Objective To assess the benefits and harms of pregabalin in the management of neuropathic pain.
Design Rapid review and meta-analysis of phase III, randomised, placebo-controlled trials.
Participants Adults aged 18 years and above with neuropathic pain defined according to the International Association for the Study of Pain criteria.
Interventions Pregabalin or placebo.
Primary and secondary outcome measures Our primary outcomes were pain (as measured using validated scales) and adverse events. Our secondary outcomes were sleep disturbance, quality of life, Patient Global Impression of Change, Clinician Global Impression scale, anxiety and depression scores, overall discontinuations and discontinuations because of adverse events.
Results We included 28 trials comprising 6087 participants. The neuropathic pain conditions studied were diabetic peripheral neuropathy, postherpetic neuralgia, herpes zoster, sciatica (radicular pain), poststroke pain and spinal cord injury-related pain. Patients who took pregabalin reported significant reductions in pain (numerical rating scale (NRS)) compared with placebo (standardised mean difference (SMD) −0.49 (95% CI −0.66 to −0.32, p<0.00001), very low quality evidence). Pregabalin significantly reduced sleep interference scores (NRS) compared with placebo (SMD −0.38 (95% CI −0.50 to −0.26, p<0.00001), moderate quality evidence. Pregabalin significantly increased the risk of adverse events compared with placebo (RR 1.33 (95% CI 1.23 to 1.44, p<0.00001, low quality evidence)). The risks of experiencing weight gain, somnolence, dizziness, peripheral oedema, fatigue, visual disturbances, ataxia, non-peripheral oedema, vertigo and euphoria were significantly increased with pregabalin. Pregabalin was significantly more likely than placebo to lead to discontinuation of the drug because of adverse events (RR 1.91 (95% CI 1.54 to 2.37, p<0.00001), low quality evidence).
Conclusion Pregabalin has beneficial effects on some symptoms of neuropathic pain. However, its use significantly increases the risk of a number of adverse events and discontinuation due to adverse events. The quality of the evidence from journal publications is low.
- systematic review
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.
Strengths and limitations of this study
We used the Cochrane criteria to assess the risk of bias.
This is the first review that rates the quality of the evidence for each outcome assessed.
The review may be prone to sampling bias, and we may have missed potentially eligible studies.
We did not assess the extent to which different doses of pregabalin influenced the outcomes.
Pregabalin is a gabapentinoid licenced for treatment of neurological disorders. It is one of the earlier drugs approved by the US Food and Drug Administration (2004) for the treatment of painful diabetic neuropathy and postherpetic neuralgia (PHN).1 Pregabalin is thought to exert its analgesic action through antagonistic activity at the voltage gated Ca2+ channels where it binds to the alpha-2-delta subunit.1 2
Prescriptions of pregabalin (and gabapentin) have markedly increased over the last few years. In the USA, prescriptions for pregabalin rose from 39 million in 2012 to 64 million in 2016 (annual prescription costs increased from approximately $2 billion to $4.4 billion over the same period). 3 In the UK, pregabalin use increased 350% over a 5-year period between 2008 and 2013.4 In England alone, there were over 6.2 million prescriptions of pregabalin across GP practices in 2017 costing about $440 million.5
Pregabalin is recommended as first-line pharmacological agent for management of neuropathic pain.6 There is, however, some evidence of increased mortality attributed to pregabalin in the UK,7 and this has led some authors to caution clinicians about the risk of harms when prescribing.8 The risks are thought to be particularly acute for patients who use heroin and those who misuse gabapentinoids. Indeed, the UK government is soon to classify the drug as a class C controlled substance because of its abuse potential and increased reports of deaths attributed to its use.9 Practising clinicians have also recently called for the evidence for the effectiveness of pregabalin to be re-examined in the light of its potential to cause harms.3 4
Rapid reviews use accelerated methods to identify and synthesise the evidence from the literature in order to meet the needs of target audiences including policy makers, healthcare professionals and patient groups.10 The objective of this rapid review was therefore to evaluate the evidence for benefits and harms of pregabalin in the treatment of neuropathic pain in adults, using evidence from published randomised clinical trials (RCTs).
