Letter to the Editor Regarding "Gabapentin has Longer-Term Efficacy for the Treatment of Chronic Pelvic Pain in Women: A Systematic Review and Pilot Meta-analysis".

Gabapentin is a drug with evidence of efficacy in neuropathic pain that has been increasingly used in other chronic pain conditions without robust supporting data. An adequately powered study showed no benefit of gabapentin over placebo for this indication at 3 months with high rates of side effects. This is a response article where we highlight our concerns about the suggestion that gabapentin may have longer-term efficacy for the treatment of chronic pelvic pain in women. Dear Editor, We were concerned to read the article entitled “Gabapentin has Longer-Term Efficacy for the Treatment of Chronic Pelvic Pain in Women: A Systematic Review and Pilot Meta-analysis”, published recently by Fan and colleagues [ 1 ], which concluded that gabapentin could be a potential treatment option for chronic pelvic pain in women. We believe that the methodology used in this systematic review is inappropriate, and thus draws erroneous conclusions with the potential to expose women with chronic pelvic pain to unnecessary risk. Gabapentin is a drug with evidence of efficacy in neuropathic pain [ 2 ] that has increasingly been used in other chronic pain conditions despite a lack of robust data to support its use [ 3 – 6 ]. Not only are there a range of unpleasant side effects associated with gabapentin, but there is also a very real risk of addiction and abuse potential as seen in both European and US populations [ 7 , 8 ]. We recently published the findings of a large multicentre, randomised, placebo-controlled trial exploring the efficacy and safety of gabapentin in women with chronic pelvic pain and no obvious pelvic pathology [ 9 ]. We compared gabapentin to placebo in 306 women (153 per group) and found no significant difference in any of our primary or secondary outcomes. However, we did observe a higher proportion of serious adverse events and other known side effects in those using gabapentin. The trial was designed after completion of a pilot study [ 10 ] comparing gabapentin to placebo in 47 women ( n  = 22 vs. n  = 25, respectively). In their systematic review and meta-analysis, Fan and colleagues combine data from our two trials with two further studies: a pilot study comparing gabapentin to amitriptyline ( n  = 20 vs. n  = 20) and to a combination of gabapentin and amitriptyline ( n  = 16) [ 11 ] (although only the first comparison is included in this review); and a single-centre trial comparing gabapentin to placebo in 60 women (30 per group) [ 12 ]. Only our trial is considered to have a low risk of bias [ 9 ], with the two remaining studies generously assessed as unclear [ 10 , 12 ] and of having a high risk of bias [ 11 ]. The meta-analysis reports two primary outcomes (change in pain scores at 3 and 6 months) and a number of secondary outcomes. The only outcomes for which a significant difference between the gabapentin and control groups was identified were change in pain scores at 3 and 6 months and the adverse effects of dizziness and somnolence. Importantly, our full randomised controlled trial (RCT) assessed outcomes at 13–16 weeks of treatment and thus the meta-analysis of data reported at 6 months does not include data from this study, the largest of the datasets. Nonetheless, the authors conclude that “whilst the change in pain scores failed to meet the threshold for a MCID (minimally clinically important difference) during the 3-month period, the changes were higher than MCID during the 6-month period”. We would additionally argue that the use of change in pain scores is inappropriate as it does not take account of baseline pain severity. Given that our large adequately powered trial with low risk of bias [ 9 ] found no benefit of gabapentin over placebo at 3 months and identified a high rate of side effects (and that gabapentin has such recognised abuse and addictive potential that it has been reclassified as a controlled drug in the UK), we do not feel it is appropriate to suggest that studies with longer dosing periods be carried out on the basis of meta-analysis of data from three studies including only 147 women in total.