Predictors of exacerbation in myasthenia gravis after minimal symptom expression: a bicenter cohort study

Background Predicting exacerbation risk in clinically treated patients with myasthenia gravis (MG) is essential for personalised intervention with neurotherapies. This study investigated whether dynamic monitoring of anti-acetylcholine receptor antibody (AChR-ab) levels could predict exacerbation in MG patients following minimal symptom expression (MSE), and developed a validated model for risk stratification to inform treatment decisions. Methods We conducted a bicenter cohort study enrolling AChR-ab+ adult MG patients who achieved MSE. The derivation cohort included 339 patients from Huashan Hospital, and the validation cohort comprised 60 patients from West China Hospital. Exacerbation was defined as an increase of ≥ 2 points in the MG Activities of Daily Living score. Independent predictors were identified through univariate and multivariate logistic regression analyses, and a prediction model was subsequently developed and validated using discrimination and calibration metrics. Results Longitudinal AChR-ab levels strongly correlated with disease progression, with a median time to exacerbation of 35.1 months post-MSE. Independent predictors included disease duration to MSE achievement, AChR-ab change, thymoma, comorbid immune-related diseases, and history of myasthenic crisis. The model demonstrated excellent discrimination in derivation (AUC = 0.886) and validation cohorts (AUC = 0.829), with good calibration in both datasets. Conclusion Dynamic AChR-ab monitoring provides strong predictive value for MG exacerbation, and our validated prediction model offers clinicians a practical tool for personalized risk stratification and therapeutic decision-making in patients achieving disease stability.