Doi:10.1016/j.pmr.2005.12.006

aDepartment of Psychiatry, University of Washington, 1959 Pacific Street, bDepartment of Rehabilitation Medicine, University of Washington, 1959 Pacific Street, The evolution of contemporary clinical pain care dates back to the pub- lication of Melzack and Wall’s gate control theory in the journal Science in1965 . Around that time, most of the attention was given to the ‘‘gate’’ inthis theory of pain, which referred to inhibitory interneurons within the sub-stantia gelatinosa of the dorsal horn of the spinal cord. The gate closedwhen stimulation of large diameter fibers by touch or vibration activatedthese inhibitory interneurons, decreasing central neurotransmission fromsmall diameter C fibers. At that time, little attention was paid to the roleof descending control from the brain. Melzack and Wall did note this intheir model, but little was known about descending pain modulation in1965. Most of the pain research in the 40 years since that time has focusedon the central control component of the gate control theory. Because centralcontrol is so important, antidepressants and anticonvulsants have a prom-inent role in the treatment of chronic pain problems.
Persons attending medical school in the 1970s and 1980s were still being taught that the neuroanatomy of pain essentially consisted of two ascendingsystems: the neospinothalamic and paleospinothalamic systems. It is nowclear that this is only half of pain neuroanatomy. Equally important arethe descending systems traveling from the central nervous system (CNS)to the periphery that continually modulate transmission in the ascendingsystems. The descending systems have opioid and nonopioid components.
Activity in the opioid system can be blocked through use of the opioid * Corresponding author.
E-mail address: (M.D. Sullivan).
1047-9651/06/$ - see front matter Ó 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.pmr.2005.12.006 antagonist naloxone. The nonopioid component of the system uses manydifferent neurotransmitters. Two of the important neurotransmitters activewithin the system are serotonin and norepinephrine. These substances arethe same neurotransmitters altered by commonly used antidepressants. Nor-epinephrine-containing cell bodies are found in the dorsolateral pontine teg-mentum. Serotonin-containing cell bodies are generally contained in therostroventral medulla . Currently, antidepressant medication is the pri-mary means of augmenting transmission in these systems.
The old thinking about pain bequeathed to us from Descartes taught that pain was primarily a function of peripheral tissue damage. Many studieshave documented that clinical pain is only loosely related to the amountof tissue damage in the periphery. The severity of clinical pain is best under-stood as a function of the overall threat to the organism rather than theamount of tissue damage in the part that hurts. The organism is continuallyadjusting the sensitivity of its pain system given its overall life situation. Insituations of great threat to survival, such as those explored by stress-induced analgesia experiments, the pain system is adjusted to low sensitivity.
In other situations of less life-threatening distress, the sensitivity of the sys-tem can be adjusted upward .
There are three basic reasons for considering antidepressant medications for patients with chronic pain. First, psychiatric disorders are common inpatients with severe or disabling chronic pain. Second, sleep disturbance iscommon, even in patients who do not meet criteria for psychiatric disorders.
Third, there is evidence that certain classes of antidepressants produce painrelief separate from relief of depression or other psychiatric disorders.
The DSM-IV defines major depression as 2 weeks or more of depressed mood or loss of pleasure. Patients with chronic pain may deny depressedmood and may attribute their lack of pleasure to the pain itself; however, ac-cording to the DSM-IV decision rules, these patients still qualify as meetingthe symptoms of anhedonia. In addition to these core symptoms, it is requiredthat a patient have four of the following symptoms to quality for the diagno-sis of major depression: weight loss or gain, insomnia or hypersomnia, psy-chomotor agitation or retardation, fatigue or loss of energy, worthlessnessor guilt, trouble concentrating or deciding, or thoughts of death or suicide.
Two percent to 4% of the US general population meet the criteria for majordepression at any time. Among ambulatory medical patients, this prevalenceis doubled to 5% to 9% of patients. Among medical inpatients, the rate isdoubled again to 15% to 20% who meet the criteria for major depressionat any point in time. It has been difficult to obtain stable prevalence estimates ANTIDEPRESSANT & ANTICONVULSANT MEDICATION of major depression from pain clinic populations, largely because these aresuch select groups. Prevalence rates have varied from 10% to 100%, withmost estimates being greater than 50%. Certainly, major depression is com-mon in patients with chronic pain and is often underdiagnosed owing to thestigma associated with psychiatric disorders and an often adversarial medi-cal/legal environment focused on responsibility for the initial injury.
One of the most frequently asked questions by clinicians about depression and chronic pain is whether the depression causes the pain or the pain causesthe depression. The appropriate answer to this question is yes. There are now13 studies of the prevalence of pain complaints in depressed patients. Therange of reported rates is between 15% and 100%, with a mean of 65% ofdepressed patients complaining of pain. Most of these studies were done inpsychiatric settings, but similar rates have been evident in primary care stud-ies. There have also been 41 studies of the prevalence of depression in patientswith chronic pain. Rates reported include 38% in pain clinics, 35% in psy-chiatry clinics, 52% in rheumatology clinics, 78% in dental clinics, 23% inobstetric-gynecology clinics, and 27% in primary care clinics These studies demonstrate that depression and pain strongly reinforce each other. Multiple studies have shown that patients with pain have a de-pression risk two to five times that of the general population This risk isespecially true for patients who have multiple pain complaints, multiple epi-sodes of chronic pain, or especially severe pain. Patients with pain and de-pression have more pain complaints, more pain intensity, and more painchronicity than those without depression .
Another psychiatric disorder with high prevalence in selected chronic pain populations is panic disorder. Panic disorder is defined as recurrent un-expected panic attacks in which at least one of the attacks has been followedby a month or more of persistent concern about additional attacks or worryabout the implications of the attack or a significant change in behavior relatedto the attacks. Panic attacks themselves are defined as discrete periods of in-tense fear or discomfort in which four or more symptoms develop abruptlyand reach a peak within 10 minutes. These symptoms include palpitations,sweating, trembling, shortness of breath, choking, chest pain, nausea, dizzi-ness, derealization, paresthesias, and hot flashes, as well as fears of losing con-trol, going crazy, or dying. In the general US population, 1% to 2% of adultsmeet the criteria for panic disorder at any given time. If patients presenting toan emergency room with chest pain are assessed for panic disorder, 16% to25% will meet these criteria. If one goes a bit ‘‘further down the career’’ ofa potential cardiac patient to patients who have had coronary angiographyand are found to have normal coronary vessels, nearly half of these patientshave panic disorder. Similarly, 28% of patients with irritable bowel syndromehave been found to have panic disorder, and 13% to 15% of patients withchronic headaches, particularly migraine headaches, have panic disorder A significantly underexplored problem among patients with chronic pain is posttraumatic stress disorder (PTSD). The prevalence of DSM-IV PTSD in the US general population is estimated to be approximately 8%, with 6%of men and 12% of women meeting the criteria. If one examines particularsubsets of patients with chronic pain, high rates of PTSD are evident .
