Long COVID linked to myasthenia gravis exacerbation

For patients with myasthenia gravis, long COVID has been reported to be a risk factor for disease exacerbation in a recent study from China.

In a cohort of 180 patients (mean age 42.1 years, 65.6 percent female) with concurrent COVID-19 recruited at the Second Affiliated Hospital of Guizhou University of Traditional Chinese Medicine, myasthenia gravis exacerbation occurred in 34.4 percent during the 12-month follow-up. [Respir Med 2026;251:108554]

Disease exacerbation was severe in 43.5 percent, moderate in 51.6 percent, and mild in 4.8 percent. About 84 percent of the exacerbations were recorded within the first month post-COVID-19. Six deaths (3.3 percent) were documented, all of which were associated with premortem myasthenia gravis exacerbation. The median progression-free interval (ie, time from COVID-19 onset to the first occurrence of exacerbation) was 3 days.

Notably, myasthenia gravis exacerbation occurred more frequently among patients with vs without long COVID (64.6 percent vs 23.5 percent). Consistently greater proportions of patients with vs without long COVID exhibited disease exacerbations without full recovery at the 1-month (47.9 percent vs 12.9 percent), 6-month (43.8 percent vs 9.8 percent), and 12-month follow-up (35.4 percent vs 6.8 percent; p<0.001 for all).

Long COVID emerged as the strongest predictor of myasthenia gravis exacerbation (adjusted hazard ratio [aHR], 2.55, 95 percent confidence interval [CI], 1.47–4.42; p=0.001), followed by acetylcholinesterase (AChE) inhibitor use (aHR, 2.38, 95 percent CI, 1.39–4.06; p=0.002), and severe COVID-19 infection (aHR, 2.30, 95 percent CI, 1.30–4.10; p=0.004).

“Overall, our findings align with emerging evidence on long COVID’s systemic autoimmune sequelae, advocating for early antiviral therapies (eg, monoclonal antibodies) to mitigate prolonged immune activation in myasthenia gravis patients,” according to the authors.

“Mechanistically, SARS-CoV-2 may perpetuate immune dysregulation through cytokine storm, molecular mimicry, or persistent viral reservoirs as observed in other autoimmune and neurodegenerative disorders. In addition, autonomic and respiratory sequelae (eg, dysautonomia, reduced ventilatory reserve, sleep fragmentation) may increase physiologic load on bulbar and respiratory muscles, amplifying fatigability and precipitating clinical exacerbations,” they explained.

Good predictive performance

The authors developed applied machine learning (ML) models to evaluate the utility of long COVID for predicting myasthenia gravis exacerbations. Compared with other models (Ridge, LASSO, and Support Vector Machine), the random forest algorithm demonstrated the best performance, with an area under the curve (AUC) of 0.83 for predicting myasthenia gravis exacerbation after SARS-CoV-2 infection.

“This performance is on par with established ML predictors of acute COVID-19 severity, where random-forest models commonly reach AUCs of 0.80–0.83 for ICU admission or mechanical ventilation,” they noted. “It also exceeds the typical discrimination reported for autoimmune flare prediction (eg, systemic lupus erythematosus), where AUROCs of approximately 0.66 are frequently reported.”

The random forest ML model validates the primary findings and serves as a clinically actionable tool for early intervention, according to the authors.

“Decision curve analysis further confirmed their utility across threshold probabilities, highlighting translational potential in resource-limited settings. We recommend future external validation in multicentre datasets to assess transportability across care settings and viral variant periods,” they said.