ABSTRACT
Purpose: Hand osteoarthritis (OA) is more common in women. Hand OA incidence increases further in females around the age of 50, the typical age of menopause. Estrogen-deficient states are associated with increased musculoskeletal pain and inflammation and with increased rates of symptomatic OA. Estrogen replacement and selective estrogen receptor modulators (SERMs) can improve pain and structure in some pre-clinical models of OA associated with estrogen loss, and in exploratory analysis from hormone replacement therapy (HRT) trials. However, no randomised clinical trials (RCTs) of HRT had been performed in symptomatic OA populations, specifically hand OA. By carrying out a RCT feasibility study of a form of HRT (conjugated estrogens (CE)-bazedoxifene) in post-menopausal women with painful hand OA, we set out to determine the feasibility and acceptability of this. We also aimed to generate proof-of concept data on likely outcomes, calculate a sample size and refine methods for a full trial. Method(s): We recruited females aged 40-65 years and 1-10 years after final menstrual period with definite hand OA and >=2 painful hand joints across three primary/secondary care sites and from the community. Medical exclusions included those typical for clinical HRT use. Design was parallel group, double-blind 1:1 randomisation of CE-bazedoxifene or placebo, taken orally once daily for 24 weeks, then tapering for 4 weeks before study end at Week 28. Primary feasibility outcomes were rates of eligible participant identification, recruitment, randomisation, retention, compliance, and likelihood of unblinding. Adverse events (AEs) were collected. Secondary clinical outcomes included the anticipated primary outcome in a full trial of mean hand pain over 14 days prior to each visit, scored on a 0-10 numerical rating scale (NRS) where 10 is worst pain possible, as well as hand function, appearance and menopause symptoms. Progression criteria to a full RCT were: (i) recruitment >=30 participants across all sites in 18 months (or proportionate to time open);(ii) a drop-out rate of <=30% of randomised individuals;and (iii) acceptability to the majority of participants, including acceptable AE rates. All clinical outcomes were analysed on an intention-to-treat basis. Though not powered to detect a treatment difference, change and treatment effects (the difference in the outcome between the two groups) were indicated with 95% CIs, with all models adjusted for clinical subtype of painful hand joint, study site, and baseline values. The sample size for a full trial was estimated using the standard deviation (SD) of week 24 mean hand pain. Result(s): Due to the COVID-19 pandemic, the recruitment window was reduced to 12-15 months. From May 2019 to December 2020, 434 enquiries/referrals were received. Of 96 telephone pre-screens, 35 individuals were potentially eligible and of these, 33 gave consent to participate. Of the remaining, 250/401 (62%) were ineligible, whilst 55/401 (14%) chose not to proceed, with the most common reason being not wanting to take HRT. 28/35 (80% (95%CI 63%,92%)) eligible participants were randomised to study medication. All 28 participants completed all follow-ups with high compliance (100% active, 13/14[93%] placebo) and outcome measure completeness (100%, mean hand pain). All three AE-related treatment withdrawals were on placebo when unblinded. No serious AEs occurred. Participants/investigators were well blinded (participant blinding index 0.50[95%CI 0.25 to 0.75]). All three prespecified progression criteria were therefore met for a full trial. The treatment effect difference over 24 weeks in mean hand pain between active and placebo was -0.71 (95% CI -2.20 to 0.78) (Fig 1A). During tapering/stopping medication, mean hand pain increased by 1.31 points in the active arm compared with 0.17 in the placebo arm, indicating a possible effect of cessation of medication (Fig 1A). Furthermore, 6/13 (46%) participants in the active group reported worsening pain at week 28 compared with week 24, but only 2/12 (17%) were worse on withdrawing placebo (Fig 1B). The sample size for a full trial was estimated as 296 (based on MCID 0.8 on NRS, SD 2.0, 90% power, 10% drop-out, alpha 5%). Conclusion(s): This first study of a RCT of HRT for painful hand OA met its progression criteria, indicating that a full trial of an HRT in this population is feasible and acceptable. Although not powered to detect an effect, there was a trend towards improvement in hand pain on treatment and worsening of hand pain on tapering in the active arm only. This adds to proof-of-concept data in this area, justifying more work.ISRCTN12196200. Funded by Research for Patient Benefit programme, National Institute for Health Research (UK) PB-PG-0416-20023 [Formula presented]Copyright © 2023