540 The Journal of Rheumatology 2005; 32:3 What WOMAC Pain Score Should Make a Patient Eligible for a Trial in Knee Osteoarthritis? JOYCE GOGGINS, KRISTIN BAKER, and DAVID FELSON ABSTRACT. Objective. To evaluate different Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) pain thresholds as eligibility criteria for a knee osteoarthritis (OA) trial and their effect on number of subjects recruited. Methods. We screened subjects with knee pain using the Likert version of the WOMAC and scored all subjects based on the severity of pain with each of the 5 WOMAC activities. For each of 4 alter- native definitions of eligibility, we tested how many subjects would be eligible for a trial. Results. Two hundred thirty-four subjects with chronic knee pain completed the WOMAC pain sur- vey. If we required a score of ≥ 4 and at least 2 activities with at least moderate pain, we found 128 of these subjects were eligible. If we required only one activity with moderate pain, the number increased to 139 (by 9%), and further to 161 (by 26%) if we required the same overall WOMAC score but no activity with at least moderate pain. The most common activity causing moderate or greater pain was going up or down stairs. Conclusion. The number of subjects recruitable for an OA trial depends on the WOMAC pain threshold required. Raising the threshold will lower the number of subjects modestly, but include more persons with moderate to severe pain. Lowering it may include many with only mild pain with activity. (J Rheumatol 2005;32:540–2) Key Indexing Terms: KNEE OSTEOARTHRITIS CLINICAL TRIALS TRIAL METHODS WESTERN ONTARIO AND MCMASTER UNIVERSITIES OSTEOARTHRITIS INDEX From the Boston University Clinical Epidemiology Research and Training Unit and the Arthritis Center, Boston University Medical Center, Boston, Massachusetts, USA. Supported by US National Institutes of Health grant AR47785. J. Goggins, MPH; K.R. Baker, PhD; D.T. Felson, MD, MPH. Address reprint requests Dr. D.T. Felson, Boston University School of Medicine, A203, 715 Albany Street, Boston, MA 02118, USA. E-mail: dfelson@bu.edu Submitted July 26, 2004; revision accepted October 13, 2004. Osteoarthritis (OA) is a common disease that lacks effective treatments. Most symptomatic OA requiring treatment involves the knee and hip, and investigators and industry are energetically developing and testing new treatments, many of which focus on relieving symptoms. As our understand- ing of the mechanisms of this disease improves, additional treatments will become possible. In trials of OA, the most commonly used outcome meas- ure to evaluate symptoms is the Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC) 1 . In a recent Medline search of knee OA trials published in 2002 and 2003 that focus on symptom improvement, we found that 26 of 42 trials (62%) used WOMAC to evaluate treat- ment response. WOMAC has been well validated 2,3 and appears to have better psychometric properties and scaling than other instru- ments, such as the Lequesne Index 4,5 . WOMAC is also more sensitive to change than generic instruments such as the Medical Outcome Study Short Form-36 6 . In trials of agents that alleviate symptoms of OA, WOMAC proved to be suf- ficiently sensitive to change to detect therapeutic effects. While WOMAC is now widely used to evaluate efficacy of treatments in OA drug trials, little guidance is available on whether or how to use WOMAC to determine eligibility for inclusion in a trial 7,8 . As recently suggested in rheuma- toid arthritis (RA), the outcome measure used to define effi- cacy should also be the measure used to determine a sub- ject’s eligibility for the trial 7,8 . If the instrument used to characterize eligibility is poorly correlated with WOMAC, eligible subjects may not obtain high enough WOMAC scores to show improvement. If the WOMAC is used to evaluate efficacy, it should also be used to define eligibility; but specific levels of WOMAC scores have not been evaluated for relevance in this regard, and we are unaware of published discussion of this issue. If the WOMAC threshold for eligibility is too high, few sub- jects will be eligible. On the other hand, admitting subjects with mild pain or low WOMAC scores, although increasing trial recruitment, might make it hard to detect improvement in patients who start with only minimal symptoms. Further, a trial in such persons may not produce results generalizable to those with more severe pain. Also, different compart- ments are affected by knee OA, and pain with climbing or descending may represent pain from the patellofemoral joint. Persons with pain only in this compartment would not be appropriate as subjects of treatments targeted to either the whole joint or the tibiofemoral articulation. This considera- tion adds complexity to the determination of eligibility, Personal non-commercial use only. The Journal of Rheumatology Copyright © 2005. All rights reserved. www.jrheum.org Downloaded on January 22, 2022 from