Chronic, high-dose, oral prednisone has been the mainstay of myasthenia gravis treatment for decades and has proven to be highly beneficial in many, toxic in some way to all, and not effective in a significant minority. No patient characteristics or biomarkers are predictive of treatment response leading to many patients suffering adverse effects with no benefit. Presently, measurements of treatment response, whether taken from clinician or patient perspective, are appreciated to be limited by lack of good correlation, which then complicates correlation to biological measures. Treatment response may be limited because disease mechanisms are not influenced by corticosteroids, limits on dosage because of adverse effects, or individual differences in corticosteroids. This review evaluates potential mechanisms that underlie lack of response to glucocorticoids in patients with myasthenia gravis.