Further qualification of a therapeutic responder index for patients with chronic low back pain.

Abstract

Previously, a preliminary patient responder index (RI) in chronic low back pain (CLBP) was developed and validated in 5 placebo-controlled clinical trials. The resulting RI was a > 30% improvement in CLBP and patient global assessment (PGA), and no worsening (< 20%) in the Roland Morris Disability Questionnaire (RMDQ) total score. Our objective was to provide further characterization of the preliminary RI in a trial with an active control. Data from a 6-week randomized, double-blind study of celecoxib compared to tramadol hydrochloride was analyzed to determine differences by treatment group on the CLBP RI and its components, to compare the CLBP RI with each of its individual components, and to reanalyze the original cutoff points for the responder criteria. Of the celecoxib arm, 50.7%, and of the tramadol hydrochloride arm, 43.7% were classified as responders under the CLBP RI (p = 0.043). The PGA is the most important component in the RI (45% of the sample failed to reach the > 30% improvement criteria on the PGA compared to 34% on the low back pain visual analog scale and only 11% on the RMDQ. The agreement among the CLBP RI with each of its 3 components was largest for the PGA component (κ coefficient 0.849) and smallest for the RMDQ component (κ coefficient 0.207). The RI appears to be particularly sensitive to the cutoff point used for improvement in the PGA component. Further testing of the index in trials with other active comparators is required to gain a fuller understanding of its performance.