Motor vehicle accident patients who have persistent pain show a 30% to50% rate of PTSD Injured workers referred for rehabilitative treatmenthave a 35% rate of PTSD . Fibromyalgia patients have been reported tohave a 20% current rate of PTSD and a 42% lifetime rate of PTSD As is true for pain and depression, many possible relationships are pos- sible between chronic pain and PTSD. PTSD can give rise to pain as sug-gested by the high rates of chronic pain in patients with a history ofsevere childhood maltreatment. The physical trauma that gave rise to thepain complaint also can give rise to PTSD. Some evidence suggests that dis-sociation during the injury increases the risk of subsequent PTSD. The painfrom the injury itself can produce PTSD, especially if the pain inducesdissociation.
Although it may seem implausible that traumatic events such as child- hood abuse could produce chronic pain problems many years after theabuse has ceased, data from animals and humans provide plausible mecha-nisms for this. Heim and coworkers reported on 49 women aged 18 to45 years separated into four groups: (1) one group with no childhood abuseor psychiatric disorder, (2) one group with current major depression andchildhood abuse, (3) one group with no depression but having sufferedabuse, and (4) one group with current major depression and no childhoodabuse. They were able to show that the abused women showed a greater cor-ticotropin (ACTH) response to a standardized speaking stress. The abusedwomen who also had current major depression showed a six times greaterrise in ACTH than those without depression or childhood abuse. This obser-vation is consistent with animal research among rodents and primates sug-gesting that there are critical periods during development when thesensitivity of the hypothalamic pituitary adrenal axis is set. Young animalsdeprived of certain nurturing or subjected to particular stress during thesecritical periods have been demonstrated to show increased stress hormonesin response to other stressors throughout their lives.
Antidepressant medication has been shown to treat these psychiatric disor- ders effectively. Cognitive-behavioral therapy also has been shown to work.
These disorders are often missed or dismissed by clinicians who conceptualizethe disorders as secondary to the injury or chronic pain problems. It is moreappropriate to think of the disorders as comorbid illnesses that can benefitfrom treatment. There is clear evidence that treatment of concurrent psychi-atric disorders improves the outcomes for patients with chronic pain The second reason to use antidepressant medications in patients with chronic pain concerns the treatment of sleep disturbance. Fifty percent to ANTIDEPRESSANT & ANTICONVULSANT MEDICATION 80% of patients with chronic pain have significant sleep disturbances Experimental disruption of slow-wave sleep has been shown to increasepain sensitivity in most studies. Clinical studies show a reciprocal relation-ship between sleep disturbance and pain. It is well known from depressiontreatment studies that persistence of insomnia strongly predicts persistenceor relapse of depression.
Chronic pain patients are often prescribed benzodiazepines to address their sleep disturbance ; however, chronic benzodiazepine therapy doesnot correct the disturbed sleep architecture typical of many patients withchronic pain but may, in fact, further decrease slow-wave sleep. It is unclearwhether the new ‘‘Z-drug’’ sleep agents, which are partial benzodiazepine re-ceptor agonists, are any better. The most popular among these agents, zol-pidem (Ambien) and the new eszopiclone (Lunesta), do seem to have someadvantages when compared with the traditional benzodiazepines in terms ofrebound after discontinuation and cognitive impairment, but it is not clearwhether they affect sleep architecture more favorably Antidepressants almost all suppress rapid eye movement (REM) sleep, whereas they have variable effects on sleep continuity and slow-wave sleep.
In general, the tricyclic antidepressants have a more predictably positive ef-fect on sleep continuity and slow-wave sleep than do the popular selectiveserotonin reuptake inhibitors (SSRIs). This factor may be one importantreason why the tricyclics have shown better effects in some chronic pain con-ditions. In some illnesses such as fibromyalgia, controlled trials suggest thatthe optimum treatment is the combination of a tricyclic with an SSRI The authors’ interpretation of these data is that fibromyalgia patients fre-quently require full-dose antidepressant treatment, which is difficult toachieve with the tricyclic antidepressants owing to side-effect problems. Ad-dition of an SSRI allows full antidepressant dosing, and the tricyclic portionof the regimen addresses problems with sleep continuity and slow-wavesleep. This theory has not been specifically tested.
The third principal reason to prescribe antidepressants to patients with chronic pain is that there is evidence of antidepressant analgesia indepen-dent of their effects on depression. Antidepressant analgesia has been wellinvestigated over the past two decades. A 1992 review article documentedthat antidepressant analgesia could be found in over half of previous studies,although, on average, the pain relief was only 50% . At that time, it ap-peared that the most responsive pain syndromes were neuropathic pain syn-dromes such as diabetes and postherpetic neuralgia. The next mostresponsive were headache, facial, and central pain syndromes. These find-ings have been largely confirmed in subsequent studies. Whether depressionwas present or an antidepressant effect noted appeared to make no differ-ence in analgesia. The meta-analysis was unable to confirm the then popular presumption that serotonin was more important than norepinephrine. In-stead, the investigators concluded that mixed agents affecting serotoninand norepinephrine were best.
The evidence suggesting that the tricyclics were better than the SSRIs for neuropathic pain began with the report by Max and coworkers in 1992.
His team at the National Institutes of Health performed two simultaneouscrossover randomized controlled trials, with 38 subjects receiving amitripty-line or desipramine and 46 subjects receiving fluoxetine or placebo. Moder-ate or greater improvement in pain was noted by 74% of the amitriptylinegroup and 61% of the desipramine group. These differences were statisti-cally equivalent. Moderate or greater improvement in pain was noted in48% of the fluoxetine group and was not statistically significantly differentfrom the 41% in the placebo group. Interestingly, it did not appear to makea difference in the response to tricyclics whether the patients were depressed;however, the SSRIs appeared to produce analgesia only if the patients wereclinically depressed.
Subsequently, other mixed action antidepressants have been shown to be effective for neuropathic pain. Bupropion (Wellbutrin) has been shown ina trial of 41 patients with neuropathic pain to be superior to placebo overa period of 6 weeks Venlafaxine (Effexor) has also been shown to besuperior to placebo for neuropathic pain. Rowbotham and coworkers studied 244 patients with diabetic neuropathic pain and showed that the re-lief in the group receiving venlafaxine, 150 mg over a period of 6 weeks, wassuperior to that in the placebo group. An earlier study by Sindrup and co-workers of 40 patients in a placebo-controlled crossover trial showedthat venlafaxine, 225 mg, and imipramine, 150 mg, were superior to placebofor neuropathic pain Duloxetine (Cymbalta) is an antidepressant that has been approved by the Food and Drug Administration (FDA) for the treatment of diabeticneuropathic pain. Goldstein and coworkers reported on a randomizedtrial of 457 patients treated for 12 weeks with 20, 60, or 120 mg a day ofduloxetine. The 60 mg and 120 mg a day doses demonstrated greater im-provement than placebo in the 24-hour average pain score beginning 1week after randomization. Duloxetine has a similar pharmacology to ven-lafaxine in that it is a reuptake inhibitor for serotonin and norepinephrine.
