Pain trajectory, peaks, and regression to the mean in untreated patients with symptomatic knee osteoarthritis

<h2>Abstract</h2><h3>Introduction</h3> Pain remains the predominant symptom and primary treatment target in knee osteoarthritis (OA). <h3>Objectives</h3> This study aimed to document pain evolution patterns in untreated knee OA patients by assessing the prevalence and frequency of pain peaks and estimating regression to the mean. <h3>Methods</h3> Patients with symptomatic, radiographically confirmed knee OA (Kellgren-Lawrence grade ≥ 2), completed baseline questionnaires and rated their pain twice daily using a smartphone app over 60 days, enabling Ecological Momentary Assessment (EMA). Group-Based Trajectory Modeling (GBTM) identified pain trajectories over 30 and 60 days. Baseline differences and trajectory characteristics were compared, pain peaks analysed, and regression to the mean calculated as the difference between baseline and 2-month pain scores. <h3>Results</h3> Of 156 patients enrolled, 140 completed the assessment. Three trajectory patterns emerged: stable, decreasing, and increasing pain, comprising 69%, 22%, and 9% (30-day analysis), and 65%, 11%, and 24% (60-day analysis) of patients, respectively. Notably, 16% shifted from stable to increasing pain during the second month. The stable group showed fewer pain peaks and fluctuations, while the decreasing group had worse baseline scores. Pain peaks occurred in 40% of patients at a frequency of more than one every two weeks. Regression to the mean was significant only in the decreasing pain subgroup (-0.43 ± 3.16; p<0.001). <h3>Conclusions</h3> Pain in untreated knee OA follows three main short-term trajectories: stable (65%), decreasing (11%), and increasing (24%). Pain peaks are common, especially in unstable trajectory groups. Regression to the mean was observed only in patients with initially higher pain levels and decreasing trajectory.