| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Evidence has accumulated that the N-methyl-D-aspartate receptor system plays a role in continuous and particularly, in stimulus-evoked pain after nerve injury. We examined, in a randomized, double-blinded, cross-over fashion, the analgesic effect of memantine (a N-methyl-D-aspartate receptor antagonist) in a group of patients with chronic pain after surgery. We randomized 19 patients to receive either memantine or placebo in the first 5-wk treatment period. A washout period of 4 wks was followed by another 5-wk treatment period with the opposite drug. The dosage of drug was increased from 5 to 20 mg/d. Pain was recorded daily, with the use of a 010 numeric rating scale. Before and at the end of each treatment period, pain and sensitivity were also assessed by using the McGill Pain Questionnaire, allodynia to touch, brush and cold, wind-up-like pain, and thresholds to mechanical stimuli (pressure and von Frey hair). A total of 15 patients (12 amputees and three patients with other nerve injuries) completed the study. There was no difference between memantine and placebo on any of the outcome measures. We conclude that memantine at a dosage of 20 mg/d does not reduce spontaneous or evoked pain in patients with nerve injury pain.
Implications: In a randomized, double-blinded and cross-over study, the analgesic effect of memantine (a drug which reduces the excitability of sensitized neurons in the dorsal horn) was examined in 19 patients with chronic pain after nerve injury.
Surgical procedures, such as thoracotomy, mastectomy, and limb amputation, are often associated with long-term spontaneous pain and abnormal, evoked pains, such as allodynia, hyperalgesia, and wind-up-like pain. It has been reported that 70% to 85% of amputees suffer from phantom pain and stump pain persists in 5% to 10% of patients (1,2). Clinical examination of stumps have revealed abnormal sensibility in at least 50% of subjects (3). It is now clear that after nerve transsection, spontaneous activity develops at the site of injury and in dorsal root ganglion cells. These changes sweep more centrally, followed by central sensitization, caused by increased afferent barrage from the periphery. The manifestations of such sensitization are manyfold, including a reduction of pain thresholds, an exaggerated response to nonnoxious (allodynia) and noxious (hyperalgesia) stimuli, and occasionally abnormal referrals (phantom sensations) (4). Evidence has accumulated that N-methyl-D-aspartate (NMDA) receptors are involved in central sensitization. Antagonists at the NMDA receptor site, such as MK 801, reduce the excitability of sensitized neurons in the dorsal horn and block the facilitated neuronal response to fast repetitive C-fiber input (5). In animal models of neuropathic pain, NMDA receptor antagonists attenuate spontaneous pain-related behaviors and responses to evoked pains (6). The previously mentioned findings prompted the use of NMDA receptor antagonists in humans with neuropathic pain. Kristensen et al. (7) used 3-(2-carboxypiperazin-4-yl) propyl-1-phosphonic acid (CPP), a NMDA receptor antagonist, in a patient with severe neuropathic pain after surgery for varicose veins. The intrathecal administration of CPP abolished wind-up-like phenomena and the spread of pain to the left half of the body. Since then, several studies have examined the effect of NMDA receptor antagonists on experimental pain (8,9) and chronic neuropathic pain (1016). All experimental and most of the clinical studies show that both spontaneous pain and abnormal evoked pain are reduced by the administration of ketamine or dextromethorphan, anesthetic, and antitussive drugs, respectively, and both clinically available NMDA receptor antagonists. However, ketamine is only available for injection, although single cases of oral administration have been reported (1718). The clinical use of both drugs is limited by the frequency of mainly psychomimetic side effects, so there is a need for other NMDA receptor antagonists. Memantine, a derivative of amantadine, is available for long-term use in humans, and has been used since 1978 in the treatment of Parkinsons disease, dementia, and spasticity. Memantine blocks NMDA receptor sites in a noncompetitive fashion (19). Experimental studies have shown that memantine has antinociceptive properties in the secondary noxious phase of inflammation induced by the formalin test (20), and prevents development of hyperalgesia in the arthritic rat model induced by carrageenan (21). In the Chung mononeuropathy pain model, it has been shown that memantine reduces thermal and mechanical hyperalgesia (22). At present, clinical experiences are few. In a clinical double-blinded, placebo-controlled trial Eisenberg et al. (16) failed to find any effect of memantine in patients with postherpetic neuralgia. Based on the clear analgesic effect of ketamine in patients with post- amputation pain (14) and in patients with stimulus-evoked, traumatic nerve injury pain (13), we decided to perform a randomized double-blinded, placebo-controlled, cross-over study of the effect of memantine on spontaneous and evoked pain in patients with nerve injury pain.
