TamaraPringsheim, University of Calgary, Calgary, Alberta, Candadatmprings@ucalgary.ca
Werner Becker (wbecker@ucalgary.ca)
Submitted April 01, 2011
We appreciate the comments by Peterlin et al. and Gantenbein et al. The correspondents are concerned that the results of this analysis do not reflect current management of CH in clinical practice.
Like most systematic reviews and meta-analyses on medication treatment, we included only randomized double-blind controlled trials (RCTs) in our analysis. Furthermore, the strength of advisements was based on trial methodological quality and this was the basis of the rating system used in our article. This rating system was adopted from the AAN Quality Criteria used for Practice Parameters.
Researchers can be more confident that a trial with appropriate methodological rigor is more likely to reflect real differences between a treatment and placebo. Civamide was studied with greater methodological rigor than verapamil and therefore received a Level B rating while verapamil received a C. We could not make a formal evidence-based advisement on prednisone because of the poor methodological quality of available RCTs. We could not advise on the use of methysergide as no RCTs have been performed with this medication.
In addition, a Level B rating does not imply that a treatment is "first-line", as Gantenbein et al. stated. A treatment that receives a Level B rating is defined as "probably effective for the given condition in the specified population." Although civamide is not yet available for clinical use, the results of the RCT suggest that it is probably effective in CH patients and will be a valuable addition to current treatment options.
Our goal was to systematically summarize the current RCT evidence that exists on CH treatment. As we noted in our Discussion, there is a disconnect between the RCT literature and current treatment of CH. We agree with Peterlin et al. that verapamil is the drug of first choice for prophylaxis of CH even though it received only a level C advisement. This does not mean we do not use it but rather that we were unable to make stronger formal advisements using the current evidence-based rating system.
We agree there needs to be a marriage between the evidence and clinical expertise and we attempted to convey this in our Discussion. However, a systematic review must objectively assess the available evidence and separate it from expert opinion. A systematic review is not a practice guideline, as it is limited in its scope by the available evidence. We hope that readers will read both the systematic review in our article and the Discussion if they are looking for guidance on current treatment of CH.
In our Discussion, we consider the use of steroids as an effective treatment for CH in clinical practice despite the lack of evidence. We refrained from making general endorsements of methysergide given the complications related to its use, drug interactions with common symptomatic therapies, and lack of evidence to support its use, while recognizing that in special scenarios this medication may be useful in patients with CH.
The systematic review and discussion in our paper highlight, as Peterlin et al. so eloquently point out, the lack of formal evidence for many commonly used treatments, and we echo their call for headache specialists to work towards better evidence for our therapies.
We appreciate the comments by Peterlin et al. and Gantenbein et al. The correspondents are concerned that the results of this analysis do not reflect current management of CH in clinical practice. Like most systematic reviews and meta-analyses on medication treatment, we included only randomized double-blind controlled trials (RCTs) in our analysis. Furthermore, the strength of advisements was based on trial methodological quality and this was the basis of the rating system used in our article. This rating system was adopted from the AAN Quality Criteria used for Practice Parameters.
Researchers can be more confident that a trial with appropriate methodological rigor is more likely to reflect real differences between a treatment and placebo. Civamide was studied with greater methodological rigor than verapamil and therefore received a Level B rating while verapamil received a C. We could not make a formal evidence-based advisement on prednisone because of the poor methodological quality of available RCTs. We could not advise on the use of methysergide as no RCTs have been performed with this medication.
In addition, a Level B rating does not imply that a treatment is "first-line", as Gantenbein et al. stated. A treatment that receives a Level B rating is defined as "probably effective for the given condition in the specified population." Although civamide is not yet available for clinical use, the results of the RCT suggest that it is probably effective in CH patients and will be a valuable addition to current treatment options.
Our goal was to systematically summarize the current RCT evidence that exists on CH treatment. As we noted in our Discussion, there is a disconnect between the RCT literature and current treatment of CH. We agree with Peterlin et al. that verapamil is the drug of first choice for prophylaxis of CH even though it received only a level C advisement. This does not mean we do not use it but rather that we were unable to make stronger formal advisements using the current evidence-based rating system.
We agree there needs to be a marriage between the evidence and clinical expertise and we attempted to convey this in our Discussion. However, a systematic review must objectively assess the available evidence and separate it from expert opinion. A systematic review is not a practice guideline, as it is limited in its scope by the available evidence. We hope that readers will read both the systematic review in our article and the Discussion if they are looking for guidance on current treatment of CH.
In our Discussion, we consider the use of steroids as an effective treatment for CH in clinical practice despite the lack of evidence. We refrained from making general endorsements of methysergide given the complications related to its use, drug interactions with common symptomatic therapies, and lack of evidence to support its use, while recognizing that in special scenarios this medication may be useful in patients with CH.
The systematic review and discussion in our paper highlight, as Peterlin et al. so eloquently point out, the lack of formal evidence for many commonly used treatments, and we echo their call for headache specialists to work towards better evidence for our therapies.