Nonlinear U-shaped association between serum uric acid and erythrocyte sedimentation rate in ankylosing spondylitis: a sex-stratified analysis of turning points

Abstrakts

Objectives Previous studies on the UA–inflammation relationship in ankylosing spondylitis have predominantly utilized traditional regression models assuming monotonic associations, masking non-linearities and sex-specific heterogeneity. This study aimed to characterize the sex-specific non-linear dose-response relationship between serum UA and ESR and to explore potential sex-specific turning points. Methods This retrospective study included 1,182 hospitalized AS patients. RCS models were employed to assess non-linear associations, with robustness validated via Bootstrap resampling. Multivariable logistic regression identified independent determinants of hyperuricemia; four progressively adjusted sensitivity models confirmed association stability; and sex-stratified analyses evaluated sex-specific heterogeneity. Results The overall cohort demonstrated a significant U-shaped UA–ESR association ( P for nonlinearity<0.001), with the fully adjusted turning point at 375 μmol/L (Bootstrap 95% CI: 365–386); the nonlinear model outperformed the linear specification (R 2 = 0.097 vs. 0.043). The turning point remained stable across all sensitivity models (375–390 μmol/L). Marked sex differences were confirmed ( P for interaction<0.001): males showed a robust U-shaped association (turning point: 371 μmol/L; R 2 = 0.081 vs. 0.021 for linear), whereas females showed a persistent linear negative correlation without a reproducible turning point ( P for nonlinearity=0.376). The U-shaped pattern was not replicated with CRP, which showed a purely linear inverse association ( P for nonlinearity=0.248), indicating that the rebound is ESR-specific. Conclusions This study demonstrates sex-specific nonlinear UA–ESR associations in AS, with a statistical turning point at approximately 371 μmol/L in males but no comparable turning point in females. These findings provide a hypothesis-generating framework for sex-stratified UA management in AS, pending prospective validation.