It differs from venlafaxine in that it is a norepinephrine reuptake inhibitorat lower doses. The clinical significance of this difference has yet to bedemonstrated.
Mirtazapine is another atypical antidepressant with some norepinephrine reuptake action. It has been shown to be effective for chronic headache.
Bendtsen and Jensen showed that mirtazapine, 15 to 30 mg for 8 weeks,was superior to placebo for prophylaxis of chronic tension headache. Anopen label study by Samborski and coworkers suggests its effective-ness in fibromyalgia. Mirtazapine is effective in improving sleep continuity.
The same antihistaminic action that promotes sleep with mirtazapine ANTIDEPRESSANT & ANTICONVULSANT MEDICATION unfortunately promotes weight gain, prompting many patients to discon-tinue mirtazapine.
St. John’s wort, a popular naturopathic remedy that has been shown to treat depression successfully in European studies, was tested by Sindrupand coworkers in a population with neuropathic pain, but it did notproduce greater analgesia than placebo.
Multiple meta-analyses of antidepressant analgesia have been reported over the past few years. Tomkins and coworkers analyzed 38 trials ofantidepressants for chronic headache. The relative risk of improvementwas 2.0 on antidepressants when compared with placebo, with 31% moreantidepressant-treated patients improving. The number needed to treat(NNT) was 3.2, and the effect size was large on average. There did notseem to be a difference in response rates by drug class or headache type. De-pression was not well monitored in these trials. In general, chronic headacheis one disorder in which the SSRI medications seem to function just as wellas the tricyclic or mixed reuptake inhibitor medications. There is a good da-tabase demonstrating efficacy for SSRIs in prophylaxis in patients who havechronic tension and recurrent migraine headaches. Holroyd and coworkersreported that stress management added significantly to an antidepres-sant regimen for chronic tension headache. Rates of achieving at least a50% reduction in the headache index were 64% for a combination of stressmanagement and amitriptyline or nortriptyline versus 38% for tricyclicsalone and 34% for stress management alone. The response rate was 29%for placebo alone.
O’Malley and coworkers performed a meta-analysis of trials testing antidepressant treatment for fibromyalgia in 2000. They reviewed 13 ade-quate randomized trials and found that there was evidence for antidepres-sant treatment of fibromyalgia with an odds ratio for improvement of 4.2;the NNT was 4. They were able to show that antidepressants improvedsleep, fatigue, pain, and well-being but not trigger points. Only one of thesetrials showed a significant correlation of analgesia with improvement in de-pression. There have been subsequent trials of antidepressants in fibromyal-gia including a recent trial with duloxetine demonstrating a superiorefficacy to placebo however, it is important to remember that patientswith fibromyalgia are heterogeneous psychiatrically. Some are relativelyeasy to treat and some are extremely difficult to treat, being resistant to psy-chopharmacologic regimens or intolerant of them. In general, the author(MDS) has found that fibromyalgia patients will improve if they can sleepand exercise.
Although the research literature concerning antidepressants and low back pain is large, many trials are poor in design. Two trials that are an exceptionto this rule were performed by Atkinson and coworkers. The first was a nor-triptyline versus placebo trial reported in 1998 The second was a trial ofmaprotiline or paroxetine versus diphenhydramine reported in 1999 The 1998 trial was able to demonstrate clearly that among older males with chronic low back pain nortriptyline produced greater pain relief thanplacebo. This difference was not related to pretreatment depression severityor depression response with treatment.
They followed this trial with a second trial to explore mechanisms. Map- rotiline, an older tricyclic drug that was the most potent norepinephrine re-uptake inhibitor at the time, was compared with paroxetine, a relativelypure serotonin reuptake inhibitor. These agents were compared with an ac-tive placebo (diphenhydramine), which is an antihistamine that producesa dry mouth and some sedation. An 8-week trial was performed randomiz-ing 103 subjects. Seventy-nine of the patients completed the trial. Each sub-ject had to have 6 months of back pain, and no one with a diagnosablemood disorder was entered into the trial. Pain relief by the highly sensitivedescriptor differential scale was 45% in the group that received maprotiline(up to 150 mg), 27% for those receiving placebo (up to 37.5 mg of diphen-hydramine), and 26% for those receiving paroxetine (up to 30 mg). Pain in-tensity diminished more markedly than pain unpleasantness. Although thisgroup had prolonged pain (median pain duration of 10 years), it was nota failed back syndrome group, because only 12% had had previous backsurgery. This finding was not simply a replication of neuropathic pain stud-ies, because only 14% had radicular pain. There was no change in self-reported or observer-rated depression in the active treatment groups. Inaddition, there was no significant correlation between the change in moodsymptoms and the change in pain intensity. On the basis of the Atkinson tri-als, the author believes that the noradrenergic antidepressants may be pref-erable not just for neuropathic pain syndromes but also for commonmusculoskeletal pain syndromes such as neck and back pain.
Not all pain syndromes show increased responsiveness to noradrenergic antidepressants. Headaches, particularly chronic daily headache, respondto SSRI antidepressants as well as tricyclic and serotonin norepinephrine re-uptake inhibitor (SNRI) medications. Nonischemic chest pain may be an-other area in which SSRIs do well. Cannon and coworkers originallyreported in 1994 that imipramine, 50 mg, was able to produce a 52% reduc-tion in chest pain episodes in a noncoronary chest pain group comparedwith a 39% reduction in a group treated with 0.1 mg twice daily of clonidineand a 1% reduction in a placebo group. Varia and coworkers reportedin 2000 on a cohort of 30 patients treated for 8 weeks with sertraline, 50 to200 mg/day, or placebo. Sixty-six percent of the sertraline-treated group meta criterion of greater than 50% pain relief versus 8% of the placebo group. Somewhat oddly, they were unable to show an effect of sertraline treat-ment on quality of life or depression measures. These trials awaitreplication.
Psychiatric disorders are common in patients with chronic pain, espe- cially if the pain is associated with significant activity impairment. Antide-pressants with norepinephrine reuptake are probably the psychotropicagents of choice for pain relief. These agents include nortriptyline, ANTIDEPRESSANT & ANTICONVULSANT MEDICATION bupropion, venlafaxine, mirtazapine, and duloxetine. SSRI and SNRI anti-depressants alone do not generally improve sleep continuity. This factormay be crucial in syndromes such as fibromyalgia and may mandate a com-bination of antidepressants with sedating agents such as the tricyclics trazo-done or mirtazapine.
