Ranking antidepressants

Abstract

We have some concerns with the design of Andrea Cipriani and colleagues' systematic review of the relative efficacy and acceptability of 12 new-generation antidepressants.1Cipriani A Furukawa TA Salanti G et al.Comparative efficacy and acceptability of 12 new-generation antidepressants: a multiple-treatments meta-analysis.Lancet. 2009; 373: 746-758Summary Full Text Full Text PDF PubMed Scopus (1265) Google ScholarResponse rate was chosen as the primary measure of efficacy to make the interpretation of results easier for clinicians. However, such analyses are ordinarily used as secondary outcome measures of efficacy, and are valuable only when significant differences have been established.2European Medicines AgencyNote for guidance on clinical investigation of medicinal products in the treatment of depression.http://www.emea.europa.eu/pdfs/human/ewp/051897en.pdfGoogle Scholar Assessment of response should be based on change in score on the Hamilton depression rating scale (HDRS) or Montgomery-Åsberg depression rating scale (MADRS), which are the most commonly used primary outcome measures and recommended in clinical trial guidelines. Responder rate analyses transform these continuous symptom scores into a dichotomous variable, thereby precluding a more precise comparison between treatments. Relying only on response rates without an assessment of the change in scores on which the rates are based can inflate differences between treatments.3Kirsch I Moncrieff J Clinical trials and the response rate illusion.Contemp Clin Trials. 2007; 28: 348-351Summary Full Text Full Text PDF PubMed Scopus (62) Google Scholar, 4National Institute for Health and Clinical ExcellenceDepression: management of depression in primary and secondary care.http://www.nice.org.uk/Guidance/CG23Google ScholarAn assessment of the robustness of the analysis by Cipriani and colleagues is therefore, paradoxically, made more difficult by the choice of outcome measure. A comparison of efficacy between drugs based on overall change in HDRS or MADRS score is therefore required before conclusions about differences can be drawn.SS declares that she has no conflicts of interest. PH received a scholarship in clinical pharmacology from Merck Sharp & Dohme in 2005. We have some concerns with the design of Andrea Cipriani and colleagues' systematic review of the relative efficacy and acceptability of 12 new-generation antidepressants.1Cipriani A Furukawa TA Salanti G et al.Comparative efficacy and acceptability of 12 new-generation antidepressants: a multiple-treatments meta-analysis.Lancet. 2009; 373: 746-758Summary Full Text Full Text PDF PubMed Scopus (1265) Google Scholar Response rate was chosen as the primary measure of efficacy to make the interpretation of results easier for clinicians. However, such analyses are ordinarily used as secondary outcome measures of efficacy, and are valuable only when significant differences have been established.2European Medicines AgencyNote for guidance on clinical investigation of medicinal products in the treatment of depression.http://www.emea.europa.eu/pdfs/human/ewp/051897en.pdfGoogle Scholar Assessment of response should be based on change in score on the Hamilton depression rating scale (HDRS) or Montgomery-Åsberg depression rating scale (MADRS), which are the most commonly used primary outcome measures and recommended in clinical trial guidelines. Responder rate analyses transform these continuous symptom scores into a dichotomous variable, thereby precluding a more precise comparison between treatments. Relying only on response rates without an assessment of the change in scores on which the rates are based can inflate differences between treatments.3Kirsch I Moncrieff J Clinical trials and the response rate illusion.Contemp Clin Trials. 2007; 28: 348-351Summary Full Text Full Text PDF PubMed Scopus (62) Google Scholar, 4National Institute for Health and Clinical ExcellenceDepression: management of depression in primary and secondary care.http://www.nice.org.uk/Guidance/CG23Google Scholar An assessment of the robustness of the analysis by Cipriani and colleagues is therefore, paradoxically, made more difficult by the choice of outcome measure. A comparison of efficacy between drugs based on overall change in HDRS or MADRS score is therefore required before conclusions about differences can be drawn. SS declares that she has no conflicts of interest. PH received a scholarship in clinical pharmacology from Merck Sharp & Dohme in 2005. Ranking antidepressants – Authors' replyWe are aware of the likely biases in the data included in our study. As is often the case, we either have to make the best use of the available, if flawed, data or ignore it. As described in our paper, we took several steps to deal with these problems and outlined the methodological limitations of our analysis. Nonetheless, we are less nihilistic than several of the correspondents and we believe that valid conclusions can be drawn. Full-Text PDF