OBJECTIVES: Myasthenia gravis (MG), a neuromuscular autoimmune disorder, causes fatigable muscle weakness. Diagnosing MG requires clinical evaluation and diagnostic tests such as autoantibody testing. Few real-world studies have explored the diagnostic process. This study examined the time between initial presentation of MG symptoms and MG diagnosis using real-world data.

METHODS: Recruitment began in August 2021 and is ongoing. Patients with MG enrolled in PicnicHealth’s research platform, consenting to collection of their medical records across U.S. health systems. Structured and unstructured data were abstracted from records using human-validated machine learning. Eligibility required MG diagnosis on or after January 1, 2010 and record of initial symptom presentation. Demographics were evaluated descriptively, and log transformed univariate linear regression models were used to examine the associations of demographics, type of presenting symptoms and antibody status with time from symptom presentation to diagnosis.

RESULTS: The 225 patients meeting eligibility criteria as of January 4, 2023 were 66% female and 78% white with a median age at enrollment of 52 (interquartile range: 39-66) years. Median age at symptom presentation was 58 (46-67) for males and 42 (30-54) for females. Most common presenting symptoms were diplopia (35%), ptosis (28%) and broad reference to symptoms of generalized MG (27%). Median time to diagnosis was 158 (16-909) days and did not significantly differ by sex, race, or type of symptoms (ocular vs generalized). 177 (79%) patients had AChR antibody data. AChR+ status was associated with significantly shorter time to diagnosis compared to AChR- (β = -1.24, p = 0.002; median days to diagnosis 103 vs. 450).

CONCLUSIONS: This study suggests that diagnosing MG in real-world clinical practice is a complex process, often taking >5 months from symptom presentation. Patients with AChR+ MG received more expedient diagnoses; surprisingly, patients with AChR- MG took an average of 11.5 months longer to diagnose.