Patients with neuropathic pain after amputation or surgery were recruited from the pain clinic at Aarhus University Hospital between January 1997 and March 1999. Inclusion criteria were age 18, insufficient effect of conventional analgesic treatment, and continuous pain intensity 3 on a numerical rating scale (NRS) from 0 to 10 (0 = no pain, 10 = unbearable pain). Exclusion criteria were kidney or liver disease, epilepsy, Parkinsons disease (memantine can enhance the effect of dopaminergic treatment), agitation, psychosis, pregnancy, or lactation. Women of fertile age were only included if they used oral anticontraception. All patients were given oral and written information about the study and the possible side effects of memantine. Informed, written consent was obtained from each participant. The study was approved by the local ethics committee and the Danish National Board of Health. Guidelines for good clinical practice were followed. Our study was a double-blinded, randomized, placebo (vehicle)-controlled, two-period cross-over trial. Patients were seen in the pain clinic before and at the end of each treatment period, a total of four times for each patient. Patients were randomized to receive either memantine or placebo in the first 5-wk treatment period. A washout period of 4 wks was followed by another 5-wk treatment period with the opposite drug. Dosage of drug was increased from 5 to 20 mg/d during the treatment period as follows: first week, 5 mg in the morning; second week, 10 mg in the morning; third week, 10 mg in the morning and 5 mg in the afternoon; fourth and fifth weeks, 10 mg in the morning and 10 mg in the afternoon. The dose of drug (20 mg) was chosen based on other studies in which memantine produced clinical effects in patients with dementia (23,24). Before the study began, three patients were treated with memantine in an unblinded fashion and two experienced the effect of memantine or side effects of memantine at dosages between 10 and 15 mg/d. The study drugs (memantine and placebo tablets) were randomized and prepared by the hospital pharmacy in identical containers, marked with the name of the project, treatment period (first or second), study week (first-fifth), number of tablets to be taken each day (one or two), and consecutive patient numbers. Vehicle tablets were identical except for the presence of memantine. At the end of each treatment period, blood samples were taken for analysis of Se-memantine. If it was not possible to discontinue concurrent analgesic medication, patients were required to remain on the same daily dosage throughout the study period. During the entire study period, patients were allowed to use rescue doses of paracetamol in a maximum dosage of 4 g/d. Analgesic medication was recorded daily. At the end of each treatment week, patients recorded the experience of any of the following side effects: dizziness, headache, nausea, tiredness, inner restlessness, motor restlessness, and insomnia. Safety evaluations were conducted at each of the four visits in the pain clinic and included assessment of adverse effects, compliance with drug administration, and use of concurrent medication. Routine blood tests (blood count, and liver and kidney functions) were taken at the patients study inclusion and at the end of each treatment period. Intensity of pain (mean intensity during the last 24 h) was recorded every day in the evening during the two treatment periods on a NRS from 0 to 10. Amputees recorded both stump and phantom pain. At the end of each treatment week, patients were asked to record pain relief on a NRS from 0 to 10 (0 = no pain relief, 10 = complete pain relief). The McGill Pain Questionnaire was completed before and at the end of each treatment period. Before and at the end of each treatment period, the following were measured at the affected site and at the contralateral mirror image site: allodynia to touch, cold and brush, wind-up-like pain, and thresholds to mechanical stimuli. The examined area was marked and drawn to a transparent paper to make sure that the same area was examined at all visits. Each patient was examined by the same investigator. Allodynia to touch was determined by gently stroking the affected skin with a cotton swab which was moved from outside the painful area toward the center. The patient was asked to report whether the sensation of touch changed to a sensation of pain. If allodynia was present, the area was traced on a transparent paper if small enough, otherwise the area was approximated and transferred to a standard homunculus. Allodynia to cold was examined by using a thermal roll (made of steel with a steel handle) cooled in a box filled with ice. The thermal roll was applied to the skin area and the evoked