Longitudinal pain trajectories and biopsychosocial outcomes in refractory versus steroid-sensitive oral lichen planus: a retrospective cohort study

Oral lichen planus (OLP) can cause persistent pain and impaired quality of life, but its long-term symptomatic course remains insufficiently characterized. We compared longitudinal pain trajectories and endpoint patient-reported outcomes according to subsequent corticosteroid treatment-response classification. Of 346 patients meeting electronic medical record-based eligibility criteria, 285 completed the January–February 2026 survey and formed the analytic cohort. Longitudinal Visual Analog Scale (VAS) scores were analyzed using a linear mixed-effects model with patient-specific random intercepts and slopes. Refractory OLP (ROLP) classification was derived retrospectively from treatment responses documented during follow-up. Clinically significant oral candidiasis was modeled as a visit-specific time-varying covariate. A day-90 landmark sensitivity analysis classified early treatment response using objective lesion response alone and analyzed subsequent VAS measurements. Endpoint oral health-related quality of life and psychological symptoms were assessed using OHIP-14 and HADS. The median longitudinal clinical observation period was 3.0 years (IQR, 2.3–3.8), and 65 patients (22.8%) were subsequently classified as ROLP. VAS scores declined more slowly in subsequently classified ROLP than steroid-sensitive patients (Time × ROLP: β = 0.54; 95% CI, 0.29–0.79; p < 0.001). The corresponding interaction remained positive in the day-90 objective-response landmark analysis (β = 0.39; 95% CI, 0.11–0.67; p = 0.006). Clinically significant oral candidiasis was associated with higher concurrent VAS scores (β = 1.49; 95% CI, 1.23–1.75; p < 0.001) and occurred more frequently in subsequently classified ROLP than steroid-sensitive patients (46.2% vs. 16.4%). Higher pain burden, candidiasis episode count, and subsequent ROLP classification were associated with poorer OHIP-14 outcomes, while higher OHIP-14 scores were associated with higher HADS-A and HADS-D scores. Subsequent ROLP classification was associated with slower pain improvement and greater functional and psychological burden. The attenuated but persistent trajectory difference in the objective-response landmark analysis supports a longitudinal distinction between treatment-response groups, although residual confounding and incorporation in the full-period classification remain possible.