[96] Benefits and harms of drugs for “neuropathic” pain

[96] Benefits and harms of drugs for “neuropathic” pain

96Chronic pain (at times presumed to be “neuropathic” in origin) is a common problem in clinical practice. It is now well recognized that the results of drug treatment are more often disappointing than not.1 Despite this, from 2005-2014 the number of British Columbians prescribed gabapentin increased 1.8 fold, pregabalin 17 fold, and duloxetine 3.6 fold (from 2008). Use of venlafaxine (mostly for depression/anxiety) has remained relatively stable.

Most gabapentin, pregabalin, and duloxetine use in B.C. is for chronic pain, driven partly by concern about problems with long-term opioid therapy. For the same reason, tricyclic antidepressants (amitriptyline, nortriptyline, imipramine, desipramine) are often prescribed for “neuropathic” pain.

In 2009 Therapeutics Letter 75 on gabapentin2 concluded: 

  • Gabapentin reduces neuropathic pain by < 1 point on a 0-10 point scale and benefits about 15% of carefully selected patients (NNT=6-8).
  • A similar proportion of people suffer harm (NNH=8).
  • A test of benefit/harm can be made after 1-2 days at a low dose (100-900 mg/day).
  • Benefit is unlikely to increase with higher doses or longer treatment.

This Letter updates information on gabapentin and critically appraises randomized clinical trials (RCT) assessing the benefits and harms of three other drugs promoted for neuropathic pain: pregabalin, duloxetine, and venlafaxine. It is based primarily on 4 Cochrane reviews.3-6 Like many systematic reviews, these either did not assess risk of bias, or did not fully reflect the implications of the risk of bias in their conclusions. We attempt to demonstrate how appreciation of the biases in RCTs can be incorporated into the conclusions of systematic reviews.


Although all pain metrics have limitations7 a 50% or greater reduction from a baseline pain score has been promoted as a more clinically relevant outcome for “neuropathic” pain because it correlates with improvements in comorbidity, function and quality of life.4 Using this outcome across all 4 Cochrane reviews, the mean number of people who must be treated for one to achieve a ≥ 50% reduction in pain (NNT) compared to placebo is about 6. This calculation is based on all doses that were statistically significantly superior to placebo. The evidence is weakest for venlafaxine, but even for gabapentin, pregabalin, and duloxetine, this NNT is likely very optimistic, as we judged the included RCTs to have a high risk of bias.

The greatest potential bias comes from the likelihood that patients and investigators were unblinded by observing drug adverse effects such as somnolence. Loss of blinding has been shown to be associated with a 68% exaggeration of relative benefits for subjective outcomes such as pain.8 In addition almost all RCTs included in the Cochrane reviews were funded by drug manufacturers. A separate Cochrane review demonstrated that industry funded studies lead to “more favourable results and conclusions” than non industry funded studies.9 Accounting for these biases, we suspect the real NNT for benefit from these drugs is at least 10.

An alternative measure of meaningful benefit is the patient’s reported global impression of change (PGIC). PGIC was not reported in any venlafaxine RCT3 and no meaningful difference was found for duloxetine.4 For gabapentin and pregabalin, the estimated NNT for “much or very much improved” PGIC ranges from 6-10.5,6 Like the ≥50% pain score reduction, this is probably overly optimistic.

The evidence of benefit for tricyclic antidepressants for neuropathic pain is weaker and it is not possible to estimate a meaningful NNT.10-13


Withdrawals due to adverse effects compared with placebo were higher with gabapentin, pregabalin, duloxetine and venlafaxine.3-6 Approximately 80% of people receiving these drugs experienced at least one adverse effect. The most common were somnolence, dizziness, and nausea. Anticholinergic effects, such as dry mouth and constipation, were common with duloxetine. The rate of adverse effects reported in Cochrane reviews almost certainly underestimate the real world rates because patients at higher risk (e.g. from impaired kidney function, alcohol use, or with other morbidities) are excluded from RCTs. Furthermore, official product monographs for these drugs report higher rates of adverse effects than do the Cochrane reviews.

The most common adverse effects reported for the tricyclic antidepressants were dry mouth, sedation and constipation.10-13 Likewise official monographs provide a better and higher estimate of the incidence of harms than the systematic reviews.

To whom do the Cochrane reviews apply?

Patients averaged 50 years of age, had moderate levels of neuropathic pain, and were free of medical conditions other than those being studied (diabetes, fibromyalgia, or post-herpetic neuralgia). RCTs varied with respect to allowed use of other analgesics from acetaminophen only to the use of multiple analgesics including opioids.

How soon is pain reduced?

In the majority of trials pain reduction compared with placebo was demonstrable within the first week. Very little additional pain reduction occurred after the second week.

Is there evidence that increasing dose improves response?