We conducted electronic searches in the following databases: MEDLINE, Embase and Cochrane Central Register of Controlled Trials (CENTRAL). We searched each database from inception until January 2018. No language restrictions were imposed (see online supplementary appendix 1 for a full search strategy). We also hand searched the bibliography of eligible studies (see online supplementary appendix 2 for the full protocol).
We included phase III, double-blinded, placebo-controlled RCTs (efficacy studies) assessing the effects of pregabalin on neuropathic pain in adults aged 18 years and above. We included studies based on the definition of the International Association for the Study of Pain definition.11 These included trials on diabetic neuropathy, HIV-related neuropathy, lumbar radiculopathy, PHN and chronic postsurgical pain. We included RCTs irrespective of study size and duration of intervention. If we included RCTs with a cross-over design, we used data from the first phase of the study. We excluded phase IV trials because they are typically unblinded. We also excluded studies that combined pregabalin with other types of pain intervention because the effects of such interventions would not be exclusively due to the actions of pregabalin; however, cointerventions used as rescue medication were allowed. Trials that randomised participants based on response to pregabalin therapy in the run-in phase were also excluded. Our main outcomes were pain (as measured using validated scales because such scales enhance the credibility of the measured outcomes12) and adverse events. Our secondary outcomes were sleep disturbance, quality of life (QOL), Patient Global Impression of Change (PGIC), Clinician Global Impression (CGI) scale, anxiety and depression, overall discontinuations and discontinuations because of adverse events.
The risk of bias for each included study was rated using the Cochrane criteria.13 Two reviewers (IJO and ETT) independently screened abstracts and determined study eligibility. Disagreements were resolved through discussion. Three reviewers (IJO (8 studies), ETT (8 studies) and JL (10 studies)) independently extracted data according to predefined criteria into customised Excel spreadsheets. The extracted data were independently verified by two reviewers (ETT and IJO). Any disagreements were resolved through discussion. For each included study, we extracted data on study ID, settings, populations, interventions, outcomes and results.
Using the random effects model (Mantel-Haenszel) of the standard meta-analysis software (RevMan V.5.3),14 we computed standardised mean differences (SMDs) and 95% CIs for continuous outcomes and risk ratios with 95% CI for binary outcomes. We used preintervention to postintervention changes to assess intervention effects between pregabalin and placebo. Where studies reported data on change from baseline but did not report SD, we imputed SDs (five studies) based on the SD of other studies included in the meta-analysis.15 We used a value of p=0.05 as our threshold for statistical significance. We assessed heterogeneity using the I2 statistic: values of 25%, 50% and 75% judged mild, moderate and substantial heterogeneity, respectively. We investigated heterogeneity using subgroup (based on central or peripheral neuropathic pain) and sensitivity (based on study quality and/or duration) analyses. We used a funnel plot to assess publication bias.
One reviewer (ETT) entered the data on benefits on RevMan, and these were independently verified by a second reviewer (IJO). One reviewer (IJO) entered the data on harms onto RevMan, and these were independently verified by a second reviewer (ETT). Using the GRADEpro software (V.3.6),16 we rated the overall quality of the body of evidence for each outcome using the Grading of Recommendation, Assessment, Development, and Evaluation (GRADE)17 criteria, which examines the following domains: study design, risk of bias, inconsistency, indirectness and imprecision.
Patient public involvement
Because this was a rapid review, we did not enlist the services of patient representatives in this research.
Our searches identified 1349 non-duplicate citations, out of which 62 articles were considered eligible (figure 1). We excluded 34 articles that did not fit our inclusion criteria (see online supplementary appendix 3 for list of excluded studies and the reasons for exclusion). In total, we included 28 studies18–45 comprising 6087 participants (table 1). The intervention duration was between 3 weeks and 20 weeks (median 8 weeks), and all the trials were industry funded.