Pain specialists typically discuss anticonvulsants primarily in relation to neuropathic pain, and the clearest evidence for their efficacy comes fromstudies of patients with this type of pain. As a first approximation, one couldsay that anticonvulsants should be used with enthusiasm in patients withneuropathic pain and with caution in patients with other types of pain; un-fortunately, this simple rule is beguiling because the boundaries around neu-ropathic pain are far from clear.
Neuropathic pain has been defined as ‘‘pain initiated or caused by a pri- mary lesion, dysfunction, or transitory perturbation in the peripheral or cen-tral nervous system’’ Some experts have argued that the term should belimited to pain caused by abnormalities in the peripheral nervous system, orat least that neuropathic pain of peripheral origin should be distinguishedfrom neuropathic pain of central origin A wide range of disorders of the peripheral nervous system can cause per- sistent pain, including peripheral polyneuropathies, entrapment neuropa-thies, radiculopathies, and traumatic injuries to nerves. Pain can also beassociated with damage to the CNS. Examples include central poststrokepain and pain in spinal cord injury. The characteristic pain syndromesthat occur in the context of these CNS lesions are sometimes called centralpain syndromes. Some research supports the efficacy of treating central painsyndromes with anticonvulsants although it is not as extensive or con-sistent as support for their use in painful disorders of the peripheral nervoussystem.
A much more difficult question is whether chronic pain syndromes asso- ciated with CNS hypersensitivity or dysfunction should be construed as neu-ropathic pain syndromes. To a large extent, the answer hinges on theinterpretation given to the word ‘‘dysfunction’’ in the International Associ-ation of Pain definition of neuropathic pain . As discussed in that text,nociception can lead to changes in the function of the CNS in the absenceof any clear evidence of damage to the CNS. If pain associated with suchfunctional changes is construed as neuropathic pain, the range of conditionsfalling under the rubric of neuropathic pain increases enormously. For ex-ample, complex regional pain syndrome and fibromyalgia have been construed by some as neuropathic pain disorders.
The issue of a narrow versus a broad conceptualization of neuropathic pain has been summarized by Rowbotham as follows , ‘‘An examinationof how the term neuropathic pain is defined reveals a conceptual split into 2partially overlapping groups of disorders: those with demonstrable pathol-ogy in the nervous system and those characterized primarily by enduringdysfunction of the nervous system. Requiring demonstrable pathology inthe nervous system in the definition of neuropathic pain is the traditionalapproach. The expansion of the definition to require only enduring nervoussystem dysfunction is less palatable because it opens the classification tomany disorders of uncertain etiology.’’ If fibromyalgia is construed as a neuropathic pain syndrome, only a short conceptual leap is needed to infer the presence of neuropathic pain in anychronic pain patient who demonstrates evidence of CNS hypersensitivity.
It is beyond the scope of this article to discuss procedures for identifyingCNS hypersensitivity in clinical populations, but one indicator of such hy-persensitivity is widespread hyperalgesia in response to palpation overmuscles. If this criterion is accepted, CNS hypersensitivity would probablybe found to be the rule rather than the exception in chronic pain syndromes.
For example, essentially all patients with myofascial pain would be con-strued as having CNS hypersensitivity .
Anticonvulsants in the treatment of neuropathic pain: research findings Anticonvulsants are thought to inhibit seizures by multiple mechanisms, including functional blockade of voltage-gated sodium channels, functionalblockade of voltage-gated calcium channels, direct or indirect enhancementof inhibitory GABAergic neurotransmission, and inhibition of glutamater-gic neurotransmission . The result is that they reduce the neuronal hy-perexcitability that is fundamental to seizure disorders.
Because neuropathic pain is also characterized by neuronal hyperexcit- ability clinicians and researchers have reasoned that anticonvulsantsmight alleviate it. This supposition is supported by a substantial amountof empirical data on the effectiveness of anticonvulsants in neuropathicpain. The multiple studies in this area have been the subject of recent sys-tematic reviews and have been considered by expert panels devoted to devel-oping guidelines for treating neuropathic pain. The discussion herein relieslargely on guidelines developed by Dworkin and coworkers , three recentCochrane systematic reviews and a systematic review by Goodmanand coworkers .
Goodman and coworkers searched for studies relevant to several anticon- vulsantsdphenytoin, valproic acid, carbamazepine, gabapentin, lamotrigine,oxcarbazepine, zonisamide, levetiracetam, tiagabine, and topiramate. Thestudies they included in their review dealt with the first five. As is typical ofsystematic reviews on therapies for chronic pain, Goodman and colleaguesdescribed significant methodologic flaws in the research on anticonvulsants ANTIDEPRESSANT & ANTICONVULSANT MEDICATION in neuropathic pain and ended up with cautious conclusions. In combinationwith the guidelines developed by Dworkin and coworkers the review per-mits the following conclusions to be reached with reasonable confidence:  The effectiveness of anticonvulsants in the treatment of neuropathic pain has been studied extensively. For example, the review by Goodmanand coworkers considered 37 studies.
 The preponderance of evidence supports the conclusion that, as a group, anticonvulsants are effective in the amelioration of neuropathic pain incomparison with placebos.
 This conclusion must be tempered by the fact that studies on anticon- vulsants have been concentrated heavily on a few neuropathic pain con-ditions, especially diabetic neuropathy, postherpetic neuralgia, andtrigeminal neuralgia. It is possible that some neuropathic pain condi-tions are responsive to anticonvulsants, whereas others are not. A relatedissue is that some observers have proposed that, for the purposes oftreatment, neuropathic pain syndromes need to be classified accordingto the specific symptoms that patients experience (eg, spontaneous burn-ing pain versus allodynia versus lancinating pain) or the pathophysio-logic mechanisms underlying the neuropathic pain The generalpoint is that research to date has not been broad enough to determinehow robust anticonvulsants are in treating the broad scope of symp-toms, pathophysiologic processes, and diagnostic entities that are sub-sumed under the broad definition of ‘‘neuropathic pain.’’  There have been virtually no head-to-head comparisons between differ- ent anticonvulsants One indirect method for comparing the effec-tiveness of different drugs is to compare the NNT for them; however,comparisons with respect to NNTs must be interpreted cautiously ). The summary NNT data suggest that carbamazepine is more effec-tive overall than gabapentin in the treatment of neuropathic pain, butthe determination of NNT for carbamazepine was based primarily onstudies of trigeminal neuralgia, whereas the NNT for gabapentin wasbased primarily on studies of diabetic neuropathy and postherpetic neu-ralgia. Given this difference, it may be meaningless to compare summaryNNTs for the two drugs. Another problem is that data from some stud-ies do not permit a reviewer to calculate the NNT. Summary NNTs arebased on only a portion of the studies that have been done with a drug.
 Owing to the previously cited methodologic issues and several others, research to date does not permit any conclusive statements to be madeabout the relative efficacy of different anticonvulsants in the treatmentof neuropathic pain. A few anticonvulsants (eg, carbamazepine, gaba-pentin, and lamotrigine) have been studied in several randomized con-trolled trials on neuropathic pain, whereas little or no systematicresearch has been done on the effectiveness of some of the newer anti-convulsants (eg, zonisamide, levetiracetam, tiagabine).
