1. Top of page
  2. Abstract
  3. Pharmacology of antidepressants and pain mechanisms
  4. Antidepressants in neuropathic pain
  5. Discussion
  6. References

Abstract: Neuropathic pain is due to lesion or dysfunction of the peripheral or central nervous system. Tricyclic antidepressants and anticonvulsants have long been the mainstay of treatment of this type of pain. Tricyclic antidepressants may relieve neuropathic pain by their unique ability to inhibit presynaptic reuptake of the biogenic amines serotonin and noradrenaline, but other mechanisms such as N-methyl-D-aspartate receptor and ion channel blockade probably also play a role in their pain-relieving effect. The effect of tricyclic antidepressants in neuropathic pain in man has been demonstrated in numerous randomised, controlled trials, and a few trials have shown that serotonin noradrenaline and selective serotonin reuptake inhibitor antidepressants also relieve neuropathic pain although with lower efficacy. Tricyclic antidepressants will relieve one in every 2–3 patients with peripheral neuropathic pain, serotonin noradrenaline reuptake inhibitors one in every 4–5 and selective serotonin reuptake inhibitors one in every 7 patients. Thus, based on efficacy measures such as numbers needed to treat, tricyclic antidepressants tend to work better than the anticonvulsant gabapentin and treatment options such as tramadol and oxycodone, whereas the serotonin noradrenaline reuptake inhibitor venlafaxine appears to be equally effective with these drugs and selective serotonin reuptake inhibitors apparently have lower efficacy. Head-to-head comparisons between antidepressants and the other analgesics are lacking. Contraindications towards the use of tricyclic antidepressants and low tolerability in general of this drug class – may among the antidepressants – favour the use of the serotonin noradrenaline reuptake inhibitors. A recent study on bupropion, which is a noradrenaline and dopamine uptake inhibitor, indicated a surprisingly high efficacy of this drug in peripheral neuropathic pain. In conclusion, antidepressants represent useful tools in neuropathic pain treatment and must still be considered as first line treatments of neuropathic pain. However, without head-to-head comparisons between antidepressants and other analgesics, it is not possible to provide real evidence-based treatment algorithms for neuropathic pain.

Neuropathic pain is pain caused by a lesion or dysfunction in the nervous system with the primary lesion or dysfunction affecting either the peripheral or central nervous system. Prominent examples of peripheral neuropathic pain conditions are painful diabetic neuropathy, post-herpetic neuralgia and post-surgical neuropathic pain and the most common central neuropathic pain conditions are post-stroke pain, pain in multiple sclerosis and spinal cord injury pain. Neuropathic pain is also present in some other patient groups, e.g. cancer patients. Therefore, the scope of neuropathic pain is broader than the classical peripheral and central neuropathic pain conditions. The neuropathic pain conditions are most often chronic in nature and represent a real challenge in clinical practice both due to their frequency, severity and the limited number of effective treatment options.

The mainstay of pharmacological treatment of neuropathic pain is antidepressant and anticonvulsant drugs, i.e. drugs developed to be used in completely different conditions. Tricyclic antidepressants were introduced in painful diabetic neuropathy based on empiric observations about 30 years ago (Davis et al. 1977), but their potential analgesic properties were already noted shortly after their introduction more than 40 years ago (Paolo et al. 1960). The benefit of antidepressants in neuropathic pain conditions has later been demonstrated in numerous controlled trials, and there is an understanding of their mode of action in neuropathic pain from their pharmacology and the pathophysiological mechanisms of neuropathic pain. It has been shown repeatedly that antidepressants work both in patients with normal and depressed mood in a number of neuropathic pain conditions (Max et al. 1987, 1988 & 1991; Leijon & Boivie 1989; Kishore-Kumar et al. 1990), i.e. these drugs possess a genuine analgesic effect. This, of course, does not exclude that they relieve pain and depression by the same neural mechanisms.

This MiniReview will go through the pharmacology of antidepressants in relation to neuropathic pain mechanisms and present the evidence for the use of antidepressants and their efficacy in different neuropathic pain conditions. This has also been the topic of previous reviews and book chapters (McQuay et al. 1996; Sindrup 1997; Sindrup & Jensen 2001; Sindrup 2003), but our understanding of how these drugs work in neuropathic pain increases and new trials on old and new antidepressants have been published.

Pharmacology of antidepressants and pain mechanisms

  1. Top of page
  2. Abstract
  3. Pharmacology of antidepressants and pain mechanisms
  4. Antidepressants in neuropathic pain
  5. Discussion
  6. References

The current knowledge of the pharmacological actions of tricyclic antidepressants (Baldessarini 2001) has slowly evolved through their over 40-year history. The unique ability of these drugs to inhibit the presynaptic reuptake of the monoamines serotonin and noradrenaline was elucidated many years ago as was their postsynaptic α-adrenergic, H1-histaminergic and muscarine cholinergic receptor-blocking effects. The tricyclic antidepressants have no effect on dopamine reuptake but may have some indirect dopaminergic action by the adrenergic effect and desensitization of dopamine D2 receptors. The classical tricyclic antidepressants differ in their effect on monoamine reuptake. Amitriptyline, imipramine and clomipramine cause a balanced inhibition of serotonin and noradrenaline reuptake in vivo. The serotonin reuptake inhibition is exerted by the compounds themselves whereas the noradrenaline reuptake inhibition comes from their respective metabolites nortriptyline, desipramine and desmethylclomipramine. Nortriptyline and desipramine are marketed as drugs (tricyclic antidepressants) themselves and together with the tricyclic antidepressant maprotiline, they are mainly causing inhibition of noradrenaline reuptake. These drugs have a similar action as the other tricyclic antidepressants at postsynaptic receptors and probably also act like them with respect to the other actions described below.

Binding to opioid receptors has been reported for tricyclic antidepressants, but their binding affinity is probably too low to be relevant in humans at therapeutic drug concentrations (Hall & Ögren 1981). Later evidence for an N-methyl-D-aspartate-receptor antagonist-like effect (Reynolds & Miller 1988; Cai & McCaslin 1992; McCaslin et al. 1992; Watanabe et al. 1993) which was paralleled by experimental analgesia (Eisenach & Gebhart 1995) was presented. The importance of this receptor interaction at therapeutic drug concentrations is debatable. The most recent discovery in the history of tricyclic antidepressant pharmacology is interaction with sodium channels. It has repeatedly been shown in experiments in vitro that tricyclic antidepressants block sodium channels and this effect is also seen in neuronal tissue (Ishii Y & Sumi 1992; Deffois et al. 1996; Pancrazio et al. 1998). Apparently, tricyclic antidepressants bind to the local anaesthetic receptor and cause blockade of the open and inactivated channel state at therapeutic drug concentrations (Wang et al. 2004). Tricyclic antidepressants seem also to block voltage-dependent calcium channels (Lavoie et al. 1990, Shimizu et al. 1992).

