Fig. 1

Associations between baseline serum levels of B-cell activating factor of the tumour necrosis factor family (BAFF), anti-Jo-1 antibodies (α-Jo-1), and C-reactive peptide (CRP) (as independent variables), and creatine kinase (CK) (as dependent variable) estimated by multivariable analysis in LISREL software [26]. a Multiple linear regression (MR) with standardised regression (path) coefficients (β)^#. The mutual associations between independent variables (BAFF, anti-Jo-1, and CRP serum levels), found by MR, are depicted with two headed arrows and their influence on the dependent variable (CK) by one headed arrows. The non-significant path coefficient (p = 0.53) is marked as n.s. b Path analysis (PA) with path coefficients (β)^† and residual variances (Θ)* for all patients. One-headed arrows in PA represent significant direct effects, suggesting possible causal relationship. Only significant standardised path coefficients (p < 0.05) are shown. The best fit parameters of PA are shown below the diagrams: chi-square statistics (χ²), root mean square error of approximation (RMSEA), comparative fit index (CFI), and parsimony normed fit index (PNFI)