Table 1NNT for CZP, gabapentin, and antidepressants in neuropathic pain: calculated from systemicreviews The NNT is calculated by first dichotomizing subjects in a placebo controlled clinical trial into those who achieved a certain degree of pain relief (eg, moderate relief or more) and thosewho did not. The percentage of subjects who achieve moderate or better pain relief is calculatedfor those who received the drug under study and for those in the placebo control group. Thedifference between these two percentages is then determined, and this number is divided into100 to determine the NNT for the drug. For example, if a study found that 60% of subjectsreceiving gabapentin achieved at least moderate pain relief versus 20% of subjects in theplacebo control group, the NNT for gabapentin would be 100/40 ¼ 2.5.
Abbreviations: CZP, carbamazepine; NNT, number needed to treat.
a Based on 2 studies, with total N ¼ 47 b Based on 2 studies, with total N ¼ 207. Dose ranged from 2400 to 3600 mg/d .
c Based on 4 studies, with total N ¼ 77 .
d Based on 4 studies, with total N ¼ 142. Dose ranged from 900 to 3600 mg/d .
e Based on 5 studies, with total N ¼ 98 .
f Based on 4 studies, with total N ¼ 91 .
g Based on 7 studies, with total N ¼ 466 h For amitriptyline only; based on 7 studies, with total N ¼ 112. Doses up to 150 mg/d .
 Assessing the relative effectiveness of different anticonvulsants is compli- cated by the fact that some of the newer ones have been investigated sorecently that the relevant studies have not been considered in systematicreviews of anticonvulsants. For example, neither of the two most com-prehensive reviews considers several recent studies on topiramate.
Clinical use of anticonvulsants in neuropathic pain A clinician who plans to use anticonvulsants in the treatment of neuro- pathic pain conditions will be heartened by evidence that as a class theyare effective; however, he or she will immediately be confronted witha host of more specific questions that are not answered conclusively by re-search. This section discusses strategies for using anticonvulsants that arebased on clinical experience rather than on research.
The first step in using anticonvulsants to treat neuropathic pain is to es- tablish that a patient has neuropathic pain. This step requires the identifica-tion of a disorder in the nervous system and reasonable clinical evidence thatthe patient’s pain is a produced by the disorder. The latter determination is ANTIDEPRESSANT & ANTICONVULSANT MEDICATION based on information about the distribution of symptoms, the quality ofpain, and the settings under which the patient experiences pain .
With regard to distribution, neuropathic pain is suggested when the pa- tient reports symptoms in a pattern that is consistent with the suspected neu-ropathy. The precise distribution, of course, depends on the neuropathy, forexample, an L5 dermatome for an individual with an L5 radiculopathy andsymptoms primarily in the feet in a patient with a diabetic neuropathy. Withregard to quality, a combination of numbness and pain is suggestive of neu-ropathic pain. Also, the pain is frequently described as burning or lancinat-ing. With respect to settings in which pain occurs, neuropathic pain is oftendistinctive in that it is worse when a patient is inactive. A patient with a di-abetic neuropathy will often experience more symptoms when trying to sleepat night than when walking during the day. Another distinctive feature isthat a patient with neuropathic pain will often describe discomfort fromthe light pressure of clothing on the body area that is symptomatic.
Antidepressants and anticonvulsants have demonstrable efficacy in the treatment of neuropathic pain and both classes of medicationhave been recommended as first-line therapy for neuropathic pain .
Indirect comparisons based on NNTs suggest that antidepressants (espe- cially amitriptyline) are more effective than anticonvulsants; however, thisconclusion is challenged by three head-to-head comparisons between anti-convulsants and amitriptyline . These studies failed to demonstrateany difference in effectiveness; therefore, the currently available researchdata do not provide a clear basis for choosing between an antidepressantand an anticonvulsant for initial therapy of neuropathic pain. The authorsrecommend that if a patient demonstrates significant depression, anxiety, orsleep disturbance, initial therapy should be with an antidepressant that af-fects the reuptake of norepinephrine and serotonin (eg, nortriptyline, dulox-etine, or venlafaxine) at a dose that is effective for treating depression. Inpatients without significant emotional distress, anticonvulsants and antide-pressants are equally appropriate as choices for initial therapy.
One recent consensus article identified not only anticonvulsants (spe- cifically gabapentin) and tricyclic antidepressants as first-line pharmacologictherapies for neuropathic pain but also three other agents: opiates, trama-dol, and a 5% lidocaine patch (specifically for postherpetic neuralgia). Noresearch permits a clinician to choose where to start in this list.
In the absence of head-to-head comparisons between different anticon- vulsants, it is reasonable to start with the ones that have been studied themostdgabapentin and carbamazepine. In particular, gabapentin is attrac-tive because it does not interact significantly with other drugs, and becauseits characteristic adverse effects are reversible with termination of the drug.
The authors recommend it for initial anticonvulsant therapy in most neuro-pathic pain. One exception to this general principle might be trigeminal neu-ralgia or any neuropathic pain condition in which lancinating paindominates. The presence of significant lancinating pain should make onestrongly consider carbamazepine. One might consider an anticonvulsantother than Gabapentin in a patient with cognitive impairment. In that set-ting, the authors have found lamotrigine to be a good choice, because itseems to cause less cognitive impairment than gabapentin.
Options for the patient who fails initial therapy with an anticonvulsant If a patient fails to benefit from a maximally tolerated dose of gabapentin, one option would be to switch to another anticonvulsant. Although there areno scientific data to guide the clinician in choosing an alternative anticonvul-sant, it seems logical to choose one (eg, carbamazepine or lamotrigine) thatinfluences Naþ ion channels, because gabapentin is thought to stabilize neu-rons primarily by its effects on Caþþ ion channels or enhancement of GA-BAergic neurotransmission . It would also be reasonable to considervalproate or one of the less studied second-generation anticonvulsants,such as oxcarbazepine , topiramate zonisamide or levetir-acetam . Another possibility is pregabalin, which was released during thelatter part of 2005. It has the disadvantage of being new but the advantage ofhaving FDA approval for use in postherpetic neuralgia and painful diabeticneuropathy Although the effectiveness of phenytoin for neuropathicpain has been studied, the results have been equivocal and the authorsdo not recommend its use .
An alternative strategy in the patient with a failed trial on gabapentin would be to switch an antidepressant, or to prescribe an opiate, tramadol,or a 5% lidocaine patch.
Although monotherapy for neuropathic pain has been reported to pro- duce significant benefit for approximately 70% of patients the re-sponses are often partial As a practical matter, one will commonlyencounter patients who continue to complain of significant pain despite tri-als on anticonvulsants, antidepressants, and other first-line agents. In thissetting, combination therapy is a reasonable strategy.
