Gabapentinoids not indicated for chronic back pain

​​a recent systematic review/meta-analysis looked at the benefits and safety ofgabapentinoids for chronic low back pain, finding not much benefit yet significant adverse effects (see http://journals.plos.org/plosmedicine/article/file?id=10.1371/journal.pmed.1002369&type=printable​ ).

Results:

--only 8 studies found:

    ​--3 compared gabapentin (GB) with placebo (n=185): minimal improvement in pain with gabapentin. mean difference 0.22 units, not statistically significant, quality of evidence very low

    ​--3 compared pregabalin (PG) alone vs other analgesics: amitriptyline, celecoxib,tramacet (a combo of tramadol and acetaminophen); n=332, studies combined. Greater improvement in cominaiton of the non-pregabalin groups (mean difference 0.42 units, quality of evidence very low, though I really doubt that 0.42 means anything clinically in a scale from 0-10). the biggest study compared 200 patients on amitriptyline​ 50mg vs PG 600mg/d for 14 weeks, finding that amitriptyline​ worked better than PG (in this study 57% on amitriptyline and 39% on PG had a >50% reduction in pain using a visual analog pain scale, with p=0.01).  The other 2 studies found some benefit for the combination of drugs (eg the combo of PG with celecoxib was somewhat better than either drug alone in another RCT), but these studies were considered to be very low quality of evidence with methodologic issues, small numbers of patients, short follow-up, risk of detection bias, and risk of allocation concealment bias. (eg the celecoxib study had only of 36 patients and with results only at 4 weeks).  

    ​--2 other studies of PG: the largest (n=423) and the one with longer followup (8 weeks) found that adding pregabalin to tapentadol added no benefit, though adding PG to transdermal buprenorphine 35 mcg/hr in an RCT of 45 patients in a 3 week trial found PG did help (again, small study, short followup). Of note, though the tapentadol study didnot find benefit, the non-PG group did receive a higher tapentadol dose of 4-500 mg/d vs 300 mg/d.

--adverse events

    --GB: dizziness (RR 1.99, ie almost twice as often), fatigue (RR 1.85), difficulties with mentation (RR 3.34), visual disturbances (RR 5.72)

        --overall, NNH (number needed to harm): respectively 7,8,6,6

    --PG: dizziness (RR 2.70), with absolute risk increase of 9%, NNH 11

Commentary:

--chronic LBP is really common (no surprise), with lifetime prevalence of 51-80%

--Pregabalin and gabapentin have been used as adjuncts for chronic pain. But, recommendations for their use are only for neuropathy, not for chronic pain (eg, see https://www.cdc.gov/mmwr/volumes/65/rr/rr6501e1.htm ). It seems likely that the reason that gabapentinoids are used for chronic pain is a "migration" of indication: seems to work for neuropathic pain, so why not try it for chronic pain? and sometimes it is hard to differentiate the two types of pain.  [But, i should add that a http://gmodestmedblogs.blogspot.com/2017/04/diabetic-peripheral-neuropathy-and-more.html of meds for diabetic peripheral neuropathy found that pregabalin had a small effect based on low-quality evidence and gabapentin was ineffective.

--and there are significant risks for abuse. See https://www.statnews.com/2017/07/06/gabapentin-opioid-abuse/ , noting that a 2016 report in the journal Addiction found that 1/5 of opiate users also used gabapentin, in Ohio GB surpassed oxycodone prescriptions (in fact in May 2017, it was the 5^th most prescribed med in the US overall!!!), is pretty frequently used in those with overdose deaths, it is pretty cheap on the streets (vs other drugs), and users report that it “makes them feel good” and acts as a sedative but also can “enhance the euphoria caused by an opioid and stave off drug withdrawals. In addition, it can bypass the blocking effects of medications used for addiction treatment, enabling patients to get high while in recovery”. Of note, it is unclear exactly what the specific harm of GB is (long-term risks of high doses are not well known), though by itself it does not appear to lead to lethal overdoses. but one concerning point, at least to me, is that by augmenting the euphoria of the opiates, it might well decrease the likelihood that users will be interested in or able to sustain getting off the opiates, with their attendant mortality and profound social effects.

--overall, the above studies were of pretty low quality overall, so hard to draw definitive conclusions.  And there were no (as in, zero) studies of PG vs placebo. But pending better quality studies, it certainly appears that adding gabapentinoids pretty clearly adds adverse effects (both to the patient, and perhaps to society in general through diversion) and seems unlikely to add benefit

--as a perhaps side-issue, it does seem from one of the larger studies noted above that amitriptyline may help (at least more than PG). Though there are few comparative studies, the ones I've seen suggest that all the tricyclics work similarly (at least one study using amitriptyline vs desipramine), and that amitriptyline has the most adverse events of any of them.  I have been using basically 2 of the tricyclics for the past 20 years or so: desipramine, which seems to have the fewest adverse events overall (dosed at 25mg for 1 week, increasing progressively to 50 then 75, by which time there usually is some effect), and nortriptyline, which does have more antihistaminic effects than desipramine and causes sleepiness (so, i prescribe this one for those with sleep problems, though it is dosed at half the others: 10mg, then to 25 then to 50 as my usual max)

so, 

--bottom line here is that there are no clear data showing benefit for the gabapentanoids for chronic low back pain, with the caveats that there have not been high-quality rigorous studies done. ​ But there are pretty impressive data on their harms..... 

--however, despite this, it does seem that these drugs are prescribed frequently for chronic low back pain, with a Canadian study finding that off-label use of PG was 75%, and the most prevalent condition for that was chronic low back pain.​

--my guess is that the drug companies have approximately zero interest in doing any more studies on these drugs, since these drugs are so frequently prescribed anyway (ie, no financial upside to them, with potential downside if the studies were negative). so, i think we are unlikely to get better data anytime soon, which to me makes it difficult to justify using these drugs for chronic low back pain (and, likely as well for other chronic non-neuropathic pain syndromes)