Longitudinal stability of molecular endotypes of knee osteoarthritis patients.

Abstract

Objective: To assess the longitudinal stability of biomarker-based molecular endotypes of knee osteoarthritis (KOA) participants from APPROACH and to evaluate the consistency of findings in an independent KOA population.

Methods: Nineteen biomarkers were measured longitudinally in 295 KOA participants from the APPROACH cohort. K-means clustering was used to identify the structural damage, inflammation, and low tissue turnover endotypes at the six-, 12-, and 24-month follow-ups. Endotype stability was defined as having the same independent endotype assignment longitudinally for patients with complete data (n = 226). Clinical and biochemical characteristics were compared between participants with longitudinally stable and unstable endotypes. The presence and longitudinal stability of the endotypes were evaluated in a different KOA population from the placebo arm of the oral salmon calcitonin trials.

Results: An average overall longitudinal endotype stability of 55% (Fleiss' Kappa of 0.53; 95% confidence interval [CI]: 0.46, 0.60) was demonstrated. An average stability of 59% (range: 54-59%) was observed for the structural damage endotype (Fleiss' Kappa 0.52; 95% CI: 0.45, 0.60), 54% (52-56%) for the inflammatory (Fleiss' Kappa 0.61; 95% CI: 0.53, 0.68), and 50% (49-52%) for the low tissue turnover endotype (Fleiss' Kappa 0.46; 95% CI: 0.39, 0.54). Participants with longitudinally unstable endotypes exhibited molecular properties of more than one endotype, which were detectable already at the first visit.

Conclusions: Our study showed for the first time that more than half of KOA participants exhibited a longitudinally stable endotype, highlighting the applicability of biomarker-based endotyping in a clinical trial setting.