Vitamin B6


Don't confide in what you are being told about B6 here (some people just don't believe in it, and would criticize the truth). Your operating doctor, and the faculty you've dealt with are experts in the field, and obviously know more about B6.. Here are a few controlled studies done proving the effectiveness of B6


OK lets take a look at that.

First of all Khorvash F et al 2012 "Treatment of carpal tunnel syndrome: A trial of vitamin B6" Pak J Med Sci 28(2) 283-286

This is a prospective randomised trial comparing Splint+Placebo with Splint+Vit B6. This is not a bad trial design but as both groups were given a treatment which is known to be effective (splinting) we would expect both groups to improve. The diagnostic criteria are well specified and these patients did all have CTS so far as one can tell. Some of the patients also received Naproxen and all were also instructed in range of motion exercises - another treatment of uncertain value. The dose of B6 used was 80mg and a well contstructed placebo was used but the method of randomisation is not described at all in the paper - a fundamental sign of risk of bias. 86 of 90 patients completed the study (3 months) which is pretty good. So far not too bad. Not a perfect study but better than many that get into print.

Now we get to the results and here it starts to go downhill I'm afraid. The primary outcome measure is said to be the NCS results. This sounds very scientific but there are some problems with this. Firstly NCS results often have very little relationship to symptoms so it may not mean much to the patient that their median nerve is slightly faster, Secondly, we customarily measure at least 4 or 5 different parameters in a set of nerve conduction studies and in this paper the authors have eventually chosen to quote just one - rather suggesting that the others perhaps produced results that did not fit their hypothesis so they have quietly ignored them. This is a perennial statistical problem when many things are measured in a group of patients. If you measure 20 things at least one of them is likely to show a 'significant' difference by chance and you can just report that one as a positive result for your study. This error is extremely widespread in CTS studies. Now we come to the worst bit - the claimed difference in the abstract is better sensory peak latency in the B6 group than in the control group after treatment p=0.002. When you look at the actual figures in the paper the after treatment peak latency in the B6 group was 4.8 msec with a standard deviation of 1.2 msec while the after treatment peak latency for the control group was 4.9 msec with a standard deviation of 1.3 msec. The p-value for this difference is actually 0.7 - ie their statistics are completely wrong!

At this point it stops being worth discussing this paper further - what they have actually shown is that the effect of B6 was virtually identical to that of placebo.


NEXT - Talebi et al 2013 "Effect of vitamin B6 on clinical symptoms and electrodiagnostic results of patients with carpal tunnel synrome" Adv Pharm Bull 3(2): 283-288

This is methodologically much poorer - smaller (20 patients treated 19 control), no placebo, unblinded. The dose of B6 used was 120mg daily for 3m and follow-up was performed at 3 months. The analysis was performed by hand rather than patient with two hands from the same patient being counted as separate data entities - a fundamental statistical flaw. As in Khorvash the method of randomisation is not described. They used a a better outcome measure in some ways, the Boston/Levine SSS/FSS score  which you can find on this site, but they analysed every question separately and they do not tabulate the actual results in the paper, providing only a table of alleged p-values for the difference between before and after. No top class journal would accept this and it doesn't say much for the Tabriz University of Medical Science who publish Adv Pharm Bull (This seems to be the house journal of the institution where the authors of this paper work). As no real data is presented you cannot tell for sure what these authors found but if their table of p-values is accepted at face value then what they have shown is probably placebo effect.

In contrast to these it is worth reading:

1) Franzblau A, Rock CL, Werner RA, et al. The relationship of vitamin B6 status to median nerve function and carpal tunnel syndrome among active industrial workers. J Occup Environ Med 1996;38:485-91.