For gabapentin, pregabalin, duloxetine and venlafaxine, RCTs demonstrated little or no benefit from doses higher than the lowest dose that was superior to placebo.3-6

Clinical implications

Evidence from 8 Cochrane reviews should temper expectations regarding the likelihood and magnitude of pain relief from gabapentin, pregabalin, duloxetine, venlafaxine, amitriptyline, nortriptyline, imipramine or desipramine. When initiating a therapeutic trial with one of these drugs in a patient, it is reasonable to start at the lowest recommended dose and assess the patient for benefit and harm at 1 week. If benefit harm ratio is unacceptable, consider stopping the drug. If insufficient but partial pain relief is achieved, increase the dose and reassess within 1 week. If functionally meaningful benefit is still absent, stop the drug and try something else. For patients who achieve clinically meaningful analgesia, use the lowest individualized effective dose to minimize adverse effects. Reassess regularly (e.g. every 2 weeks), as most patients treated with placebo also improve over time.


  • The evidence base for drug treatment of neuropathic pain is weak, due to the small magnitude of clinically meaningful effects and the high risk of bias in the RCTs.
  • Probably less than 1 in 10 patients achieve a meaningful reduction in pain.
  • Most patients experience some adverse side effects like somnolence, dizziness, nausea, dry mouth and constipation.
  • To identify patients who respond, a therapeutic trial with early assessment is essential. Reassessment of drug utility is needed to detect people with spontaneous remission or placebo response.
  • Higher doses are unlikely to achieve greater pain reduction, but are more likely to cause harm.


  1. Moore A, Derry S, Eccleston C, Kalso E. Expect analgesic failure; pursue analgesic success. BMJ. 346:f2690, 2013.
  2. Therapeutics Initiative. Gabapentin for pain. New evidence from hidden data. Therapeutics Letter. 2009; 75:1-2.
  3. Gallagher HC, Gallagher RM, Butler M, et al. Venlafaxine for neuropathic pain in adults. Cochrane Database of Systematic Reviews, 2015 Issue 8. Art. No.: CD011091. DOI: 10.1002/14651858. CD011091.pub2.
  4. Lunn MPT, Hughes RAC, Wiffen PJ. Duloxetine for treating painful neuropathy, chronic pain or fibromyalgia. Cochrane Database of Systematic Reviews 2014, Issue 1. Art. No.: CD007115. DOI: 10.1002/14651858.CD007115.pub3.
  5. Moore RA, Straube S, Wiffen PJ, et al. Pregabalin for acute and chronic pain in adults. Cochrane Database of Systematic Reviews 2009, Issue 3. Art. NO.: CD007076. DOI: 10.1002/14651858.CD007076.pub2.
  6. Moore RA, Wiffen PJ, Derry S, et al. Gabapentin for chronic neuropathic pain and fibromyalgia in adults. Cochrane Database of Systematic Reviews, 2014, Issue 4. Art. No.: CD007938. DOI: 10.1002/14651858.CD007938.pub3.
  7. Ballantyne JC, Sullivan MD. Intensity of Chronic Pain – The Wrong Metric? N Engl J Med 2015;373(22): 2098-9.
  8. Hrobjartsson A, Thomsen AS, Emanuelsson F, et al. Observer bias in randomized clinical trials with measurement scale outcomes: a systematic review of trials with both blinded and nonblinded assessors. CMAJ 2013 Mar 5;185(4):E201-11.
  9. Lundh A, Sismondo S, Lexchin J, et al. Industry sponsorship and research outcome. Cochrane Database of Systematic Reviews 2012, Issue 12. Art. No.: MR000033. DOI:10.1002/14651858.MR000033.pub2.
  10. Moore RA, Derry S, Aldington D, et al. Amitriptyline for neuropathic pain in adults. Cochrane Database of Systematic Reviews 2015, Issue 7. Art. No.: CD008242. DOI: 10.1002/14651858.CD008242.pub3.
  11. Derry S, Wiffen PJ, Aldington D, Moore RA. Nortriptyline for neuropathic pain in adults. Cochrane Database of Systematic Reviews 2015, Issue 1. Art. No.: CD011209. DOI: 10.1002/14651858.CD011209.pub2.
  12. Hearn L, Moore RA, Derry S, et al. Desipramine for neuropathic pain in adults. Cochrane Database of Systematic Reviews 2014, Issue 9. Art. No.: CD011003. DOI: 10.1002/14651858.CD011003.pub2.
  13. Hearn L, Derry S, Phillips T, et al. Imipramine for neuropathic pain in adults. Cochrane Database of Systematic Reviews 2014, Issue 5. Art. No.: CD010769. DOI: 10.1002/14651858.CD010769.pub2.