The selective serotonin reuptake inhibitors (SSRIs) are non-tricyclic drugs which are characterised by causing inhibition of serotonin reuptake without action on noradrenaline reuptake (Baldessarini 2001). This drug class is represented by fluoxetine, paroxetine, citalopram, escitalopram, sertraline and fluvoxamine. These drugs are also selective in the sense that they do not block postsynaptic receptors, or at least such actions have not yet been reported. Fluoxetine has been reported to block sodium channels, but apparently the blockade is different than the sodium channel blockade of tricyclic antidepressants (Deffois et al. 1996; Pancrazio et al. 1998). Sodium channel blockade has not been reported for the other SSRIs. There are some minor interdrug differences in pharmacology between the SSRIs, but they are probably not important.

Bupropion, another second generation non-tricyclic antidepressant antidepressant, is a noradrenaline and dopamine reuptake inhibitor without postsynaptic effects (Baldessarini 2001). Ion channel blockade has not been reported for this drug.

Serotonin noradrenaline reuptake inhibitors (SNRIs) such as venlafaxine, milnacipran and duloxetine cause a balanced inhibition of serotonin and noradrenaline (Baldessarini 2001). These drugs are sometimes called balanced inhibitors of serotonin and noradrenaline. For venlafaxine, the balance in vivo depends on the drug dose or concentration. Venlafaxine is a serotonin and a weak noradrenaline reuptake inhibitor, but with increasing drug doses noradrenaline reuptake inhibition will increase mainly due to increasing concentrations of the metabolite R-O-desmethylvenlafaxine (Muth et al. 1991). Venlafaxine has no postsynaptic effects but it blocks sodium channels (Khalifa et al. 1999) although the characteristics of the blockade is different from that of tricyclic antidepressants. Duloxetine is itself a potent balanced inhibitor of serotonin and noradrenaline reuptake with no significant effect on a range of postsynaptic receptors or sodium channels (Wong & Bymaster 2002). The balanced monoamine reuptake inhibition has also been shown in vivo (Wong & Bymaster 2002).

The basic pharmacology of different antidepressants is summarised in table 1.

Table 1.  Pharmacological profile of antidepressant drugs tried in neuropathic pain.
Amitriptyline Imipramine ClomipramineNortriptyline Desipramine MaprotilineVenlafaxine DuloxetineBupropionFluoxetine Paroxetine Citalopram
  1. TCA=tricyclic antidepressants, SNRI=serotonin noradrenaline reuptake inhibitors, DNRI=dopamine noradrenaline reuptake inhibitors, SSRI=selective serotonin reuptake inhibitors. +: action present. (+): action weak. −: action not present. ?: not known.

Reuptake inhibitionSerotonin+−/(+)++
Receptor blockadeα-Adrenergic++
Musc. cholinergic++
Ion channel blockadeSodium++(+)/−?(+)/−/?

Several of the pharmacological actions can be linked to mechanisms of neuropathic pain (Sindrup & Jensen 1999; Woolf & Mannion 1999; Jensen et al. 2001) and endogenous pain modulation.

Serotonin and noradrenaline are together with endogenous opioids and γ-aminobuturic acid neurotransmitters in the network of neurones which from centres in the brain and brainstem modulate the activity in the nociceptive pathway at the dorsal horn of the spinal cord and more rostrally in the central nervous system. Presynaptic reuptake inhibition of the monoamines serotonin and noradrenaline by antidepressants will increase the levels of these amines in the synaptic clefts and this is expected to enhance the pain suppression induced by this system. In neuropathic pain, there may be a disinhibition due to degeneration of the pain-modulating system. The degeneration could be a result of excessive stimulation of the system by excitatory amino acids released e.g. by barrages of spontaneous activity in nociceptive neurones entering the dorsal horn and by increased activity in the rest of the pain pathway.

Neuronal hyperexcitability caused by a series of signalling substances appears to play a major role in neuropathic pain. NMDA receptor-operated ion channels on second order neurones in the dorsal horn of the spinal cord represent one such family of excitatory mechanisms which render neurones hyperexcitable to stimulation. This hyperexcitability causing both spontaneous and stimulus evoked pain may be blocked by the NMDA receptor antagonist-like effect of tricyclic antidepressants, although this effect may not be fully present at therapeutic drug concentrations (see above).

In the peripheral nervous system, spontaneous activity caused by excessive expression of sodium channels may lead both to spontaneous pain and hyperexcitability. Apparently, tricyclic antidepressants have the potential to block sodium channels and thus interfere with these sodium channels in the diseased peripheral nervous system. It is now clear that sodium channels in the central nervous system are involved in neuropathic pain, since the unspecific sodium channel blocker lidocaine relieves central pain (Attal et al. 2000; Finnerup et al., 2005).

The possible calcium channel blocking effect of tricyclic antidepressants may also be important. The anticonvulsant drug gabapentin which clearly relieves neuropathic pain is now believed to act by blocking calcium channels through its binding to the α2-δ-subunit of the channel complex. Binding to the α2-δ-subunit of calcium channels reduces presynaptic release of neurotransmitters into the synaptic cleft otherwise induced by action potentials arriving at the nerve terminal. It is of course speculative if the apparent calcium channel blocking effect of tricyclic antidepressants will do the same.

It is clear that several of the actions of tricyclic antidepressants and the specific actions of the more selectively acting antidepressants have the potential to interfere with neuropathic pain mechanisms and thus relieve neuropathic pain. Fig. 1 gives an overview of the mechanisms and sites of action of tricyclic antidepressants in pain treatment.


Figure 1. Some suggested mechanisms and sites of action of tricyclic antidepressants (TCA) in neuropathic pain on peripheral nerves, in the dorsal horn of the spinal cord and at supraspinal levels. NA=noradrenaline, 5-HT=serotonin, DOPA=dopamine, NMDA=N-methyl-D-aspartate.

Download figure to PowerPoint

Antidepressants in neuropathic pain

  1. Top of page
  2. Abstract
  3. Pharmacology of antidepressants and pain mechanisms
  4. Antidepressants in neuropathic pain
  5. Discussion
  6. References


Randomised, controlled trials are regarded as the only legitimate evidence of the effect of antidepressants in neuropathic pain. The quality of controlled trials of drug effects can be evaluated by a standard instrument (Jadad et al. 1996). Scores are from 2 to 5 with points given for randomisation, double-blinding, the description of these and account of patients and drop-outs. Table 2 lists all studies on antidepressants in neuropathic pain as identified by a MEDLINE search and personal archives of the authors.

Table 2.  Randomised, double-blind, placebo-controlled trials of antidepressants in neuropathic pain.
Active drug Daily drug doseStudy Quality ratingDesign Patient nos.Outcome>50% pain reliefaNNT (95% CI)Drop outs side effectsbNNH (95%CI)
  • a

    NNT=Numbers Needed to Treat to obtain one patient with more than 50% pain relief.

  • b

    NNH=Number Needed to Harm, i.e. number of patients to be treated for one drop out due to side effects.

