Characteristics of eligible patients with knee osteoarthritis accepting versus declining participation in a randomised trial investigating the effect of weight loss versus knee arthroplasty to explore generalisability: A cross-sectional study.

BACKGROUND

The INtensive diet versus Knee Arthroplasty (INKA) trial is a randomised trial assessing weight loss as an alternative to knee arthroplasty (KA) in obese patients with severe knee osteoarthritis (OA) awaiting KA (NCT05172843). The external validity of the INKA trial may be hampered if the patients who participate differ from those who decline participation.

OBJECTIVE

To compare baseline characteristics between patients who enrol in the INKA trial and those who decline participation (i.e., non-INKA [nINKA] group).

METHODS

We applied a cross-sectional study design, collecting and comparing baseline characteristics among all patients eligible for enrolment in the INKA trial from two clinics in Copenhagen. Imbalance between accepting (INKA) and declining (nINKA) groups was assessed using standardised differences (StdDs). We were prespecified that StdD values < 0.20 would indicate a clinically insignificant imbalance between groups, whereas values > 0.80 indicate incomparability.

RESULTS

Of the 913 patients scheduled for KA, 888 were screened for INKA trial eligibility. Of the 217 eligible patients, 92 (42%) were enroled in the INKA trial, while 37 (17%) participated in the nINKA cross-sectional sample only. Patients enroled in INKA had on average a less severe Oxford knee score (OKS) of 22.0 (standard deviation = 6.7) compared to declining participants in nINKA with 18.6 (7.2), corresponding to an StdD of 0.50, and an absolute difference of 3.45 (95% confidence interval = 0.64-6.26, p = 0.017). A consistent similar pattern was noted across all secondary patient-reported outcomes applied in the INKA trial.

CONCLUSIONS

We observed discrepancies in patient-reported outcomes, with those who declined enrolment reporting more severe symptoms. These differences, however, were below the minimally important difference between groups for OKS, which is set to 4.84 points.

LEVEL OF EVIDENCE: Level II-III cross-sectional study in a randomised control trial.