pain was graded on a NRS from 0 to 10 (0 = no pain, 10 = unbearable pain). Allodynia to brush was examined by using an electric toothbrush applied to the skin area. Stimulation was discontinued after 60 s or before, if the pain became unbearable. Evoked pain was graded on a NRS from 0 to 10. Wind-up-like pain was examined by repeatedly tapping the affected skin with a stiff von Frey hair (447 g) at a rate of 3 taps/s by using a computer-controlled solenoid (Laboratory for Experimental Pain Research, Aaborg University, Denmark). The von Frey hair did not bend during stimulation. Stimulation was discontinued after 60 s or earlier, if it became unbearable. Evoked pain was graded on a NRS from 0 to 10. Touch detection threshold (TDT) and pain detection threshold (PDT) were determined by using von Frey hairs consisting of 20 monofilaments of increasing diameter calibrated to deliver a specific force on the skin (0.004447 g monofilaments, Semmes-Weinstein, Stoelting, IL). The filaments were applied in an ascending and descending order of magnitude. The force required to bend the filament at the TDT and PDT was recorded. The pressure tolerance threshold was determined by using a hand-held pressure algometer. A circular probe with an area of 0.25 cm2 was used. The pressure application rate was 20 kPka/s. When the pressure tolerance threshold was reached, the patient pressed a push-button that immediately froze the digital display that could be read by the examiner. Each value was determined three times and the average of the three measurements was defined as threshold. Statistical analysis of this cross-over trial focused on estimation of the difference in effect of treatment with memantine/placebo. For estimates of effect, 95% CI are shown and data are presented as mean ± SD or median. A P <0.05 was considered statistically significant. For data on intensity of pain, an average of seven daily pain scores were calculated for each patient, given one value per week. The estimate of effect was calculated as the difference between NRS value during treatment with memantine and NRS value during treatment with placebo. Scores for each week were analyzed both separately and combined, estimating the effect of increasing the dose of memantine/placebo. Also, an average effect based on all five treatment weeks was estimated. For data concerning the McGill Pain Questionnaire, wind-up-like pain, and allodynia to cold and brush, the effect was calculated as the difference between values obtained before and after each treatment period with memantine/placebo. Because data concerning TDT, PDT, and PPT were highly skewed, the effect was calculated as the ratio between the two values (before and after treatment with memantine/placebo) and data were then analyzed on a log scale. The results were transferred back to the original scale, implying that difference in effect is given as the ratio between effect of treatment during memantine/placebo.
We included 19 patients in the study who had pain after amputation (n = 15) or surgery (n = 4) (mean age = 51 yr ± 14; men = 11; women = 8; mean pain intensity at inclusion = 5.4 ± 1.8 [NRS 010]). Patient characteristics are shown in Table 1. Four patients dropped out or were withdrawn during the study period. Among the 15 patients who completed both treatment periods, nine received memantine and six received placebo as the first drug.
In all patients, the most intense pain was chosen for analysis. In amputees with both stump and phantom pain, stump was most intense in six patients, and phantom pain was most intense in six patients. Figure 1 shows the estimated effect of memantine compared with placebo and the estimated dose-response relationship. There was no effect of memantine on intensity of pain. Four patients reported pain relief graded as 1, 2, 4, and 5, respectively, on a NRS from 0 to 10, during the fifth treatment week with memantine (max dosage 20 mg/d) and two patients reported pain relief graded as 2 and 5 on a NRS during the fifth treatment week with placebo. One patient reported pain relief during both treatments. Rescue doses of paracetamol were used by six patients. Total consumption was 110 g during treatment with memantine and 118 g during treatment with placebo.
Table 2 contains the estimated effects of treatment with memantine compared with placebo on spontaneous and evoked pains. As can be seen, there was no statistically significant effect on any of the responses except for PDT on the unaffected side where the effect of memantine was only a factor 0.32 (0.110.94) of placebo. Allodynia to touch was found in eight patients, mean area (cm2) was 164 ± 104 before and 166 ± 106 after treatment with memantine and 158 ± 119 before and 166 ± 120 after treatment with placebo.