Unfortunately, combination therapy involving anticonvulsants has re- ceived only minimal attention in research. Simpson found that amongdiabetic neuropathy patients who failed to respond to gabapentin, patientswho subsequently received a combination of gabapentin and venlafaxinereported more pain relief than did controls who received gabapentinand a placebo. Gilron and coworkers recently studied the effects ofmorphine, gabapentin, and a combination of morphine plus gabapentinin a crossover trial of patients with diabetic neuropathy and postherpeticneuralgia. Although both morphine and gabapentin led to pain reduction, ANTIDEPRESSANT & ANTICONVULSANT MEDICATION the combination of the two agents was the most effective regimen. Thesestudies suggest that combination therapy involving anticonvulsants maybe effective, but more research is needed to determine what combinationsare optimal.
The most logical drugs to combine would be an antidepressant and an anticonvulsant, because these two classes of drugs are thought to affect neu-ropathic pain by different mechanisms. The authors frequently see neuro-pathic pain patients who are receiving this type of combination therapyfrom their primary care providers. Typically, the anticonvulsant is pre-scribed for pain control, and the antidepressant is prescribed because ofthe patient’s emotional dysfunction.
Anticonvulsants in other painful conditions A considerable amount of research has been done on the efficacy of anti- convulsants in central pain syndromes, especially poststroke pain and painin spinal cord injury Although the evidence for efficacy of anticonvul-sants in these disorders is less impressive than it is in neuropathic pain fromdisorders of the peripheral nervous system, it is reasonable to give patientswith any central pain syndrome a trial of anticonvulsants.
The role of anticonvulsants in migraine headache has been studied exten- sively, and the results are generally favorable, especially for valproate .
As discussed previously, some experts have proposed that fibromyalgia is a neuropathic pain syndrome. Regardless of the utility of this hypothesis, itis clear that fibromyalgia involves CNS hypersensitivity To the extentthat this hypersensitivity is a reflection of hyperresponsiveness of neuronsin nociceptive pathways, the use of anticonvulsants in patients with fibro-myalgia could be rationalized in the same way that it is rationalized forthe treatment of pain that is unequivocally neuropathic. Moreover, as dis-cussed by Curatolo and colleagues elsewhere in this issue, CNS hypersensi-tivity is by no means restricted to fibromyalgia; in fact, it probably occurs inmost chronic pain disorders. There is a conceptual rationale for consideringanticonvulsants in the treatment of many patients with chronic pain.
Obviously, such an expansive use of anticonvulsants needs empirical sup- port rather than just theoretical plausibility. Some data support the efficacyof anticonvulsants in musculoskeletal pain syndromes with CNS hypersen-sitivity. In particular, pregabalin, a new anticonvulsant, has been shown toreduce pain in fibromyalgia in a dose-dependent manner .
In contrast, studies of anticonvulsants in musculoskeletal conditions such as chronic low back pain have yielded inconsistent results . Thesestudies are difficult to evaluate in relation to the hypothesis that anticonvul-sants are effective for disorders characterized by CNS hypersensitivity, be-cause participants in the studies were not systematically assessed for thepresence of indicators of CNS hypersensitivity. In an ideal study, patientswith a common musculoskeletal condition such as axial low back pain would be divided into groups who did and did not show clinical evidence ofCNS hypersensitivity, and both groups would be treated with an anticonvul-sant. The expectation would be that only the patients demonstrating CNShypersensitivity would benefit from the anticonvulsant therapy.
Clinicians in the authors’ community frequently prescribe anticonvulsants (especially gabapentin) for patients with refractory musculoskeletal pain.
In the absence of definitive research data, the authors’ clinical experience sug-gests that a trial on an anticonvulsant is worth performing for a patient withchronic musculoskeletal pain and evidence of evidence of CNS hypersensi-tivity; however, these patients are often refractory to essentially all therapies,including anticonvulsants. If an anticonvulsant is prescribed, it is incumbenton the physician to monitor the patient’s response carefully, and to discon-tinue the drug if it does not produce clear clinical improvement.
The clinical information available on the patient in Case 1 in the intro- ductory article suggests that his right lower extremity pain reflects a centralpain syndrome. Although central pain syndromes represent a heterogeneousgroup of disorders with respect to the underlying CNS dysfunction, there isevidence that at least some of them respond to tricyclic antidepressants, lamotrigine and possibly gabapentin . It should be assumedthat a patient with a history of prolonged coma from a traumatic brain in-jury would be sensitive to cognitive adverse effects from any centrally actingdrug. This consideration implies that tricyclic antidepressants and gabapen-tin should be used with caution. Among the anticonvulsants, lamotriginewould probably be the best choice for this patient.
Anticonvulsants and several antidepressants, including tricyclics, SSRIs, and dual reuptake inhibitors, have been shown to be effective in reducingpain associated with diabetic neuropathy; therefore, the patient in Case 2provides the clinician with a plethora of pharmacologic options involvingantidepressants or anticonvulsants. There is no obvious starting point fortrials on these drugs, except that tricyclic antidepressants should be usedwith caution in an older individual with cardiovascular disease.
Tricyclic antidepressants have repeatedly demonstrated effectiveness in the treatment of fibromyalgia, and one study has shown duloxetine to be ef-fective. SSRIs and venlafaxine have demonstrated benefit in some studies.
Among the anticonvulsants, only pregabalin has been shown to benefit ANTIDEPRESSANT & ANTICONVULSANT MEDICATION patients with fibromyalgia Based on research to date, it would be rea-sonable for the clinician to have the patient in Case 2 undergo trials withseveral different sedating antidepressants and pregabalin.
It is not obvious that antidepressants or anticonvulsants are likely to help the patient in Case 4 with persistent spinal pain in the absence of any evi-dence of radiculopathy; however, if this patient demonstrated widespreadhyperalgesia, the clinician might conclude that CNS sensitization was con-tributing to his pain, and perhaps that the patient had a fibromyalgia-likecondition. In that situation, it would be reasonable to consider medicationsthat have been shown to help fibromyalgia (see Case 3), as long as the clini-cian is aware that there are no well-controlled studies validating the effec-tiveness of such medications for the type of patient described in Case 4.
[1] Melzack R, Wall PD. Pain mechanisms: a new theory. Science 1965;150:971–9.
[2] Fields HL. Pain modulation: expectation, opioid analgesia and virtual pain. Prog Brain Res [3] Craig AD. A new view of pain as a homeostatic emotion. Trends Neurosci 2003;26(6):303–7.
[4] Bair MJ, Robinson RL, Katon W, et al. Depression and pain comorbidity: a literature re- view. Arch Intern Med 2003;163(20):2433–45.
[5] Gureje O, Von Korff M, Simon GE, et al. Persistent pain and well-being: a World Health Organization study in primary care. JAMA 1998;280(2):147–51.