Twenty-three studies examined the effectiveness of pregabalin in treatment of peripheral neuropathic pain including diabetic peripheral neuropathy (DPN), PHN and herpes zoster (table 1). Five studies examined the effectiveness of pregabalin for treating central neuropathic pain including sciatica (radicular pain), poststroke pain and spinal cord injury-related pain. Twenty-five studies were conducted in two or more centres. Outcome measures for pain included numerical rating scale (NRS), visual assessment scale (VAS), Short-Form McGill Pain Questionnaire visual assessment scale (SF-MPQ VAS) and SF-MPQ personal pain intensity (SF-MPQ PPI) index (see table 1 for full characteristics of included studies). The overall risk of bias in the included studies was moderate to high (figures 2 and 3). This was mainly due to inadequate reporting of blinding procedures, selective outcome reporting and financial conflicts of interest among study authors (see online supplementary appendix 4 for the risk of bias judgements).
Twenty-one studies provided adequate data on pain using the NRS or variants of it to allow meta-analysis. Meta-analysis showed a significant reduction in pain scores with pregabalin compared with placebo (SMD −0.49 (95% CI −0.66 to −0.32, p<0.00001, I2=88%; figure 4)). Visual inspection of a funnel plot showed that the studies were almost symmetrically distributed around the mean difference for all trials (online supplementary figure S1); trim and fill analyses showed that the subsequent addition of studies with smaller sample sizes did not change the direction of effect. The effect was significant for peripheral neuropathic pain (p<0.00001), but not for central neuropathic pain (p=0.08; online supplementary appendix table 1). The overall quality of the evidence was very low (Summary of Findings (SoF) table 2). Sensitivity analyses revealed similar direction of effects (online supplementary appendix table 2). Four studies that measured pain using NRS did not provide adequate data for meta-analysis; three of these reported significant reductions in pain scores favouring pregabalin over placebo, while one reported no significant difference between groups (see online supplementary appendix table 3).
Three studies measured pain using the VAS, and all showed significant reduction in pain scores favouring pregabalin over placebo (online supplementary appendix table 3). Nine studies measured pain using SF-MPQ VAS, and all reported significant reduction in pain scores favouring pregabalin over placebo. Four studies measured pain using SF-MPQ PPI index, and all reported significant reduction in pain scores favouring pregabalin over placebo.
Figure 5 shows that pregabalin was significantly more likely to cause adverse events compared with placebo (RR 1.33 (95% CI 1.23 to 1.44, p<0.00001, I2=52%). This translates into an absolute effect of 145 (95% CI 101 to 194) more adverse events per 1000 treated. The overall quality of the evidence was low (SoF table 3). Sensitivity analyses revealed similar direction of effects (online supplementary appendix table 2). The risk of experiencing individual adverse events of weight gain, somnolence, dizziness, peripheral oedema, fatigue, visual disturbances, ataxia, non-peripheral oedema, dry mouth, vertigo and euphoria were significantly increased with pregabalin compared with placebo (see online supplementary appendix table 1 and supplementary figures S2 to 12). Pregabalin was also significantly more likely to cause discontinuation because of adverse events (RR 1.91, 95% CI 1.54 to 2.37, p<0.00001, I2=0%); the quality of the evidence was low (SoF table 3; online supplementary appendix table 1; and online supplementary figure S13). Sensitivity analyses by study duration revealed similar direction of effects, but there was no significant difference with higher quality studies (online supplementary appendix table 2).