The draft of this Therapeutics Letter was submitted for review to 60 experts and primary care physicians in order to correct any inaccuracies and to ensure that the information is concise and relevant to clinicians.
The Therapeutics Initiative is funded by the BC Ministry of Health through a grant to the University of BC. The Therapeutics Initiative provides evidence-based advice about drug therapy, and is not responsible for formulating or adjudicating provincial drug policies .
  • Sara Himelstein
    Posted at 14:49h, 19 January Reply

    Great to have a look at this info on pain management as this is very central to pre-dialysis work with patients.
    let me identify myself as a social worker. I write today because I see this area as being prescribed several meds for the same pain.
    My question is: are there any studies about docs prescribing because of patient demand and on docs needing to see themselves as doing something for patient?

    • Aaron Tejani
      Posted at 07:52h, 27 January Reply

      Thanks for the post Sara. There is literature on the pressure to prescribe but it is conflicting. Some studies indicate prescribers are pressured to prescribe yet others suggest patients want information and don’t necessarily want a drug. I will send you what I found.
      Aaron M Tejani
      Researcher, Therapeutics Initiative

  • Bob Paddock
    Posted at 18:23h, 19 January Reply

    My late wife Karen was taking this crap, Gabapentin (Neurontin) that is on this FDA warning list:

    From FDA warning: “Manufacturers of antiepileptic drugs (AEDs) or anticonvulsant drugs will update product labeling to include a warning about an increased risk of suicidal thoughts or actions and will develop a Medication Guide to help patients understand this risk.” – http://www.fda.gov/Drugs/DrugSafety/PostmarketDrugSafetyInformationforPatientsandProviders/ucm100190.htm

    Karen hung herself while on this medication to stop Chronic Pain from as she could no longer stand the excruciating headache caused by Intracranial Hypotension due to Cerebrospinal Fluid (CSF) Leaks. A condition that is more common that many think (for example Actor George Clooney had/has a CSF Leak and considered suicide), yet is so unknown that some doctors argue the condition does not even exist.

    Karen’s Journal is now required reading at Duke School of Medicine. http://www.kpaddock.com/book

    “Karen’s first-hand account of her illness gave an honest, heart-wrenching depiction of what it is like to live with debilitating pain day-to-day.” – Cover Story of The Derrick Newspaper Sept 8th 2014 – http://www.kpaddock.org

    Now tell me how good these drugs are again please…


    • Aaron Tejani
      Posted at 07:56h, 27 January Reply

      Dear Bob,
      Our sympathies to you and your family. Thank you for sharing your story. Change and awareness can only come from people speaking up.
      We encourage others to do the same.
      Aaron M Tejani
      Researcher, Therapeutics Initiative

  • Dr. John Kent
    Posted at 01:18h, 20 January Reply

    I’m a family physician practicing in Israel, apparently now the world’s capital for medical marijuana (it’s very hard to get a valid license, although one clear criterion is neuropathic pain). I’d appreciate a TI letter along the same lines as the above, detailing if possible the risks and benefits of this therapy. There might be some research bias if the studies were funded by drug lords instead of drug companies, which are only slightly richer but in many ways quite similar.

  • Aaron Tejani
    Posted at 07:58h, 27 January Reply

    Dear Dr. Kent,

    We appreciate the post. As yet we have not reviewed the literature on this topic. We encourage any of our other readers to post comments if they have information regarding this topic.

    Aaron M Tejani
    Researcher, Therapeutics Initiative

  • Eve Sample
    Posted at 14:54h, 12 February Reply

    Unfortunately options for neuropathic pain are limited . Opioid, first line, often are insufficient. Physicians should be encouraged to get a methadone license to prescribe for pain.
    About 50% of palliative patients I now see, have methadone prescribed for neuropathic pain with benefit- possibly because of its activity at other receptors besides the opioid ones.
    In my experience gabapentin causes drowsiness and clouded sensation with little or no pain relief.
    Studies and research needs to be supported by the Federal Government in this area.

    • Aaron Tejani
      Posted at 14:57h, 12 February Reply

      Dear Eve,

      Thanks for comment. I know you are a strong advocate for rational drug therapy and am glad to hear you continue to share your knowledge.
      Aaron M Tejani
      Researcher, Therapeutics Initiative

  • Aaron
    Posted at 18:07h, 28 January Reply

    Why is it that almost the entire medical field is ignorant to the fact that the dependence and withdrawal from drugs like gabapentin and lyrica can be just as bad (if not worse) than benzodiazapines.

    Gabapentin has destroyed many lives, robber people of years of their lives leaving many dissabled, not to mention the people who have taken their own lives because of the severity of withdrawal and the awful side effect, it has even caused perminent movement dissorders in some.

    This drug keeps new synapses from forming, and has shown to induce both parkinsons and altheimerz like dissorders, not to mention catatonia and psychosis.

    The industry did not learn the lesson from barbituates when benzos came along, how on earth could this happen a 3rd time? Gabapentinoids are dangerous drugs that throw your entire body chemistry out of balance as well as cause brain damage in as little as 21 days…


Post A Comment