# Hematological Indices as Potential Biomarkers of Disease Activity in Ankylosing Spondylitis: LASSO-Based Multivariable Modelling

## TL;DR

This study explores how blood-related measurements can help track disease activity in ankylosing spondylitis, finding that certain indices like PLR improve prediction when combined with traditional markers.

## Contribution

The study introduces a LASSO-based model showing that hematological indices like PLR improve disease activity prediction in ankylosing spondylitis.

## Key findings

- PLR is independently associated with active disease in ankylosing spondylitis.
- A model combining hematological indices achieved a good discriminative ability (AUC 0.77).
- White blood cell count and ESR also showed significant associations with disease activity.

## Abstract

Background and Objectives: Reliable laboratory markers that accurately reflect disease activity in ankylosing spondylitis (AS) are limited. Conventional acute-phase reactants do not consistently correlate with clinical activity. Composite hematological indices derived from complete blood count may better capture systemic inflammatory burden. In this study, we aimed to investigate hematologic parameters in AS and to assess their relationships with disease activity. Materials and Methods: This retrospective observational study included 196 patients with AS. Disease activity was defined as a Bath Ankylosing Spondylitis Disease Activity Index (BASDAI) ≥4. Demographic variables, laboratory parameters, hematological indices, and extra-articular manifestations were evaluated. Variable selection was performed using least absolute shrinkage and selection operator (LASSO) regression with ten-fold cross-validation. Variables with non-zero coefficients were entered into a multivariable logistic regression model. Model performance was assessed using receiver operating characteristic (ROC) curve analysis. Results: Ninety-seven (49%) patients had active disease. LASSO regression identified erythrocyte sedimentation rate (ESR), white blood cell count, red cell distribution width (RDW), platelet-to-lymphocyte ratio (PLR), and selected extra-articular manifestations as relevant predictors. In multivariable logistic regression, ESR (OR 1.03, 95% CI 1.00–1.06), white blood cell count (OR 1.23, 95% CI 1.04–1.46), and PLR (OR 1.01, 95% CI 1.003–1.020) were independently associated with active disease, while RDW showed a borderline association. The model demonstrated good discriminative ability (AUC 0.77, 95% CI 0.69–0.84). Conclusions: PLR is independently associated with disease activity in ankylosing spondylitis and improves discrimination when incorporated into a multivariable model. Easily accessible hematological indices may complement traditional inflammatory markers in the assessment of disease activity in routine clinical practice.