A Network Analysis of Panic Disorder, Agoraphobia, and Generalized Anxiety Disorder in 463 Patients From a...

Panic disorder, agoraphobia, and generalized anxiety disorder (GAD) frequently co-occur and share overlapping symptoms, yet it remains unclear whether they reflect distinct or interconnected symptom systems. This study examined the network structure of these disorders using clinician-administered diagnostic data.

A total of 463 adults completed the Mini International Neuropsychiatric Interview (M.I.N.I.) conducted by trained clinicians. Eighteen items from the panic disorder, agoraphobia, and GAD modules were retained after a multistep selection procedure ensuring clinical relevance, endorsement variability, and nonredundancy. A binary Ising network was estimated using eLASSO with EBIC model selection. Network accuracy, stability, and edge differences were evaluated through nonparametric bootstrapping.

The estimated network revealed two well-defined symptom clusters corresponding to (1) panic-agoraphobia and (2) GAD. Within the panic-agoraphobia cluster, physiological symptoms (e.g., palpitations, shortness of breath, sweating, dizziness) were tightly interconnected, and catastrophic cognitions (fear of dying, fear of losing control) were moderately linked to bodily sensations. Agoraphobia symptoms were strongly connected to each other but relatively peripheral to other panic symptoms. The GAD cluster was anchored by difficulty controlling worry, which emerged as the most central symptom and showed strong associations with restlessness, sleep disturbance, fatigue, and irritability. Notably, no direct edges were found between panic-agoraphobia and GAD symptoms, suggesting distinct anxiety systems.

These findings indicate that fear-based and worry-based anxiety symptoms form separable yet clinically relevant structures. Focusing on core processes like excessive worry and interoceptive regulation could enhance the specificity of interventions and more effectively disrupt anxiety maintenance mechanisms.