  • c

    Additional data provided by author.

  • d

    d Study includes questionable neuropathic pain conditions.

Central post-stroke pain
Amitriptyline, 75 mgLeijon & Boivie (1989), 4Cross-over, 15Ami>pla10/151/141.7 (1.2–3.1)0/150/15
Spinal cord injury pain
Amitriptyline,  average 50 mgCardenas et al. (2002), 4Parallel, 84Ami=pla8/442/407.6 (3.8–∞)  7/442/409.2 (4.2–∞)
Postherpetic neuralgia
Amitriptyline,  average 73 mgWatson et al. (1982), 4Cross-over, 24Ami>pla16/241/241.6 (1.2–2.4)1/240/2424 (8.0–∞)
Amitriptyline,  average 65 mgMax et al. (1988), 3Cross-over, 34Ami>pla15/345/254.1 (2.1–82.1)5/353/3015 (4.2–∞)
Desipramine,  average 167 mgKishore-Kumar et al. (1990), 3Cross-over, 19Desi>pla12/192/191.9 (1.3–3.7)5/233/2113.4 (3.3–∞)
Nortriptyline,  mean 89 mg Desipramine,  mean 63 mgcRaja et al. (2002), 5Cross-over, 56TCA>Pla18/56a4/57a4.0 (2.6–8.9)a7/591/579.9 (5.3–84.6)
Painful polyneuropathy
Imipramine, 100 mgKvinesdal et al. (1984), 4Cross-over, 12Imi>pla7/120/121.7 (1.2–3.3)1/130/1313 (4.5–∞)
Nortriptyline, 30 mgGomez-Perez et al. (1985), 4Cross-over, 18Nor>pla16/181/181.2 (1.0–1.5)0/180/18
Amitriptyline,  average 90 mgMax et al. (1987), 4Cross-over, 29Ami>pla15/291/292.1 (1.5–3.5)3/322/3134.2 (6.2–∞)
Imipramine,  average 200 mgcSindrup et al. (1990a), 4Cross-over, 20Imi>pla17/193/201.3 (1.1–1.9)7/290/204.1 (2.5–11.7)
Clomipramine, 75 mgcSindrup et al. (1990b), 4Cross-over, 19Clo>pla10/191/192.1 (1.4–4.4)3/240/208 (3.9–∞)
Desipramine, 200 mgcSindrup et al. (1990b), 4Cross-over, 19Des>pla7/191/193.2 (1.8–13.0)3/230/207.7 (3.7–∞)
Desipramine,  average 201 mgMax et al. (1991), 3Cross-over, 20Des>pla11/202/202.2 (1.4–5.1)2/241/242.4 (5.6–∞)
Imipramine, 150 mgcSindrup et al. (1992a), 4Cross-over, 18Imi>pla8/182/183.0 (1.7–16.2)1/220/2022 (7.6–∞)
Amitriptyline, 75 mgVrethem et al. (1997), 4Cross-over, 33Ami>pla22/338/332.4 (1.6–4.8)3/360/3312 (5.8–∞)
Maprotiline, 75 mgVrethem et al. (1997), 4Cross-over, 33Map>pla14/338/335.5 (2.5–∞)  1/340/3334 (11.6–∞)
Imipramine, 150 mgSindrup et al. (2003), 5Cross-over, 29Imi>pla14/292/292.4 (1.6–4.8)0/372/40
Paroxetine, 40 mgcSindrup et al. (1990a), 4Cross-over, 20Par>pla10/203/202.9 (1.6–12.4)0/200/20
Fluoxetine, 40 mgMax et al. (1992), 3Cross-over, 46Flu=pla22/4619/4615.3 (3.7–∞)3/542/5454.0 (10.2–∞)
Citalopram, 40 mgSindrup et al. (1992a), 4Cross-over, 15Cit>pla3/151/157.5 (2.7–∞)2/180/189 (3.9–∞)
Venlafaxine, 225 mgSindrup et al. (2003), 5Cross-over, 30Ven>pla8/302/295.1 (2.6–68.8)4/402/4020 (6.1–∞)
Venlafaxine,  75–225 mgRowbotham et al. (2004), 4Parallel, 244Ven>pla78/16327/816.9 (3.7–58.6)14/1633/8120.5 (9.2–∞)
St. John's WortSindrup et al. (2000), 5Cross-over, 47SJW=pla9/472/476.7 (3.6–44.4)1/501/521300 (18.3–∞)
Postmastectomy pain
Amitriptyline, 100 mgcKalso et al. (1995), 3Cross-over, 15Ami>pla8/152/152.5 (1.4–10.6)4/200/205 (2.7–40.5)
Venlafaxine,  37.5 – 75 mgTasmuth et al. (2002), 4Cross-over, 13Ven>pla11/13NANA1/150/1315 (5.2–∞)
Amitriptyline,  25–100 mgKieburtz et al. (1998), 5Parallel, 98Ami=pla23/4624/5050 (4–∞)3/461/5022 (7.9–∞)
Amitriptyline,  25–75 mg Mixed patientsShlay et al. (1998), 4Parallel, 110Ami=pla27/5824/50NANANA
Clomipramine,  25–100 mgPanerai et al. (1990), 3Cross-over, 24Clo>plaNANANA0/271/27
Nortriptyline,  25–100 mgPanerai et al. (1990), 3Cross-over, 24Nor>PlaNANANA2/271/2727 (6.3–∞)
dBupropion, 300 mgSemenchuck et al. (2001), 3Cross-over, 41Bup>pla30/414/411.6 (1.3–2.1)2/411/4045.1 (9.5–∞)

There is evidence that tricyclic antidepressants will relieve peripheral neuropathic pain conditions such as painful polyneuropathy of diabetic and non-diabetic aetiology, postherpetic neuralgia, postmastectomy pain syndrome and groups of patients with different peripheral neuropathic pain conditions. The evidence is high in postherpetic neuralgia with 4 trials including in total 134 patients and in particular in diabetic neuropathy with 10 trials including in total 198 patients for both disease entities in cross-over designs. There is evidence for effect of both noradrenergic (maprotiline, desipramine, nortiptyline) and balanced serotonergic and noradrenergic (amitriptyline, imipramine, clomipramine) tricyclic antidepressants. Apparently, there is no effect of tricyclic antidepressants in HIV neuropathy, although tricyclic antidepressant doses in these trials were at the lower end of the range. These data may not be representative for the effect in peripheral neuropathic pain due to the complexity of the disease which is also known to cause affection of the central nervous system.

One small trial indicated that central post-stroke pain will also be relieved by tricyclic antidepressants, whereas one large trial failed to find an effect of tricyclic antidepressant in spinal cord injury pain. However, the study on spinal cord injury pain (Cardenas et al. 2002) probably used a too low dose of amitriptyline (average 50 mg/day). The drug concentrations reported were below the level which was found to be associated with pain relief in central post-stroke pain (Leijon & Boivie 1989).

