Background Myasthenia gravis (MG) is a rare and recurrent disease. The purpose of this study was to investigate the risk factors for relapse in MG patients after their first attack and establish a clinical predictive model. We conducted a retrospective study of 86 MG patients, followed and reviewed the clinical data of patients from the first onset to the first relapse, including age of onset, site of first symptom, MGFA at onset, thymoma, surgical resection of the thymoma, infection history, irregular drug use, combination of other autoimmune diseases, AChR antibody, and anti-Musk antibody, etc. The R software was used for statistical analysis. Univariate analysis and multivariate analysis were used to analyze risk factors. The clinical predictive model was established by Logistic regression analysis.
Results Within 2 years after the first attack, 61.2% of MG patients relapsed. MGFA at onset, irregular drug use and infection history were independent risk factors for MG relapse within 2 years after the first attack ( p < 0.05). The clinical prediction model has good discrimination and calibration.
Conclusion The relapse of MG is affected by a variety of factors. The clinical predictive model that was established in this study can help clinicians predict the probability of relapse in MG patients, identify early high-risk relapse patients, and serve for high-quality clinical management.