Criterion Validity and Responsiveness of Estimated Cardiorespiratory Fitness Models in Patients with Inflammatory Joint Disease

Cardiorespiratory fitness (CRF) is an excellent marker of overall health. This study aimed to assess criterion validity and responsiveness of estimated CRF models (eCRF) in patients with inflammatory joint disease (IJD). CRF was measured directly as peak oxygen uptake (VO(2peak)) by a Cardiopulmonary Exercise Test (CPET), while one generic eCRF model (eCRF(GEN)) and two disease-specific eCRF models (eCRF(ALT) and eCRF(PGA)) were used to estimate CRF at baseline and after 3 months in 55 Norwegian patients with IJD. Moderate correlations were observed between eCRF(GEN), eCRF(ALT), eCRF(PGA), and VO(2peak) at baseline (ICC 0.60, 0.64 and 0.62, respectively) and 3 months (ICC 0.62, 0.65 and 0.57, respectively). All eCRF models overestimated measured VO(2peak), and there was large variability in agreement of individual measurements at baseline and at 3 months. Weak correlations were observed for responsiveness of eCRF(GEN) (ICC 0.39), eCRF(ALT) (ICC 0.40) and eCRF(PGA) (ICC 0.39). Mean differences between change in eCRF models and change in VO(2peak) were small, but the wide limits of agreement exceeded the pre-defined clinically acceptable margins. The eCRF models possessed adequate ability to detect ≥3.5 mL/kg/min improvement in VO(2peak). eCRF may suffice for group-level assessment, but caution is advised when applying eCRF to individual patients with IJD.