It is controversial whether SSRIs relieve neuropathic pain. Two small trials (n=20 and n=15) indicated a small but significant effect of paroxetine and citalopram in painful diabetic neuropathy (Sindrup et al. 1990a, b & c & 1992a & b), whereas fluoxetine in a large trial (n=46) apparently had no effect on this neuropathic pain condition (Max et al. 1992). The very long half-life of active fluoxetine metabolites may to some extent have hampered the latter trial. There are no data on SSRIs in any other neuropathic pain conditions.

The noradrenergic and dopaminergic drug bupropion relieved neuropathic pain in a group of patients with mixed aetiology in a single trial (Semenchuck et al. 2001).

Recently, a number of trials on SNRIs in neuropathic pain have been published. One trial showed that venlafaxine relieves painful diabetic neuropathy (Rowbotham et al. 2004) and another trial reported relief in painful polyneuropathy of different aetiologies (Sindrup et al. 2003). Promising data showing a clearcut pain relief have also been reported in abstract form for the SNRI duloxetine (Detke et al. 2003).


In clinical practice, drug efficacy is of major importance. Statistically significant pain-relieving effects of drugs reported from different clinical trials may differ tremendously with respect to actual efficacy. One way to elucidate drug efficacy in a clinically relevant fashion is to calculate numbers needed to treat (NNT) to obtain e.g. one patient with at least 50% pain relief (Cook & Sackett 1995; McQuay et al. 1996). In this way, different types of ratings of pain or pain relief, or other measures of effect are translated into a common currency which is considered clinically relevant and can be pooled and compared between different trials, drugs and conditions to give a rough estimate of overall efficacy. NNT is calculated as the reciprocal values of the difference in response rate (more than 50% pain relief) on active and placebo (corresponds to the reciprocal value of an absolute risk reduction), i.e. NNT is corrected for placebo response and can only be calculated from placebo-controlled trials. It can be argued that NNT for 30% pain relief is also clinically important (Farrer 2001) but we have chosen to use NNT for 50% pain relief (hereafter NNT). Comparison of the results of the new large scale parallel-group trials with the old cross-over studies must also be done cautiously, since in the latter, intention to treat analyses cannot be used. This will tend to overestimate the efficacy in the cross-over versus the parallel-group trial. There are also other problems with NNT, e.g. patient selection criteria, dosing and cut-off point for pain relief may differ between studies, and the ideal would of course be to have head-to-head comparisons in large studies.

NNT values for different antidepressants in different neuropathic pain conditions as calculated from data in table 2 are shown in fig. 2 and 3.


Figure 2. Numbers Needed to Treat (NNT) with 95% confidence intervals to obtain one patient with more than 50% pain relief for tricyclic antidepressants (TCA) in peripheral and central neuropathic pain. The size of the NNT points corresponds to the number of patients exposed to active drug plus the number exposed to placebo in the underlying clinical trials. Data from HIV-neuropathy not included.

Download figure to PowerPoint


Figure 3. Numbers Needed to Treat (NNT) with 95% confidence intervals to obtain one patient with more than 50% pain relief for antidepressants (AD) in peripheral neuropathic pain. The size of the NNT points corresponds to the number of patients exposed to active drug plus the number exposed to placebo in the underlying clinical trials. Data from HIV-neuropathy not included. TCA=tricyclic antidepressants, NA=noradrenaline, 5-HT=serotonin, SNRI=serotonin noradrenaline reuptake inhibitors (data from low dose excluded), DNRI=dopamine noradrenaline reuptake inhibitors, SSRI=selective serotonin reuptake inhibitors.

Download figure to PowerPoint

In peripheral neuropathic pain excluding HIV neuropathy, the NNT for tricyclic antidepressants is 2.3 (2.1–2.7) with no major difference across the different disease entities. For tricyclic antidepressants with balanced reuptake inhibition of noradrenaline and serotonin, the NNT is 2.2 (1.9–2.6) and the relatively noradrenergic tricyclic antidepressants have an NNT of 2.5 (2.1–3.3). SSRIs have in painful diabetic polyneuropathy an NNT of 6.8 (3.4–441) and the superiority of tricyclic antidepressants is also found in head to head comparison (Sindrup et al. 1990a & b). The NNT for the SNRI venlafaxine in painful polyneuropathy is 5.5 (3.4–13.5). However, the largest study on venlafaxine included two dose levels and only the highest dose level (150–225 mg/day) relieved pain, and if the data from the low dose group (75 mg/day) are excluded, we end up with an NNT of 4.6 (2.9–10.6). The head-to-head comparison of the tricyclic antidepressant imipramine and venlafaxine found a trend of better effect of imipramine but it failed to reach statistical significance (Sindrup et al. 2003). None of the preliminary reports on duloxetine include data to be used in NNT calculation. Lack of sufficient data for this purpose is also a problem in venlafaxine in post-mastectomy pain and clomipramine/nortriptyline in mixed patients with peripheral neuropathic pain (table 2). Bupropion in mixed aetiology patients with peripheral neuropathic pain has an NNT of 1.6 (1.3–2.1), but this is based on only one trial with 41 patients and the NNT must be regarded as a point estimate.

The positive results with amitriptyline in central post-stroke pain (NNT 1.7, 1.2–3.1) and the negative results with the same drug in spinal cord injury pain add up to an NNT of 4.0 (2.6–8.5) for tricyclic antidepressants in central neuropathic pain. Again, the amitriptyline value in central post-stroke pain is only a point estimate relying on one trial with 15 patients.

Side effects.

Tricyclic antidepressants cannot be used in patients with cardiac conduction disturbances, cardiac incompensation and epilepsy. Side effects including dry mouth, sweating, dizziness, orthostatic hypotension, fatigue, constipation and problems with micturition are often bothersome and will lead to discontinuation of tricyclic antidepressants in a number of patients. The SSRIs and SNRI are better tolerated but drugs from these groups also cause side effects. The SSRIs may induce nausea, vomiting and dyspepsia and the same type of side effects are seen with the SNRIs. Venlafaxine may also lead to rising in blood pressure. Bupropion may induce seizures and epilepsy is, as for tricyclic antidepressants, a contraindication towards this compound.

Some of the side effects deserve some attention. The orthostatic hypotension induced by tricyclic antidepressants may cause falls and hip fractures, which has been found in population-based studies (Ray et al. 1991). In a recent study it was shown that both tricyclic antidepressants and SSRIs are associated with increased risk of hip fractures (Hubbard at al. 2003). The most worrisome side effect of tricyclic antidepressants is sudden death supposedly related to cardiac arrhythmia. A retrospective cohort study found that SSRIs and tricyclic antidepressants in doses less than 100 mg/day did not increase the risk of sudden cardiac death, whereas higher doses of tricyclic antidepressants were associated with 2–3 times increased relative risk (Ray et al. 2004). Pre-excisting risk factors such as cardiovascular disease and age did not increase sudden cardiac death for the tricyclic antidepressant doses less than 100 mg/day. These side effects of tricyclic antidepressants may favour the use of SNRIs.