There was no significant difference in the risk of serious adverse events (RR 0.9; 95% CI 0.66 to 1.24, p=0.50, I2=0%; SoF table 3; online supplementary appendix table 1; and online supplementary figure S14); the quality of the evidence was moderate. Sensitivity analyses showed a significant effect in favour on pregabalin with three higher quality studies, but there was no difference based on study duration (online supplementary appendix table 2). In total, six deaths were reported across four trials, five in pregabalin group and one in placebo (RR 0.86, 95% CI 0.18 to 4.06, p=0.85, I2=0%).
Twenty-one studies measured sleep interference using the NRS sleep interference scale or variants of it. Pregabalin significantly reduced sleep interference scores compared with placebo (SMD −0.38, 95% CI −0.50 to −0.26, p<0.00001, I2=32%); the quality of the evidence was moderate (SoF table 4; online supplementary appendix table 1; and online supplementary figure S15). Fourteen studies reported sleep interference outcome measures with the NRS scale but did not provide adequate data for statistical pooling; 12 of these reported significant reductions in sleep interference scores favouring pregabalin over placebo, while two studies reported no significant difference between groups (online supplementary appendix table 3). Seven studies measured sleep outcomes using the Medical Outcomes Study Sleep Scale (MOS-Sleep). We could not pool results from these studies because of insufficient data. All the studies reported significant improvements in sleep scores in favour of pregabalin over placebo (online supplementary appendix table 3).
Quality of life
Four studies assessed QOL using EuroQoL-5 dimensions scores or variants of it. Two of these reported significant improvements with pregabalin compared with placebo, while the other two reported no significant differences between groups (online supplementary appendix table 3).
Patient Global Impression of Change
Thirteen studies reported this outcome. Ten studies reported significant improvements in PGIC scores with pregabalin compared with placebo, while three studies found no significant differences between groups (online supplementary appendix table 3). We could not pool results from these studies because insufficient data were published.
Clinician Global Impression of Change
Six studies reported this outcome; four of these reported significant improvements with pregabalin compared with placebo, while two found no significant differences between groups (online supplementary appendix table 3).
Anxiety and depression scores
Four studies were pooled for this outcome. There was no significant difference in HADS-Anxiety scores between groups (SMD −0.12, 95% CI −0.29 to 0.04, p=0.14, I2=44%) the quality of the evidence was moderate (SoF table 5; online supplementary figure S16). There was also no significant difference in HADS-Depression scores between groups (SMD −0.06, 95% CI −0.26 to 0.13, p=0.54, I2=60%) the quality of the evidence was low (SoF table 5; online supplementary appendix table 1 and online supplementary figure S17). One study41 that did not provide sufficient data for statistical pooling reported significant improvement in the HADS-Anxiety scores in favour of pregabalin, but no significant difference in HADS-depression scores between groups (online supplementary appendix table 1). One study40 measured anxiety using the VAS anxiety scale and reported significant improvements in QOL scores with fixed-dose and flexible-dose pregabalin compared with placebo (p=0.03 and p=0.02, respectively).
In total, there were 1203 drop-outs (approximately 20%) in the 28 trials (n=5972) that reported the data (online supplementary appendix table 1). There was no significant difference in overall discontinuation rates between groups (RR 1.09 (95% CI 0.93 to 1.28, p=0.29, I2=51%)).
Summary of the evidence
The evidence from published RCTs suggests that pregabalin reduces pain in patients with neuropathic pain. The effect is statistically significant in peripheral neuropathic pain, but not with central neuropathic pain. Pregabalin significantly increases the risk of adverse events including weight gain, somnolence, dizziness, dry mouth, peripheral oedema, fatigue, visual disturbances, ataxia, non-peripheral oedema, vertigo and euphoria. Pregabalin significantly reduces sleep interference scores compared with placebo. There was insufficient evidence to assess an effect on QOL. The evidence for PGIC and CGIC scores was mixed among studies that reported these outcomes, and there were no significant effects on HADS anxiety and depression scores compared with placebo. There were five deaths in the pregabalin arms and one in the placebo but insufficient power to detect an overall effect.
