Comparison of depressive and anxiety symptom networks by allergic rhinitis status: a nationwide population-based study using KNHANES

Introduction: Allergic rhinitis (AR) is a chronic inflammatory condition affecting a significant portion of the global population, with clinical evidence increasingly linking it to various psychiatric comorbidities, yet the structural relationship between its comorbid symptoms remains poorly understood[1]. Traditional psychiatric research has largely relied on total sum-scores, which obscure the complex, symptom-level dynamics. Network analysis provides a sophisticated alternative, conceptualizing mental disorders as a complex system of interacting symptoms[2]. Despite its potential, large-scale, population-based studies comparing the structural connectivity of these symptom networks based on allergic status are currently lacking. This study aimed to compare the connectivity and central symptoms of the anxiety-depression network between individuals with AR and non-allergic controls using a representative national sample. We hypothesized that AR status would significantly alter the network’s global strength and identify specific gateway symptoms that facilitate symptom-to-symptom contagion.

Methods: We conducted a cross-sectional analysis using data from the Korea National Health and Nutrition Examination Survey (KNHANES), assessing anxiety-depression symptoms via the Patient Health Questionnaire-9(PHQ-9) and Generalized Anxiety Disorder 7-item(GAD-7) scales. Participants were categorized into the AR group (physician-diagnosed) and a non-AR control group. Using propensity score matching, the AR group was matched with the control group in a 1:4 ratio based on age and sex. Gaussian Graphical Models (GGM) were estimated for both groups using the ‘bootnet’ package in R. The Network Comparison Test (NCT) was the primary statistical method used to evaluate differences in global strength and network invariance through 1,000 permutations. Additionally, bridge centrality measures and stability tests (case-dropping bootstrap) were conducted to calculate Centrality Stability (CS) coefficients, identify bridge symptoms between the anxiety-depression clusters, and ensure the robustness of the central nodes.

Results: Analysis revealed distinct network characteristics between the two groups. The Network Comparison Test demonstrated that the non-allergic group had significantly higher overall connectivity (Global Strength = 12.367) compared to the allergic group (Global Strength = 11.791) (p = 0.001). While network topology showed no significant difference (p = 0.40), ‘depressed mood’ (PHQ9_2) emerged as a critical bridge symptom with significantly higher betweenness centrality in the AR group. Specific node-level analysis revealed that the AR group exhibited lower strength and closeness centrality in several symptoms, particularly in sleep disturbances (PHQ9_3), fatigue (PHQ9_4), and excessive worry (GAD7_3), compared to the non-AR group. The stability analysis confirmed high reliability for the strength centrality index, with a CS-coefficient of 0.517 in the AR group, the non-AR group also maintained acceptable stability (CS-coefficient 0.361).

Conclusion: This nationwide study demonstrates that the mental health network of individuals with allergic rhinitis is characterized by lower global connectivity but a higher reliance on specific gateway nodes like ‘depressed mood.’ These factual findings suggest that while symptom contagion may be less intense in AR patients, targeted interventions on bridge symptoms are essential to prevent the escalation of comorbid anxiety and depression.

References

1. Ya-Kui Mou, Han-Rui Wang, Wen-Bin Zhang, Yu Zhang, Chao Ren, Xi-Cheng Song. Allergic Rhinitis and Depression: Profile and Proposal
2. Qiuge Zhao, Yuzhen Zhang, Lili Ji, Zhaoqian Pan. Network analysis of anxiety and depressive symptoms among patients with cardiovascular disease. BMC Public Health (2025);25:1085