BACKGROUND: Although several evidence-based analyses have found that in patients with moderate-to-severe psoriasis, biological agents for subcutaneous use show no difference in efficacy (no proof of difference), there is limited evidence on whether or not there is also proof of no difference. METHODS: The clinical material was the same as that reported in previous meta-analyses. Our methodology was a combination of meta-analysis and equivalence testing. The endpoint was the rate of Psoriasis Area and Severity Index (PASI) 75 achievement. The agents examined for equivalence testing included etanercept, high-dose etanercept, adalimumab, ustekinumab, and high-dose ustekinumab. The equivalence margin [±18 % as risk difference (RD)] was derived from statistical power information from the original trials; a more conservative margin at ±10 % was also tested. RESULTS: Our analysis involved 16 randomized trials including 8,257 patients. Ten head-to-head indirect comparisons were made and the respective values of RD were estimated. The 95 % confidence intervals for RDs remained within the margins of ±18 and ±10 % in six and two cases, respectively. A post hoc margin of about ±25 % was satisfied in all cases. All analyses assumed α = 2.5 %. CONCLUSION: Our study indicates that some of these five biological treatments can be equivalent, but results are strongly influenced by the margins adopted. Our findings can be helpful to develop local acquisition tenders on these drugs.
BACKGROUND: Although several evidence-based analyses have found that in patients with moderate-to-severe psoriasis, biological agents for subcutaneous use show no difference in efficacy (no proof of difference), there is limited evidence on whether or not there is also proof of no difference. METHODS: The clinical material was the same as that reported in previous meta-analyses. Our methodology was a combination of meta-analysis and equivalence testing. The endpoint was the rate of Psoriasis Area and Severity Index (PASI) 75 achievement. The agents examined for equivalence testing included etanercept, high-dose etanercept, adalimumab, ustekinumab, and high-dose ustekinumab. The equivalence margin [±18 % as risk difference (RD)] was derived from statistical power information from the original trials; a more conservative margin at ±10 % was also tested. RESULTS: Our analysis involved 16 randomized trials including 8,257 patients. Ten head-to-head indirect comparisons were made and the respective values of RD were estimated. The 95 % confidence intervals for RDs remained within the margins of ±18 and ±10 % in six and two cases, respectively. A post hoc margin of about ±25 % was satisfied in all cases. All analyses assumed α = 2.5 %. CONCLUSION: Our study indicates that some of these five biological treatments can be equivalent, but results are strongly influenced by the margins adopted. Our findings can be helpful to develop local acquisition tenders on these drugs.
Authors: Jeffrey S Heier; David M Brown; Victor Chong; Jean-Francois Korobelnik; Peter K Kaiser; Quan Dong Nguyen; Bernd Kirchhof; Allen Ho; Yuichiro Ogura; George D Yancopoulos; Neil Stahl; Robert Vitti; Alyson J Berliner; Yuhwen Soo; Majid Anderesi; Georg Groetzbach; Bernd Sommerauer; Rupert Sandbrink; Christian Simader; Ursula Schmidt-Erfurth Journal: Ophthalmology Date: 2012-10-17 Impact factor: 12.079
Authors: Nicolai Haase; Anders Perner; Louise Inkeri Hennings; Martin Siegemund; Bo Lauridsen; Mik Wetterslev; Jørn Wetterslev Journal: BMJ Date: 2013-02-15