[6] Gureje O, Simon GE, Von Korff M. A cross-national study of the course of persistent pain in primary care. Pain 2001;92(1–2):195–200.
[7] McWilliams LA, Cox BJ, Enns MW. Mood and anxiety disorders associated with chronic pain: an examination in a nationally representative sample. Pain 2003;106(1–2):127–33.
[8] Sharp TJ. The prevalence of post-traumatic stress disorder in chronic pain patients. Curr Pain Headache Rep 2004;8(2):111–5.
[9] Hickling EJ, Blanchard EB. Post-traumatic stress disorder and motor vehicle accidents.
[10] Asmundson GJG, Norton G, Allerdings M, et al. Post-traumatic stress disorder and work- related injury. J Anxiety Disord 1998;12:57–69.
[11] Roy-Byrne P, Smith WR, Goldberg J, et al. Post-traumatic stress disorder among patients with chronic pain and chronic fatigue. Psychol Med 2004;34(2):363–8.
[12] Heim C, Newport DJ, Heit S, et al. Pituitary-adrenal and autonomic responses to stress in women after sexual and physical abuse in childhood. JAMA 2000;284(5):592–7.
[13] Guzman J, Esmail R, Karjalainen K, et al. Multidisciplinary biopsychosocial rehabilitation for chronic low back pain. Cochrane Database Syst Rev 2002;1:CD000963.
[14] Smith MT, Haythornthwaite JA. How do sleep disturbance and chronic pain interrelate? [15] Neutel CI. The epidemiology of long-term benzodiazepine use. Int Rev Psychiatry 2005; [16] Stiefel F, Stagno D. Management of insomnia in patients with chronic pain conditions. CNS [17] O’Malley PG, Balden E, Tomkins G, et al. Treatment of fibromyalgia with antidepressants: a meta-analysis. J Gen Intern Med 2000;15(9):659–66.
[18] Onghena P, Van Houdenhove B. Antidepressant-induced analgesia in chronic nonmalignant pain: a meta-analysis of 39 placebo-controlled studies. Pain 1992;49(2):205–19.
[19] Max MB, Lynch SA, Muir J, et al. Effects of desipramine, amitriptyline, and fluoxetine on pain in diabetic neuropathy. N Engl J Med 1992;326(19):1250–6.
[20] Semenchuk MR, Sherman S, Davis B. Double-blind, randomized trial of bupropion SR for the treatment of neuropathic pain. Neurology 2001;57(9):1583–8.
[21] Rowbotham MC, Goli V, Kunz NR, et al. Venlafaxine extended release in the treatment of painful diabetic neuropathy: a double-blind, placebo-controlled study. Pain 2004;110(3):697–706.
[22] Sindrup SH, Bach FW, Madsen C, et al. Venlafaxine versus imipramine in painful polyneu- ropathy: a randomized, controlled trial. Neurology 2003;60(8):1284–9.
[23] Goldstein DJ, Lu Y, Detke MJ, et al. Duloxetine vs. placebo in patients with painful diabetic neuropathy. Pain 2005;116(1–2):109–18.
[24] Bendtsen L, Jensen R. Mirtazapine is effective in the prophylactic treatment of chronic ten- sion-type headache. Neurology 2004;62(10):1706–11.
[25] Samborski W, Lezanska-Szpera M, Rybakowski JK. Open trial of mirtazapine in patients with fibromyalgia. Pharmacopsychiatry 2004;37(4):168–70.
[26] Sindrup SH, Madsen C, Bach FW, et al. St. John’s wort has no effect on pain in polyneu- [27] Tomkins GE, Jackson JL, O’Malley PG, et al. Treatment of chronic headache with antide- pressants: a meta-analysis. Am J Med 2001;111(1):54–63.
[28] Holroyd KA, O’Donnell FJ, Stensland M, et al. Management of chronic tension-type head- ache with tricyclic antidepressant medication, stress management therapy, and their combi-nation: a randomized controlled trial. JAMA 2001;285(17):2208–15.
[29] O’Malley PG, Balden E, Tomkins G, et al. Treatment of fibromyalgia with antidepressants: a meta-analysis. J Gen Intern Med 2000;15(9):659–66.
[30] Goldenberg DL, Burckhardt C, Crofford L. Management of fibromyalgia syndrome. JAMA [31] Arnold LM, Lu Y, Crofford IJ, et al. A double-blind multicenter trial comparing duloxetine to placebo in the treatment of fibromyalgia with or without major depressive disorder. Ar-thritis Rheum 2004;50:2974–84.
[32] Atkinson JH, Slater MA, Williams RA, et al. A placebo-controlled randomized clinical trial of nortriptyline for chronic low back pain. Pain 1998;76(3):287–96.
[33] Atkinson JH, Slater MA, Wahlgren DR, et al. Effects of noradrenergic and serotonergic antidepressants on chronic low back pain intensity. Pain 1999;83(2):137–45.
[34] Cannon RO 3rd, Quyyumi AA, Mincemoyer R, et al. Imipramine in patients with chest pain despite normal coronary angiograms. N Engl J Med 1994;330(20):1411–7.
[35] Varia I, Logue E, O’Connor C, et al. Randomized trial of sertraline in patients with unex- plained chest pain of noncardiac origin. Am Heart J 2000;140(3):367–72.
[36] Merskey H, Bogduk N, editors. Classification of chronic pain. 2nd edition. Seattle (WA): [37] Woolf CJ. Dissecting out mechanisms responsible for peripheral neuropathic pain: implica- tions for diagnosis and therapy. Life Sci 2004;74:2605–10.
[38] Nicholson BD. Evaluation and treatment of central pain syndromes. Neurology 2004; [39] Ghai B, Dureja GP. Complex regional pain syndrome: a review. J Postgrad Med 2004;50(4): [40] Grabow TS, Christo PJ, Raja SN. Complex regional pain syndrome: diagnostic controversies, psychological dysfunction, and emerging concepts. Adv Psychosom Med 2004;25:89–101.
[41] Crofford LJ. The relationship of fibromyalgia to neuropathic pain syndromes. J Rheumatol [42] Price DD, Staud R. Neurobiology of fibromyalgia syndrome. J Rheumatol 2005;32(Suppl ANTIDEPRESSANT & ANTICONVULSANT MEDICATION [43] Rowbotham MC. Is fibromyalgia a neuropathic pain syndrome? J Rheumatol 2005; [44] Robinson JP, Arendt-Nielsen L. Muscle pain syndromes. In: Braddom R, editor. Physical medicine and rehabilitation. 3rd edition. In press.