Comparison with the existing literature
We have identified several published reviews assessing the effectiveness of pregabalin the management of neuropathic pain, and our results are partly consistent with these. Zhang et al 46 and Wang et al 47 showed that pregabalin was more efficacious than placebo for treatment of DPN-associated pain and PHN-associated pain respectively; however, the two reviews did not base their results on changes from baseline between groups. Semel et al 48 and Freeman et al 49 also concluded that pregabalin was more effective than placebo for neuropathic pain; however, both reviews did not account for the quality of the included primary studies. Finnerup et al 50 concluded that there was modest evidence supporting the use of pregabalin for treatment of neuropathic pain; although the authors used GRADE criteria to assess the strength of recommendation, they did not report the quality of the evidence. In an overview of Cochrane reviews, Wiffen et al 51 concluded that there was clinical trial evidence supporting the use of pregabalin for treatment of some aspects of neuropathic pan; however, the authors did not rate the quality of the evidence for the outcomes reported.
Two reviews52 53 that examined the safety profile of pregabalin concluded that pregabalin use was significantly more associated with adverse events than placebo; however, both reviews did not rate the quality of the evidence for the outcomes reported.
Comparison with existing guidelines
We identified several guidelines that recommend the use of pregabalin for treatment of neuropathic pain, and some of their specifications are consistent with our results. For instance, the European Federation of Neurological Societies guideline54 based on data from comparative studies recommended pregabalin as first-line treatment for neuropathic pain; however, the guidance assessed only the level, but not the quality, of the evidence, and also notes that there are too few large-scale comparative studies to make definite conclusions about the benefits and harms. Similarly, the American Academy of Neurology, the American Association of Neuromuscular and Electrodiagnostic Medicine and the American Academy of Physical Medicine and Rehabilitation guidance55 recommend pregabalin as first-line treatment based on levels (and not quality) of the evidence; however, they guidance recommends that clinical trials of longer duration should be conducted. The Canadian Pain Society guidance56 recommends pregabalin as first-line treatment for neuropathic pain but acknowledges that paucity of longer duration trials limit the conclusions that can be drawn about its benefits and harms on the long term.
Strengths and limitations
This rapid review has limitations due to its streamlined methods and search strategy. First, the rapid review methodology employed could have introduced selective outcome reporting bias; nevertheless, most of the outcomes reported in this review have been listed as outcomes of interest to be considered when designing trials of neuropathic pain interventions.57 There is a risk that our review may be prone to sampling bias and that we may have missed potentially eligible studies, which could have been identified by searching clinical trials registries and grey literature. However, we comprehensively searched the literature and used standard criteria to assess the risk of bias and rate the quality of the evidence. It has also been reported that generally the conclusions of rapid reviews and full reviews do not greatly differ,10 and enhanced rapid reviews where data are independently checked by a second reviewer could help policy makers with quicker access to the evidence base.58 This review therefore provides the most up-to-date comprehensive summary of the available literature, as it accounts for study quality and reports clinically meaningful patient outcomes. We did not assess the extent to which different doses of pregabalin influenced the outcomes assessed; in addition, the benefits and harms of pregabalin were not analysed according to specific neuropathic pain conditions; only two subgroups (central and peripheral neuropathic pain) were assessed.
Implications for research
The quality of the included studies examining efficacy of pregabalin for pain was rated as low or very low according to the GRADE framework. This highlights the need for larger, robust, high-quality clinical trials to be conducted, with particular attention paid to minimising selective reporting of outcomes. Concerns about selective reporting could be mitigated if drug manufacturers enabled access to clinical study reports (CSRs), especially as industry-sponsored trials are likely to skew reports in favour of benefits over harms.59 60 This would allow for a more comprehensive assessment of the benefits and harms of pregabalin. Of note, all the included trials were industry sponsored, and an overwhelming majority of the authors of the include studies had financial ties to the pharmaceutical industry. Of note, the results of the only published charity-funded phase IV placebo-controlled trial that assessed the effectiveness of pregabalin in management of neuropathic (radicular) pain contrast our meta-analysis results; there was no significant difference in pain scores between groups.61 Independent and publicly funded trials assessing the benefits and harms of pregabalin should be conducted. Only a few studies assessed the effect of pregabalin in improving QOL, anxiety and depression and CGIC. Future trials should further assess the role of pregabalin for these outcomes. Studies investigating the type of neuropathic pain pregabalin relieves (eg, stimulus-dependent pain such as hyperalgesia or allodynia) or spontaneous pain could be an area of consideration for future research.
