Low- to medium-dose glucocorticoids have been shown to have not only anti-inflammatory but also disease-modifying properties in rheumatoid arthritis. The evidence for the benefit of its early use in combination with disease-modifying antirheumatic drugs underlines the need for a close evaluation of their risk–benefit ratio. Over time, numerous myths and fears about glucocorticoid toxicity in rheumatoid arthritis have arisen from observational studies, and many concerns have been unduly extrapolated from observations with higher-dose treatment. Furthermore, we cannot exclude the possibility of a powerful effect of bias by indication in these studies. Low- to medium-dose glucocorticoid regimens continued to be evaluated in randomized clinical trials, particularly in early disease, but these studies also have relevant methodological limitations in assessing safety, particularly due to small size and/or short duration. At present, the evidence on which to support clear recommendations about glucocorticoid toxicity remains remarkably weak. A large prospective pragmatic trial dedicated to the toxicity of low-dose glucocorticoids is dearly needed. Meanwhile, adherence to recommendations on standardized methodologies for registration and report of glucocorticoid adverse events is essential for improving our knowledge and competence in the best management of these important medications.