This study looked quite hard for evidence of B6 deficiency in people with CTS and failed to find any so the original claim by Ellis et al that CTS patients are deficient in B6 does not really hold water. That does not however mean that 'extra' B6 might not help so we have to look for therapeutic trials. The two above are quite new but, as I have explained above, not very convincing. Indeed one of them actually seems to actively support the idea that B6 is ineffective when you look at the numbers. The topic has been subjected to attempts at meta-analysis however:


2) O'Connor et al "Non-Surgical treatment (other than steroid injection) for carpal tunnel syndrome. Cochrane Database Syst Rev 2003 (1) - freely available online

This review found 1 fairly good, and one moderately good placebo controlled study of B6 - there was no evidence of efficacy for most outcomes though one B6 group did report less "finger swelling and movement discomfort" (see notes above about measuring lots of things)


3) Piazzini et al "A systematic review of conservative treatment of carpal tunnel syndrome" Clin Rehabil 2007; 21: 299-314

These authors found the same two papers as O'connor to include in their review but their conclusion was more robust "Vitamin B6 is ineffective"

There is a common trend here - the better the methodology of the study being looked at the less likely it is to show any evidence of benefit for B6. No-one ever said that evidence based medicine was easy - there are endless traps and pitfalls in designing, carrying out, publishing and meta-analysing clinical trials. We know how to minimise the risks of most of them now - after decades of effort and study. JB


For the most part, I have been taking 100mg of B6 for quite some time. In response to your warning awhile back, I did reduce the dose from 200 mg/day to 100mg.

In a mostly unrelated matter, I sustained a painful neurological injury to my left eye from accidentally looking into an outdoor LED bulb from only 1 and 1/2 feet away. Have been diagnosed by a Neuro-Opthalmologist with Photo-Occulodynia with a strong photosensitivity to the highly concentrated blue light radiation emanating from LED bulbs. My primary Opthalmologist called it periocular neuropathic pain. At any rate, the pain took nine months to heal. I continued with B6 for the entire time.

Thus, I have no idea if B6 helped or if I would have healed without it. It did provide a psychological benefit in that I was taking something extra that might be accelerating the healing process. The treatment for the neuro pain was Neurontin, BTW, which was very effective.

This is a fantastic site. I am greatly appreciative of the support you provided me with my CTS issue. And I could tell that your opinions are very thoughtful and worthy of trust.


To follow that train of thought, I would suggest that vitamin B6 and Carpel Tunnel Release Surgery both have uncertain outcomes. Very astute research/evidence compiling my friend, I will indeed check on those references, thanks also for the hard work that you put into this forum :)


Different degrees of uncertainty! We know what proportion of people feel much better after carpal tunnel surgery with great precision (see the surgical prgnosis page) and it is quite high. I don't think the operation would survive otherwise. What we don't know with such precision is the proportion of people with CTS who would have got better without any treatment,. However the difference between the best estimate of the rate of natural remission (about 20%) and the rate of 'success' from surgery (about 80%) is so great that the probability that surgery is an 'ineffective' treatment is extremely low, even without a proper randomised, placebo controlled trial. The argument here is that you don't need a placebo controlled trial to demonstrate the effectiveness of a parachute in preventing death from jumping out of planes. On the other hand, the study by Khorvash et al which you quoted - despite apparently being misreported - actually seems to show that it is considerably more likely than not that 80mg/day of B6 is no more effective than placebo. Yes there is uncertainty in both estimates, as there always is, but with B6 it is a small probability that the treatment is effective, while for surgery it is a very small probability that it is ineffective. If these were two horses in a race I would have little difficulty in deciding which to bet on.

My main grouse with surgery is not that it is an ineffective treatment, it is that some patients suffer unacceptable side effects (something that goes with all effective treatments - if there are no side effects the treatment probably isn't doing anything). These are an unavoidable part of doing the operation correctly, notwithstanding the fact that I have just been told by a patient in another forum that 'their' surgeon has done 1500 endoscopic carpal tunnel operations with 100% success, a statistic which I simply do not believe! I therefore try to avoid surgery where possible by using methods which have a reaosnable chance of producing improvement, but I can't honestly recommend something where the balance of probabilities favours it having no effect, especially when there are other relatively safe alternatives for which there is very good evidence, such as local corticosteroid injection.

Medicine is a game of probabilities, not certainties, but the existence of uncertainty does not mean that we know nothing! JB


The review in the Journal of Evidence Based Complementary and Alternative Medicine by Bender 2011 which is in the reference list here does in fact cover all therapeutic uses and is an interesting read if you can get access to it. I think B6 does at least make quite a good placebo - you seem to have to take more than 1000 mg daily to run a risk of nerve damage from it. JB

This site uses cookies. By continuing to browse the site you are agreeing to our use of cookies. Find out more here.