Keeping cool: a hypothesis about the mechanisms and functions of slow-wave sleep. Trends Neurosci
1990; 13: 480–487.
21 Vogel GW. A review of REM sleep deprivation. Arch Gen Psychiat
1975; 32: 749–761.
22 Somers VK, Dyken ME, Mark AL, Abbound FM. Sympathic-nerve activity during sleep in normal subjects. New Engl J Med
1993; 328: 303–307.
23 Jouvet M. Paradoxical sleep mechanisms. Sleep
1994; 17: S77–S83.
24 Moote CA, Knill RL, Skinner MI, Rose EA. Morphine disrupts nocturnal sleep in a dose-dependent fashion. Anesth Analg
1989; 68: S200.