Abstract
Background: Clinical studies of Collagen Hydrolysate have suggested efficacy in decreasing symptoms and increasing joint function. Therefore a large multi-center, randomized, double-blind, placebo-controlled trial including 389 patients with osteoarthritis (OA) of the knee was performed with 6 centers in the United States (US), 3 in the United Kingdom (UK) and 11 in Germany (GER). Patients randomly received either identically packaged 10g Pharmaceutical Collagen Hydrolysate (PCH) or 12g lactose (placebo) for 24 weeks. Centers were used as strata for randomization. Acetaminophen (ACET) rescue up to 4g/day was allowed. Primary efficacy parameters were the differences of the WOMAC pain score, the WOMAC physical function score and the Patient Global Evaluation that were measured at the beginning and the end of the study period. No statistically significant differences between treatment groups were observed for the total study population. However, when individual countries were analyzed, a statistically significant treatment advantage of PCH over placebo was observed in all 3 primary efficacy endpoints in Germany (adjusted p-value < 0.05), but neither in the US nor the UK.
Purpose: To analyse country differences of the 3 primary endpoints using univariate and multivariate statistical methods.
Methods: Since the randomization was done within the centers, the randomization is still valid for a by country analysis. Firstly, the success of the randomization was investigated. Secondly, the influences of certain factors on the effect of the treatment were exploratorily analyzed for each primary efficacy endpoint by a univariate analysis. Thirdly, a multivariate linear regression analysis (for the two continuous primary efficacy endpoints WOMAC pain score and WOMAC physical function score) or a logistic regression (for the discrete primary efficacy endpoint Patient Global Evaluation) was performed.
Results: Randomization – No differences of age, gender, and race were determined between the treatment groups in all countries.
Univariate analysis – Effects of the placebo, drop-out rate, protocol violations, ACET consumption, and baseline differences in all primary efficacy endpoints show significant differences between Germany and the USA.
Multivariate analysis – The drop-out rate was the prognostic factors with the highest effect for all primary efficacy endpoints. Furthermore treatment, country, baseline differences, and ACET intake were also significant prognostic factors.
Conclusions:
-
Although a unique study protocol for all countries was used, the outcomes of PCH compared to placebo were different between the three countries.
-
In multi-national studies strategies for controlling possible cultural influences have to be developed.
-
Results of studies carried out in one country should only be cautiously transferred into another country.
Theses abstracts were prepared by Professor Roger Lemaire. Correspondence should be addressed to EFORT Central Office, Freihofstrasse 22, CH-8700 Küsnacht, Switzerland.