Drop-outs due to side effects during clinical trials with antidepressants in neuropathic pain are detailed in table 2. From these data, numbers needed to harm (NNH) can be calculated as the reciprocal value of the difference in drop out rates on active and placebo treatment which provides a measure of tolerability of the drugs. The overall NNHs are 13.6 (9.8–22.5) for tricyclic antidepressants, 19 (8.1-∞) for SSRIs and 21.5 (11.2–270) for SNRIs and bupropion together. The somewhat better tolerability of SSRIs and SNRIs than of tricyclic antidepressants is reflected in these figures. Treatment discontinuation may be more frequent in daily clinical practice than in the setting of a clinical trial.


Dosing of tricyclic antidepressants deserves special attention due to the large interindividual variation in pharmacokinetics with a 30 times difference in steady state serum concentrations on 100 mg imipramine (Reisby et al. 1977) and a similar variability for the other tricyclic antidepressants (Baldessarini 2001). The variability is caused by a genetic polymorphism of the drug metabolising enzyme CYP2D6 (Brøsen & Gram 1989). Several studies have indicated a relation between the pain-relieving effect and the serum drug concentration (Sindrup et al. 1990c; Leijon & Boivie 1989; Rasmussen et al. 2004). Thus, standard dosing may cause toxicity in some patients due to the relatively low therapeutic index and leave others at subtherapeutic drug levels. Dosing according to effect and side effect is not expected to be successful, since side effects are often present even at subtherapeutic concentrations and not all patients will obtain a pain-relieving effect at all. Dosing guided by measurements of serum drug concentrations is suggested to improve therapeutic outcome, i.e. a start dose of 50 mg/day and dose adjustment according to a drug level measured after 2–3 weeks on the start dose.

Some of the SSRIs are also metabolised by CYP2D6, but the pharmacokinetic variability is less pronounced (Rasmussen & Brøsen 2000) and the therapeutic index is probably higher. Fluoxetine and paroxetine are potent inhibitors of CYP2D6 and fluoxetine has a metabolite with a very long half life (Rasmussen & Brøsen 2001). Therefore, citalopram may be preferred among the SSRIs tried in neuropathic pain to avoid drug-drug interactions. Dosing according to effect and side effects is feasible and this also holds for the SNRIs. The studies on venlafaxine showed that a dose of 75 mg/day was ineffective whereas 225 mg/day relieved pain (Rowbotham et al. 2004), and low serum drug levels were associated with non-response (Sindrup et al. 2003). This fits with the experimental data showing that noradrenaline reuptake inhibition is first present at higher drug concentration and the noradrenergic effect is expected to be important for the analgesic effect. The preliminary data on duloxetine indicate that 60–120 mg/day provides pain relief whereas 20 mg/day is ineffective.


  1. Top of page
  2. Abstract
  3. Pharmacology of antidepressants and pain mechanisms
  4. Antidepressants in neuropathic pain
  5. Discussion
  6. References

The effect of tricyclic antidepressants in peripheral neuropathic pain is based on solid evidence with numerous trials on a range of different compounds in a number of important neuropathic pain conditions. The data on SSRIs are limited and equivocal, and under all circumstances it is indicated that their efficacy at best is moderate. The amount of data on SNRIs is increasing and suggests that drugs from this class could become a real alternative to the tricyclic antidepressants. The efficacy seems not to be quite as good as for the tricyclic antidepressants, but the SNRIs do not have any obvious contraindications and they are better tolerated. They will also be easier to use in clinical practice, since dosing according to effect and tolerability is feasible. The noradrenergic and dopaminergic drug bupropion is promising, but further studies are needed to support the current evidence obtained from only one trial.

The data on antidepressants in central neuropathic pain are limited, and it seems not to be justified to anticipate that efficacy of drugs in peripheral neuropathic pain is reflected in similar efficacy in central neuropathic pain. Although pain mechanisms may overlap between these groups, some of the basic pain mechanisms are supposed to be different. Therefore, we are at thin ice when these drugs are prescribed for disease entities belonging to this group. On the other hand, there are no other treatments with sufficient evidence at the present time. Studies have shown small treatment effects of lamotrigine in central post-stroke pain (Vestergaard et al. 2001) and of a cannabinoid in pain in multiple sclerosis (Svendsen et al. 2004). The central pain conditions are regarded to be relatively treatment-resistant. This may explain the sparse data in spite of a frequent clinical problem with e.g. central post-stroke pain experienced by as many as one in every 13 stroke sufferers (Andersen et al. 1995).

How do antidepressants work in neuropathic pain?

Enhancement of pain suppression via monoaminergic links in diffuse noxious inhibitory control induced by inhibition of presynaptic reuptake of serotonin and noradrenaline was suggested as a mechanism by which antidepressants could relieve neuropathic pain many years ago (Max et al. 1987; Sindrup et al. 1989). Taking all the present data into account, it is supposed that both of these reuptake mechanisms play a role, since tricyclic antidepressants with balanced reuptake inhibition tend to work better than noradrenergic tricyclic antidepressants and SNRIs are more efficacious than SSRIs. Further, it can be deduced that the noradrenergic mechanism appears to be most important. However, the apparent better effect of tricyclic antidepressants than of SNRIs indicates that other mechanisms contribute to the effect of tricyclic antidepressants and it is obvious to suggest the blockade of NMDA receptors and sodium channels as the candidate mechanisms. Therefore, the mechanism of action of tricyclic antidepressants in neuropathic pain is probably multimodal with contribution of monoamine reuptake inhibition and blockade of NMDA receptors and sodium channels. Bupropion provided substantial pain relief in a single study (Semenchuck et al. 2001) and this raises the possibility that a dopaminergic effect could enhance the pain-relieving effect obtained with serotonergic and noradrenergic mechanisms. This deserves further evaluation in clinical trials.

How does the efficacy compare with other treatments?

Although neuropathic pain is considered difficult to treat, several treatment options with reasonable efficacy are available at the present time. The anticonvulsant drug gabapentin has been extensively studied in peripheral neuropathic pain and relieves painful diabetic neuropathy with NNT 4.3 (2.8–8.6) and postherpetic neuralgia with NNT 4.3 (3.3–6.1) (Finnerup, personal communication). A limited amount of data shows that lamotrigine, another anticonvulsant drug, relieves peripheral neuropathic pain with NNT 4.0 (2.1–4.2) (Finnerup, personal communication). In head-to-head comparison gabapentin and amitriptyline had similar analgesic effect but the trial including 28 patients with painful diabetic neuropathy clearly was under powered to detect even clinically relevant differences in pain relief (Morello et al. 2000). The opioid drugs oxycodone and tramadol also relieved peripheral neuropathic pain; painful polyneuropathy with NNT 2.6 (1.7–6.0) and 3.5 (2.4–6.4) and postherpetic neuralgia with NNT 2.5 (1.7–5.1) and 4.8 (2.6–26.9) (Finnerup, personal communication). In postherpetic neuralgia a head-to-head comparison showed a trend of better effect of opioids than of tricyclic antidepressants (Raja et al. 2002), which contrasts the opposite rank order indicated by the NNT method. Probably there are no major differences in efficacy although higher doses of opioids could provide superior pain relief, whereas tricyclic antidepressant-dosing is probably already relatively high.

