Latent class analysis identifies distinct pain phenotypes in newly diagnosed systemic juvenile idiopathic arthritis.

BACKGROUND

Patients with systemic juvenile idiopathic arthritis (sJIA) exhibit highly heterogeneous pain manifestations, which significantly impact their quality of life and disease prognosis. An understanding of the pain phenotypes for this disorder and their influencing factors is crucial for individualized pain management.

OBJECTIVE

To explore the pain phenotypes of newly diagnosed sJIA patients via latent class analysis (LCA), analyse the influencing factors of these phenotypes, and evaluate the impacts of different pain phenotypes on short-term inpatient outcomes.

METHODS

A retrospective cohort study was conducted by collecting the electronic health records of 165 patients who were first diagnosed with sJIA at the Children's Hospital of Chongqing Medical University from January 2018 to July 2024. Patient pain characteristics, laboratory indicators, and inpatient outcome data were extracted. LCA was used to identify pain phenotypes, and multivariate logistic regression was used to analyse the influencing factors. The Lanza-Tan-Bray method and the data combination analysis technique were applied to evaluate the relationships between pain phenotypes and clinical outcomes.

RESULTS

LCA categorized the pain phenotypes of sJIA patients into three distinct classes, including (1) Class 1: inflammation-related moderate to severe pain with functional impairment (53.9% of patients); (2) Class 2: mild intermittent pain with extra-articular symptoms (19.4% of patients); and (3) Class 3: no joint pain with mild functional impairment (26.7% of patients). The analysis revealed that age (P = 0.023) and serum IL-10 levels (P = 0.047) were significant factors influencing pain phenotypes. Significant differences were observed among different pain phenotypes in terms of hospital stay duration, intrahospital department transfer rates, and pain status at discharge.

CONCLUSION

Pain in sJIA patients can be classified into three distinct phenotypes, which are influenced by factors such as age and IL-10 levels. The identification of these pain phenotypes has important clinical significance for developing individualized pain management strategies.