This secondary analysis used data from the BERLimmun cohort, a prospective registry including patients with MS, MOGAD, AQP4-NMOSD, isolated syndromes, and healthy controls. Eighteen clinical and patient-reported outcome measures (P(h)ROMs) covering cognition, mobility, fatigue, pain, mood, spasticity, and quality of life were analyzed. Variables with <30% missing data and <90% uniform responses were retained, standardized, and imputed. Pairwise Spearman correlations and hierarchical clustering were applied. Cluster-level composite scores were compared across diagnosis, sex, age, and disease-modifying therapy status to assess differential symptom patterns.

Hierarchical clustering of 193 items from 13 patient- and 5 physician-reported outcome measures identified 26 symptom clusters grouped into seven domains: motor–functional, cognitive–executive, mood–fatigue/psychosocial, sensory–pain, visual–oculomotor, autonomic/bladder–bowel, and composite/overlapping. PROMs largely separated from PhROMs (MSFC, BICAMS, SARA, MOT) except for the EDSS, indicating complementary coverage. Specific PROMs (WALK12, BPI, ABCD) showed high domain specificity, while general disability scales (HALEMS, EQ5D5L, PROMIS) were broadly distributed. Fatigue divided into mental and physical components, and psychological symptoms (depression, self-blame, isolation, motivation) were dispersed across clusters. Traditional fatigue scales (FSMC, FSS) did not capture physical fatigue differences across diagnoses, while sexual and bladder–bowel domains varied by diagnosis, age, and disease duration.

Hierarchical clustering of PROMs and clinical assessments revealed distinct yet overlapping symptom domains across MS, MOGAD, and AQP4-NMOSD. Item-level PROM analysis improved sensitivity to group differences and identified clinically relevant subsymptoms, supporting more precise disability assessment in neuroinflammatory diseases.