The number of reported side effects was similar during all treatment weeks, so data are shown for the fifth treatment week (max dosage, 20 mg/d) (Table 3). Five patients reported 10 side effects during the fifth treatment week with memantine and four patients reported 11 side effects during the fifth treatment week with placebo. Only one patient graded a side effect as severe during treatment with memantine (nausea). Another patient reported three side effects as severe during treatment with placebo (headache, tiredness, insomnia). One patient dropped out in the first treatment period (memantine). She never experienced an effect or side effects of treatment. Three patients were withdrawn from the study. One patient complained of a severe headache after 3 days of treatment (first treatment period, placebo), another patient got a skin rash after a few days of treatment (first treatment period, placebo), and a third patient developed a drop-foot which resolved after 3 mo (first week, second treatment period, memantine). A patient broke her leg in the last week of the second treatment period, so data from her planned last visit at the pain clinic are missing. The last two events were registered as serious adverse events and reported to the authorities.
Routine blood tests remained unchanged throughout the study period. Memantine was present in all blood samples from the memantine treatment period. The mean concentration of memantine was 80 ± 35 ng/mL. The largest measured concentration (159 ng/mL) was not associated with an effect or side effects of treatment.
This study of patients with traumatic nerve injury pain (15 amputees and 4 with other types of nerve injury) failed to demonstrate any benefit of memantine in a maximum dosage of 20 mg/d. The study is thus, in line with a double-blinded parallel study by Eisenberg et al. (16) in which 24 patients with postherpetic neuralgia (PHN) were randomized to receive either memantine (10 mg/d increasing to 20 mg/d) or placebo for 5 weeks. Neuropathic pains are irrespective of underlying causes or anatomical location and are characterized by a series of symptoms and signs which include spontaneous and evoked types of pain to different stimuli. We measured continuous pain, pain tolerance to pressure, touch and pain detection thresholds to pinprick, pain evoked by brush, cold, and repetitive von Frey hair stimulation. This approach allowed us to determine whether memantine had an effect on any of the constituent symptoms of neuropathic pain. The failure to show analgesic or other benefits in the studied patient population may have several explanations, such as a small number of patients, small dose of drug or plasma concentrations, or lack of effect at presumed receptor sites. Only 15 patients completed our study. However, the size of the confidence limits indicates that the failure to find an effect is unlikely to be caused by a Type 2 error. The lack of effect at a dosage of 20 mg/d does not exclude that the drug might be effective at a larger dose. After completion of our study, four patients were offered treatment with memantine at dosages up to 40 mg/d. None of the patients had a clear analgesic effect of treatment. Also, in a recent continuing trial, 24 patients with PHN or painful diabetic neuropathy were treated with memantine up to 55 mg/d (in PHN, the mean dosage was 42 mg/d and in diabetic neuropathy, the mean dosage was 29 mg/d) without an effect on pain (MB Max, written communication, February 2000). Our analysis of serum memantine showed values in a range similar to that found in pharmacokinetic studies after memantine at a dose of 20 mg (Merz, Frankfurt, Germany, investigators brochure, July 11, 1995), so it is unlikely that an insufficient dose or reduced bioavailability can account for the lack of effect. The lack of effect may also be caused by the slow dose escalation we used. The reason for this gradual increase was an attempt to avoid side effects; however, side effects were few, tolerable, and reported with the same frequency during treatment with memantine and placebo. Alternatively, memantine may be less likely to respond to manifestations of central sensitization compared with ketamine and dextromethorphan. We consider this possibility unlikely, because memantine, ketamine, and dextromethorphan all have a noncompetitive action at NMDA receptor sites. Future studies may document an analgesic effect of memantine at larger doses or in other types of neuropathic pain. The observation that central sensitization induced by nerve injury depends not only on NMDA receptor activity, but also on a variety of other factors (neurokinins, prostaglandins, adenosine, and opioids) suggests that a combination of different drug classes may be a useful approach. Stubhaug et al (25) showed combined ketamine and morphine had a better effect than each drug alone. Thus, it is possible that memantine may be effective combined with other drugs that also act at sodium channels or at opioid receptor sites.
Supported, in part, by Grant 120828-1 from the Danish Medical Research Council, Grant 78000 Danish Cancer Society, and the Danish Pain Research Center. We thank H. Hartvig, Department of Biostatistics, University of Aarhus, Aarhus, Denmark, for statistical advice.
Merz, Germany, kindly delivered memantine tablets and analyzed Se-memantine.
This article has been cited by other articles:
|
||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|