There are also data on some other drugs, e.g. other anticonvulsants and NMDA antagonists, but the evidence is limited due to either low efficacy or very limited data. Thus, in comparison with other drugs for neuropathic pain, tricyclic antidepressants are in the upper end of the efficacy range, SNRIs in the mid range and SSRIs in the lower end all as evaluated by the NNT method. Some of the other treatment options may be better tolerated at least as compared to tricyclic antidepressants, but these treatments as the antidepressants, will cause side effects in the majority of patients. The overall NNH based on drop outs due to side effects in neuropathic pain trials is 26.8 (13.7–698) for gabapentin, 23.0 (9.8–∞) for oxycodone and 9.0 (6.0–17.5) for tramadol (Finnerup, personal communication).

Combination with other treatments.

The effect of combining antidepressants with other drugs in neuropathic pain is largely unknown. Adding venlafaxine to gabapentin, if the latter provided insufficient pain relief in painful diabetic neuropathy, resulted in significant additional effect (Simpson 2001). When treatment combinations are used, it is suggested to employ combinations of drugs that do not overlap too much in pharmacological action. Thus, tricyclic antidepressants and venlafaxine could be combined with gabapentin and with opioids except tramadol which, besides the opioid effect, also interferes with the monoaminergic system. Combination of different antidepressant classes should be avoided as should be combination of any antidepressant with tramadol. In all, it must be kept in mind that combination treatments in most cases will not be evidence-based.

In conclusion, in peripheral neuropathic pain, antidepressants of the categories tricyclic antidepressants and SNRIs are fairly effective whereas SSRIs have a low efficacy. The efficacy of tricyclic antidepressants and SNRIs are either better or in the same range as other drugs used for this type of pain, i.e. gabapentin or the opioids oxycodone and tramadol. The side effects of tricyclic antidepressants, in particular the increased risk of sudden cardiac death with higher doses, may favour the choice of SNRIs or drugs from other classes. The usefulness of antidepressants in central neuropathic pain has not been finally settled.


