Statistics from Altmetric.com
Recent adequately powered, randomised controlled trials have shown the efficacy of corticosteroids in Bell's palsy. These contradict the recent study by Tong et al published in Acupuncture in Medicine and the 2004 Cochrane review on the subject. The authors believe that the study by Tong et al has serious methodological flaws that make its results invalid in determining the efficacy of corticosteroids or acupuncture for idiopathic facial palsy.
We would like to express our concerns regarding the randomised controlled trial from Tong et al, which aimed to show the efficacy of steroids or acupuncture for idiopathic peripheral facial paralysis.1
The authors mention in the Introduction that there is no good evidence that medication is effective for Bell's palsy and refer to the Cochrane database of systematic reviews. Since there is no citation, we assume that they are referring to the 2004 Cochrane review, which included four trials with a total of 179 patients, and concluded that there was no evidence for the efficacy of corticosteroids.2 Since that review, there have been two recently published, adequately powered, randomised controlled trials that contradict the findings of the 2004 Cochrane review and the results of the trial by Tong et al. Both of these trials had significantly greater numbers than the Cochrane review and showed that prednisolone significantly improves the chances of complete recovery and shortens the time to recovery in patients with Bell's palsy.3 4
We believe that this study has major flaws in its methodology that seriously compromise its internal validity. It is apparent that the sample size of this trial was not calculated before the start of the study. This resulted in a small sample size that did not have the power to detect the effect of steroids or acupuncture in the treatment of Bell's palsy. The authors have not explained why there were disparate numbers of patients recruited in the three groups with almost twice the number in the steroid group as there were randomised to receive acupuncture. Block randomisation could have been used to avoid the significant imbalance in sample size among the three groups.
Although the authors have mentioned that poor prognostic factors include poor initial grade of severity, diabetes mellitus and hypertension, they have only made mention of initial grade of severity in the three groups. Since there were significantly milder cases in the control group, this was likely to affect the final results when a comparison was made with acupuncture or corticosteroids. This is a major flaw in the study and could have been avoided if randomisation of patients was stratified by the initial grade of facial palsy. Also, no mention was made of the numbers of diabetic or hypertensive patients in the three groups, so it is not possible to say if these groups were comparative.
The Copenhagen Facial Nerve Study showed that 85% of people with idiopathic peripheral facial nerve palsy regained normal function within 3 weeks.5 As this has been previously established in the literature, most interventional study designs involve early recruitment of affected patients, and the two recent studies quoted above started intervention within 72 h of onset of symptoms. In our view, it is unfair to assess the effect of any attempted intervention by broadening the recruitment period to 2 weeks as Tong et al have done in this study.
Although the acupuncture points used were according to Traditional Chinese Medicine, they would also be acceptable to Western Medical acupuncturists, since they were over the muscles supplied by the facial nerve. However, it is debatable whether an adequate acupuncture dose was given.6 The use of 20 min of manual acupuncture eliciting de qi given three times each week may be adequate for chronic conditions such as knee pain. However, we would like to have seen the use of electroacupuncture in an acute condition such as idiopathic facial palsy when time is of the essence in affecting a change, especially when starting as late as 2 weeks after onset as the authors have done in this study.
Competing interests None.
Provenance and peer review Not commissioned; not externally peer reviewed.
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.