[45] Czapinski P, Blaszczyk B, Czuczwar SJ. Mechanisms of action of antiepileptic drugs. Curr [46] Soderpalm B. Anticonvulsants: aspects of their mechanisms of action. Eur J Pain 2002; [47] Jensen TS. Anticonvulsants in neuropathic pain: rationale and clinical evidence. Eur J Pain [48] Dworkin RH, Backonja M, Rowbotham MC, et al. Advances in neuropathic pain. Arch [49] Wiffen PF, McQuay HF, Moore RA. Carbamazepine for acute and chronic pain. Cochrane [50] Wiffen PF, McQuay HF, Edwards JE, et al. Gabapentin for acute and chronic pain. Co- chrane Database Syst Rev 2005;3:CD005452.
[51] Wiffen P, Collins S, McQuay H, et al. Anticonvulsant drugs for acute and chronic pain. Co- chrane Database Syst Rev 2005;3:CD001133.
[52] Goodman F, Glassman P, Qiufei M, et al. Drug class review on antiepileptic drugs in bipolar mood disorder and neuropathic pain. Santa Monica (CA): Rand; 2004.
[53] Dib JG. Focus on topiramate in neuropathic pain. Curr Med Res Opin 2004;20(12):1857–61.
[54] Raskin P, Donofrio PD, Rosenthal NR, et al. Topiramate vs. placebo in painful diabetic neuropathy: analgesic and metabolic effects. Neurology 2004;63:865–73.
[55] Thienel U, Neto W, Schwabe SK, et al. Topiramate in painful diabetic polyneuropathy: find- ings from three double-blind placebo-controlled trials. Acta Neurol Scand 2004;110(4):221–31.
[56] Saarto T, Wiffen PF. Antidepressants for neuropathic pain. Cochrane Database Syst Rev [57] Namaka M, Gramlich CR, Ruhlen D, et al. A treatment algorithm for neuropathic pain.
[58] Dallochio C, Buffa C, Mazzarello P, et al. Gabapentin vs. amitriptyline in painful diabetic neuropathy: an open label pilot study. J Pain Symptom Manage 2000;20(4):280–5.
[59] Leijon G, Boivie J. Central post-stroke painda controlled trial of amitriptyline and carba- [60] Morello CM, Leckband SG, Stoner CP, et al. Randomized double-blind study comparing the efficacy of gabapentin with amitriptyline on diabetic peripheral neuropathy pain. ArchIntern Med 1999;159(16):1931–7.
[61] Belliotti TR, Capiris T, Ekhato V, et al. Structure-activity relationships of pregabalin and analogues that target the alpha2-delta protein. J Med Chem 2005;48:2294–307.
[62] Kochar DK, Jain N, Agarwal RP, et al. Sodium valproate in the management of painful neu- ropathy in type 2 diabetesda randomized placebo controlled study. Acta Neurol Scand2002;106(5):248–52.
[63] Pappagallo M. Newer antiepileptic drugs: possible uses in the treatment of neuropathic pain and migraine. Clin Ther 2003;25(10):2506–38.
[64] Dogra S, Beydoun S, Mazzola J, et al. Oxcarbazepine in painful diabetic neuropathy: a ran- domized, placebo-controlled study. Eur J Pain 2005;9(5):543–54.
[65] Atli A, Dogra S. Zonisamide in the treatment of painful diabetic neuropathy: a randomized, double-blind, placebo-controlled pilot study. Pain Med 2005;6(3):225–34.
[66] Price MJ. Levetiracetam in the treatment of neuropathic pain: three case studies. Clin J Pain [67] Hempenstall K, Nurmikko TJ, Johnson RW, et al. Analgesic therapy in postherpetic neural- gia: a quantitative systematic review. PLoS Med 2005;2(7):164.
[68] Freynhagen R, Strojek K, Griesing T, et al. Efficacy of pregabalin in neuropathic pain evaluated in a 12-week, randomised, double-blind, multicentre, placebo-controlled trialof flexible- and fixed-dose regimens. Pain 2005;115(3):254–63.
[69] Dworkin RH, Corbin AE, Young JP, et al. Pregabalin for the treatment of postherpetic neu- ralgia: a randomized, placebo-controlled trial. Neurology 2003;60:1274–83.
[70] Sabatowski R, Galvez R, Cherry DA, et al. Pregabalin reduces pain and improves sleep and mood disturbances in patients with post-herpetic neuralgia: results of a randomized,placebo-controlled clinical trial. Pain 2004;109:26–35.
[71] Richter RW, Portenoy R, Sharma U, et al. Relief of painful diabetic peripheral neuropathy with pregabalin: a randomized, placebo-controlled trial. J Pain 2005;6(4):253–60.
[72] Chadda VS, Mathur MS. Double-blind study of the effects of diphenylhydantoin sodium on diabetic neuropathy. J Assoc Physicians India 1978;32:403–6.
[73] Saudek CD, Werns S, Reidenberg MM. Phenytoin in the treatment of diabetic symmetrical polyneuropathy. Clin Pharmacol Ther 1977;22:196–9.
[74] Vinik A. Clinical review: use of antiepileptic drugs in the treatment of chronic painful dia- betic neuropathy. J Clin Endocrinol Metab 2005;90:4936–45.
[75] Simpson DA. Gabapentin and venlafaxine for the treatment of painful diabetic neuropathy.
J Clin Neuromusc Dis 2001;3:53–62.
[76] Gilron I, Bailey JM, Dongsheng T, et al. Morphine, gabapentin, or their combination for neuropathic pain. N Engl J Med 2005;352:1324–34.
[77] Crofford LJ, Rowbotham MC, Mease PJ, et al. Pregabalin 1008–105 Study Group. Prega- balin for the treatment of fibromyalgia syndrome: results of a randomized, double-blind, pla-cebo-controlled trial. Arthritis Rheum 2005;52(4):1264–73.
[78] McCleane GJ. Does gabapentin have an analgesic effect on background, movement and re- ferred pain? A randomized, double-blind, placebo-controlled study. Pain Clin 2001;13:103–7.
[79] McCleane GJ. Gabapentin reduces chronic benign nociceptive pain: a double-blind, pla- cebo-controlled cross-over study. Pain Clin 2000;12:81–5.
[80] Leijon G, Boivie J. Central post-stroke painda controlled trial of amitriptyline and carba- [81] Panerai AE, Monza G, Molivia P, et al. A randomized, within-patient, cross-over, placebo- controlled trial on the efficacy and tolerability of the tricyclic antidepressants chlomipramineand nortriptyline in central pain. Acta Neurol Scand 1990;82(1):34–8.
[82] Vestergaard K, Andersen G, Gottrup H, et al. Lamotrigine for central poststroke pain: a ran- domized controlled trial. Neurology 2001;56:184–90.
[83] Tai Q, Kirshblum S, Chen B, et al. Gabapentin in the treatment of neuropathic pain after spinal cord injury: a prospective, randomized, double-blind, crossover trial. J Spinal CordMed 2002;25:100–5.

Source: http://nyulmc.com/pmr/residency/resources/PMR%20clinics%20NA/PMR%20clinics%20NA_pain/antidepressant%20and%20anticonvulsant%20medication%20for%20chronic%20pain.pdf