Sample size calculations for detecting disease-modifying osteoarthritis drug effects on the incidence of end-stage knee osteoarthritis in clinical trials: Data from the Osteoarthritis Initiative.

OBJECTIVE: We previously published data representing calculations for sample sizes assuming that the reduction of the incidence of knee joint replacement (KJR) would be an endpoint to prove efficacy of a disease-modifying drug in osteoarthritis (DMOAD). The sample sizes required for such hypothetical studies appeared to be high, rendering those studies unrealistic in the clinical research setting for practical reasons. The purpose of this work is to calculate sample sizes for hypothetical trials for DMOAD efficacy using a proxy for reaching end-stage knee osteoarthritis (esKOA) as an endpoint.

METHODS: Based on a sub-population of the Osteoarthritis Initiative, the cumulative incidence for both esKOA and KJR were calculated for a period of four years. The sample sizes of hypothetical DMOAD trials were then calculated for particular sub-cohorts of the OAI subpopulation, subdividing the groups according to age, Kellgren-Lawrence (KL) grades and gender.

RESULTS: Both the incidence for esKOA and for KJR over the four year period increase along with rising age, severity of OA and being female. The sample sizes to detect DMOAD efficacy are considerably smaller if reduction of the incidence of esKOA is chosen as an endpoint instead of reduction of the incidence of KJR.

CONCLUSION: In the future, generating health-economic data may become increasingly important to gain reimbursement. By choosing esKOA as an endpoint in DMOAD trials, we are able to show in our work that clinical trials in the field of OA are feasible, merely including a few hundred study participants per study arm.