  1. Top of page
  2. Abstract
  3. Pharmacology of antidepressants and pain mechanisms
  4. Antidepressants in neuropathic pain
  5. Discussion
  6. References
  • Andersen, G., K. Vestergaard, M. Ingemann-Nielsen & T.S. Jensen: Incidence of central poststroke pain. Pain 1995, 61, 187193.DOI: 10.1016/0304-3959(94)00144-4
  • Attal, N., V. Gaude, L. Brasseur, M. Dupuy, F. Guirimand, F. Parker & D. Bouhassira: Intravenous lidocaine in central pain: a double-blind, placebo-controlled, psychophysical study. Neurology 2000, 54, 56474.
  • Baldessarini, R. J.: Drugs for the treatment of psychiatric disorders. In: Goodman & Gilman's The Pharmacological Basis of Therapeutics, 10th. Eds.: J. G.Hardman, L. E.Limbird & A. GoodmanGilman. McGraw Hill, New York, USA, 2001, pp. 447483.
  • Brøsen, K. & L. F. Gram: Clinical significance of the sparteine/debrisoquine oxidation polymorphism. Eur. J. Clin. Pharmacol. 1989, 36, 537547.
  • Cai, Z. & P. P. McCaslin: Amitriptyline, cyproheptadiene and carbamazepine, in concentrations used therapeutically, reduce kainite- and N-methyl-D-aspartate-induced intracellular Ca2+ levels in neuromnal culture. Eur. J. Pharmacol. 1992, 219, 5357.
  • Cardenas, D. D., C. A. Warms, J. A. Turner, H. Marshall, M. M. Brooke & J. D. Loeser: Efficacy of amitriptyline for relief of pain in spinal cord injury: results of a randomized controlled trial. Pain 2002, 96, 365373.
  • Cook, R. J. & D. L. Sackett: The number needed to treat: a clinical useful measure of treatment effect. Brit. Med. J. 1995, 310, 452454.
  • Davis, J. L., S. B. Lewis, J. E. Gerich, R. A. Kaplan, T. A. Schultz & J. D. Wallin: Peripheral diabetic neuropathy treated with amitriptyline and fluphenazine. J. Amer. Med. Assoc. 1977, 238, 22912292.
  • Deffois, A., D. Fage & C. Carter: Inhibition of synaptosomal veratridine-induced sodium influx by antidepressants and neuroleptics used in chronic pain. Neurosci. Lett. 1996, 220, 117120.
  • Detke, M., D. Goldstein, Y. Lu, S. Iyengar & T. Lee: Efficacy of duloxetine in the treatment of pain associated with diabetic neuropathy. Diabetologia 2003, 46 (suppl 2), A315.
  • Eisenach, J. C. & G. F. Gebhart: Inthrathecal amitriptyline acts as an N-methyl-D-aspartate receptor antagonist in the presence of inflammatory hyperalgesia in rats. Anesthesiology 1995, 83, 10461054.
  • Farrer, J. T., J. P. Young Jr., L. LaMoreaux, J. L. Werth & R. M. Poole: Clinical importance of changes in chronic pain intensity measured on an 11-point numerical pain rating scale. Pain 2001, 94, 149158.
  • Finnerup, N. B., F. Biering-Sørensen, I. L. Johannsen, A. J. Terkelsen, G. I. Juhl, A. D. Kristensen, S. H. Sindrup, F. W. Bach & T. S. Jensen: Intravenous lidocaine relieves spinal cord injury pain: a randomised, controlled trial. Anesthesiology 2005, in press.
  • Gomez-Perez, F. J., R. A. Rull, H. Dies, G. Rodriguez-Rivera, J. Gonzlez-Barranco & O. Lozano-Castañeda: Nortriptyline and fluphenazine in the symptomatic treatment of diabetic neuropathy. A double-blind cross-over study. Pain 1985, 23, 395400.
  • Hall, H. & S.-O. Ögren: Effects of antidepressant drugs on different receptors in the brain. Eur. J. Pharmacol. 1981, 70, 393407.
  • Hubbard, R., P. Farrington, C. Smith, L. Smeeh & A. Tattersfield: Exposure to tricyclic and selective serotonin reuptake inhibitor antidepressants and the risk of hip fracture. Amer. J. Epidemiol. 2003, 158, 7784.
  • Ishii, Y. & T. Sumi: Amitriptyline inhibits striatal efflux of neurotransmitters via blockade of voltage-dependent Na+ channels. Eur. J. Pharmacol. 1992, 221, 377380.
  • Jadad, A. R., A. Moore, J. D. Carroll, C. Jenkinson, J. M. Reynolds, D. J Gavaghan & H. J. McQuay: Assessing the quality of reports of randomized clinical trials: Is blinding necessary. Controlled Clin. Trials. 1996, 17, 112.
  • Jensen, T. S., H. Gottrup, S. H. Sindrup & F. W. Bach: The clinical picture of neuropathic pain. Eur. J. Pharmacol. 2001, 429, 111.
  • Kalso, E., T. Tasmuth & P. J. Neuvonen: Amitriptyline effectively relieves neuropathic pain following treatment of breast cancer. Pain 1995, 64, 293302.
  • Khalifa, M., P. Daleau & J. Turgeon: Mechanism of sodium channel block by venlafaxine in guinea pig ventricular myocytes. J. Pharmacol. Exp. Therap. 1999, 291, 280284.
  • Kieburtz, K., D. Simpson, C. Yiannoutsos, M. B. Max, C. D. Hall, R. J. Ellis, C. M. Marra, R. McKendall, E. Singer, G. J. Dal Pan, D. B. Clifford, T. Tucker, B. Cohen & the AIDS Clinical Trial Group 242 Protocol Team: A randomized trial of amitriptyline and mexiletine for painful neuroapthy in HIV infection. Neurology 1998, 51, 16821688.
  • Kvinesdal, B., J. Molin, A. Frøland & L. G. Gram: Imipramine treatment of painful diabetic neuropathy. J. Amer. Med. Assoc. 1984, 251, 17271730.
  • Kishore-Kumar, R., M. B. Max, S. C. Schafer, A. M. Gaughan, B. Smoller, R. H. Gracely & R. Dubner: Desipramine relieves postherpetic neuralgia. Neurology 1990, 47, 305312.
  • Lavoie, P.-A., G. Beauchamp & R. Elie: Tricyclic antidepressants inhibit voltage-dependent calcium channels and Na+-Ca2+ exchange in rat brain cortex synaptosomes. Can. J. Physiol. Pharmacol. 1990, 68, 14141418.
  • Leijon, G & J. Boivie: Central post-stroke pain – a controlled trial of amitriptyline and carbamazepine. Pain 1989, 36, 2736.
  • Max, M. B., M. Culnane, S. C. Schafer, R. H. Gracely, D. J. Walther, B. Smoller & R. Dubner: Amitriptyline relieves diabetic neuropathy pain in patients with normal or depressed mood. Neurology 1987, 37, 589596.
  • Max, M. B., S. C. Schafer, M. Culnane, B. Smoller, R. Dubner & R. H. Gracely: Amitriptyline, but not lorazepam, relieves postherpetic neuralgia. Neurology 1988, 38, 14271432.
  • Max, M. B., R. Kishore-Kumar, S. C. Schafer, B. Meister, R. Gracely, B. Smoller & R. Dubner: Efficacy of desipramine in painful diabetic neuropathy: a placebo-controlled trial. Pain 1991, 45, 39.
  • Max, M. B., S. A. Lynch, J. Muir, S. E. Shoaf, B. Smoller & R. Dubner: Effects of desipramine, amitriptyline, and fluoxetine on pain in diabetic neuropathy. New Engl. J. Med. 1992, 326, 12501256.
  • McCaslin, P. P., X. Z. Yu, I. K. Ho & T. G. Smith: Amitriptyline prevents N-methyl-D-aspartate (NMDA)-induced toxicity, does prevent NMDA-induced elevations of extracellular glutamate, but augments kainite-induced elevations of glutamate. J. Neurochem. 1992, 59, 401405.
  • McQuay, H. J., M. Tramer, B. A. Nye, D. Carroll, P. Wiffen & A. Moore: A systematic review of antidepressants in neuropathic pain. Pain 1996, 68, 217227.
  • Morello, C. M., S. G. Leckbard, C. P. Stoner, D. F. Moorhouse & G. A. Sahagian: Randomised double-blind study comparing the efficacy of gabapentin with amitriptyline on diabetic peripheral pain. Arch. Int. Med. 2000, 160, 10401041.
  • Muth, E. A., J. A. Moyer, J. T. Haskin, et al.: Biochemical, neurophysiological and behavioral effects of WY-45,233 and other identified metabolites of the antidepressant venlafaxine. Drug Dev. Res. 1991, 23, 191193.
  • Pancrazio, J. J., G. L. Kamatchi, A. K. Roscoe & C. Lynch III: Inhibition of neuronal Na+ channels by antidepressant drugs. J. Pharmacol. Exp. Therap. 1998, 284, 208214.