That the median duration of intervention was 9 weeks suggests that the intermediate to longer term benefits of pregabalin for neuropathic pain are unproven. Indeed, in real-life clinical care, it has been reported that the initial benefits seen with use of the drug in patients with neuropathic pain were no longer apparent after 6–12 months of therapy.62 Therefore, RCTs that are adequately powered and with longer durations of interventions are desirable. The finding of five deaths among 891 participants on pregabalin, versus one death among 320 participants on placebo, is somewhat concerning. Given the low frequency of this outcome (coupled with the short trial durations), RCTs are unlikely to be informative; we suggest pharmacoepidemiological studies in routinely collected electronic health records and spontaneous reporting databases to assess the impact of pregabalin on mortality.
Implications for clinical practice
Very low-to-moderate quality evidence suggests that pregabalin improves some symptoms of neuropathic pain. However, it significantly increases the risk of adverse events including somnolence, oedema, visual disturbances, ataxia, vertigo and euphoria. Pregabalin also increases the risk of drug discontinuation because of adverse events. Clinicians should be cautious about prescribing pregabalin and should consider whether its benefits outweigh potential harms in individual patients.
The evidence from RCTs in journal publications suggests that pregabalin has beneficial effects on some symptoms of neuropathic pain. However, its use significantly increases the risk of adverse events and discontinuation due to adverse events. The quality of the evidence from journal publications is overall low, and the duration of trials is short. Greater transparency in the reporting of outcomes is advocated; independent and publicly funded trials assessing the effects of pregabalin in neuropathic pain should be encouraged. Allowing researchers access to full CSRs of pregabalin trials should be a priority for drug companies and regulators.
Patient consent for publication Not required.
Contributors IJO was involved with devising the review methods, conducting electronic searches, screening of abstracts, data extraction, data analysis and interpretation and codrafting of the review. ETT was involved with devising the review methods, screening of abstracts, data extraction, data analysis and interpretation and codrafting of the review. JJL was involved with data extraction, data analysis and interpretation and codrafting of the review. BG and CJH were involved with devising the review methods, data analysis and interpretation and codrafting of the review.
Funding IJO, BG and CJH are part of the Evidence Synthesis Working Group. The Evidence Synthesis Working Group is funded by the National Institute for Health Research School for Primary Care Research (NIHR SPCR) (ProjectNumber 390). JJL is supported by an NIHR In Practice Fellowship.
Disclaimer The views expressed are those of the author(s) and not necessarily those of the NIHR, the NHS or the Department of Health.
Competing interests CJH has received expenses and fees for his media work. He has received expenses from the WHO and FDA and holds grant funding from the NIHR, the NIHR School of Primary Care Research, The Wellcome Trust and the WHO. He has received financial remuneration from an asbestos case. He has also received income from the publication of a series of toolkit books published by Blackwell’s. On occasion, he receives expenses for teaching EBM and is also paid for his general practitioner work in National Health Service out of hours. CEBM jointly runs the EvidenceLive Conference with the BMJ and the Overdiagnosis Conference with some international partners that are based on a non-profit making model. BG receives funding from the Laura and John Arnold Foundation and reports personal fees from intermittent additional personal income from speaking and writing for lay audiences on problems in science and medicine including regulatory shortcomings.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement No additional data available.