  • Panerai, A. E., G. Monza, P. Movilia, M. Bianchi, B. M. Francucci & M. Tiengo: A randomized, within-patient, cross-over, placebo-controlled trial on the efficacy and tolerability of the tricyclic antidepressants chlorimipramine and nortriptyline in central pain. Acta. Neurol. Scand. 1990, 82, 3438.
  • Paolo, F., G. Darcourt & P. Corsa: Note préliminaire sur l'action de l'imipramine dans les états douloureux. Rev. Neurol. 1960, 2, 503504.
  • Raja, S. N., J. A. Haythornthwaite, M. Pappagallo, M. R. Clark, T. G. Travison, S. Sabeen, R. M. Royall & M. B. Max: Opioids versus antidepressants in postherpetic neuralgia. A randomized, placebo-controlled trial. Neurology 2002, 59, 10151021.
  • Rasmussen, B. B. & K. Brøsen: Is therapeutic drug monitoring a case for optimizing clinical outcome and avoiding interactions of the selective serotonin reuptake inhibitors? Therap. Drug Monit. 2001, 22, 143154.
  • Rasmussen, P. V., T. S. Jensen, S. H. Sindrup & F. W. Bach: TDM-based imipramine treatment in neuropathic pain. Therap. Drug Monit. 2004, 26, 352360.
  • Ray, W. A., M. R. Griffin & E. Malcolm: Cyclic antidepressants and the risk of hip fractures. Arch. Int. Med. 1991, 151, 754756.
  • Ray, W. A., S. Meredith, P. B. Thapa, K. Hall & K. Murray: Cyclic antidepressants and the risk of sudden cardiac death. Clin. Pharmacol. Therap. 2004, 75, 234241.
  • Reisby, N., L. F. Gram, P. Bech, A. Nagy, G. O. Petersen, J. Ortmann, I. Ibsen, S. J. Dencker, O. Jakobsen, O. Krautwald, I. Søndergaard & J. Christensen: Imipramine: Clinical effects and pharmacokinetic variability. Psychopharmacology 1977, 54, 263272.
  • Reynolds, I. J. & R. J. Miller: Tricyclic antidepressants block N-methyl-D-aspartate receptors: similarities to the action of zinc. Brit. J. Pharmacol. 1988, 95, 95102.
  • Robinson, L. R., J. M. Czerniecke, D. M. Ehde, T. Edwards, D. A. Judish, M. L. Goldberg, K. M. Campbell, D. G. Smith & M. P. Jensen: Trial of amitriptyline for relief of pain in amputees: results of a randomized controlled study. Arch. Phys. Med. Rehabil. 2004, 85, 16.
  • Rowbotham, M. C., V. Goli, N. R. Kunz & D. Lei: Venlafaxine extended release in the treatment of painful diabetic polyneuropathy: a double-blind, placebo-controlled study. Pain 2004, 110, 697706.
  • Semenchuck, M. R., S. Sherman & B. Davis: Double-blind, randomized trial of bupropion SR for the treatment of neuropathic pain. Neurology 2001, 57, 15831588.
  • Shimizu, M., A. Nishida & S. Yamawaki: Antidepressants inhibit spontaneous oscillations of intracellular Ca2+ concentration in rat cortical cultured neurons. Neurosci. Lett. 1992, 146, 101104.
  • Shlay, J. C., K. Chaloner, M. B. Max, B. Flaws, P. Reichelderfer, D. Wentworth, S. Hillman, B. Brizz, D. L. Cohn & the Terry Beirn Community Programs for Clinical Research on AIDS: Acupuncture and amitriptyline for pain due to HIV-related peripheral neuropathy. J. Amer. Med. Assoc. 1998, 280, 15901595.
  • Simpson, D. A.: Gabapentin and venlafaxine for the treatment of painful diabetic neuropathy. J. Clin. Neuromusc. Dis. 2001, 3, 5362.
  • Sindrup, S. H., B. Ejlertsen, A. Frøland, E. H. Sindrup, K. Brøsen & L. F. Gram: Imipramine treatment in diabetic neuropathy: relief of subjective symptoms without changes in peripheral and autonomic nerve function. Eur. J. Clin. Pharmacol. 1989, 37, 151153.
  • Sindrup, S. H., L. F. Gram, K. Brøsen, O. Eshøj & E. F. Mogensen: The selective serotonin reuptake inhibitor paroxetine is effective in the treatment of diabetic neuropathy symptoms. Pain 1990a, 42, 135144.
  • Sindrup, S. H., L. F. Gram, T. Skjold, E. Grodum, K. Brøsen & H. Beck-Nielsen: Clomipramine vs desipramine vs placebo in the treatment of diabetic neuropathy symptoms. A double-blind cross-over study. Brit. J. Clin. Pharmacol. 1990b, 30, 683691.
  • Sindrup, S. H., L. F. Gram, T. Skjold, A. Frøland & H. Beck-Nielsen: Concentration-response relationship in imipramine treatment of diabetic neuropathy symptoms. Clin. Pharmacol. Therap. 1990c, 47, 509516.
  • Sindrup, S. H., C. Tuxen, L. F. Gram, E. Grodum, T. Skjold, K. Brøsen & H. Beck-Nielsen: Lack of effect of mianserin on the symptoms of diabetic neuropathy. Eur. J. Clin. Pharmacol. 1992a, 43, 251255.
  • Sindrup, S. H., U. Bjerre, A. Dejgaard, K. Brøsen, T. Aaes-Jørgensen & L. F. Gram: The selective serotonin reuptake inhibitor citalopram relieves the symptoms of diabetic neuropathy. Clin. Pharmacol. Therap. 1992b, 52, 547552.
  • Sindrup, S. H.: Antidepressants as analgesics. In: Anesthesia. Biological foundations.. Eds.: T. L.Yaksh, C.Lynch, W. M.Zapol, M.Maze, J. F.Biebuyck, L. J.Saidman. Lippencott-Raven Publishers, Philadelphia, USA, 1997, pp. 987997.
  • Sindrup, S. H. & T. S. Jensen: Efficacy of pharmacological treatments of neuropathic pain: an update and effect related to mechanism of action. Pain 1999, 83, 389400.
  • Sindrup, S. H., C. Madsen, F. W. Bach, L. F. Gram & T. S. Jensen: St. John's wort has no effect on pain in polyneuropathy. Pain 2000, 91, 361365.
  • Sindrup, S. H. & T. S. Jensen: Antidepressants in the treatment of neuropathic pain. In: Neuropathic pain: Pathophysiology and treatment. Progress in pain Research and management. Eds.: P. T.Hansson, H. L.Fields, R. G.Hill & P.Marchettini. IASP Press, Seattle, USA, 2001, pp. 169183.
  • Sindrup, S. H.: Antidepressants and chronic pain. In: Clinical pain management. Chronic pain. Eds.: T. S.Jensen, P.Wilson, A. S.Rice. Arnold, London, UK, 2003, pp. 239249.
  • Sindrup, S. H., F. W. Bach, C. Madsen, L. F. Gram & T. S. Jensen: Venlafaxine versus imipramine in painful polyneuropathy. A randomized, controlled trial. Neurology 2003, 60, 12841289.
  • Svendsen, K. B., T. S. Jensen & F. W. Bach: Does the cannabinoid dronabinol reduce central pain in multiple sclerosis? Randomised double blind placebo controlled crossover trial. Brit. Med. J. 2004, 329, 257258.
  • Tasmuth, T., B. Härtel & E. Kalso: Venlafaxine in neuropathic pain following treatment of breast cancer. Eur. J. Pain. 2002, 6, 1724.
  • Vestergaard, K., G. Andersen , H. Gottrup, B. T. Kristensen & T. S. Jensen: Lamotrigine for central poststroke pain. A randomized controlled trial. Neurology 2001, 56, 184190.
  • Vrethem, M., J. Boivie, H. Arnqvist, H. Holmgren, T. Lindström & L.-H. Thorell: A comparison of amitriptyline and maprotiline in the treatment of painful polyneuropathy in diabetics and nondiabetics. Clin. J. Pain 1997, 13, 313323.
  • Wang, G. K., C. Russell & S.-Y. Wang: State-dependent block of voltage-gated Na+ channels by amitriptyline via local anesthetic receptor and its implication for neuropathic pain. Pain 2004, 110, 166174.
  • Watanabe, Y., H. Saito & K. Abe: Tricyclic antidepressants block NMDA receptor-mediated synaptic responses and induction of long term potentiation in rat hippocampal slices. Neuropharmacol. 1993, 32, 479486.
  • Watson, C. P., R. J. Evans, K. Reed, H. Merskey, L. Goldsmith & J. Warsh: Amitriptyline versus placebo in postherpetic neuralgia. Neurology 1982, 32, 671673.
  • Wong, D. T. & F. P. Bymaster: Dual serotonin and noradrenaline uptake inhibitor class of antidepressants – potential for greater efficacy or just hype? Prog. Drug. Res. 2002, 58, 169222.
  • Woolf, C. J. & R. J. Mannion: Neuropathic pain. Aetiology, symptoms, mechanisms and management. Lancet 1999, 353, 19591964.