We have recently participated in a careful literature search and critical evaluation of glucocorticoids, and we have revised the side-effects data of four recent controlled trials of low-dose glucocorticoids (GCs) in rheumatoid arthritis. The toxicity profile stands out as remarkably more benign than expected from most textbook recommendations. Data regarding low-dose therapy are scarce and of low quality, as no controlled trials have been designed to specifically address toxicity. Common fears of GC toxicity seem to originate from an excessive weight on anecdotal data and observations with high doses, as in organ transplantation. There is now evidence that mechanisms of action of GCs vary considerably according to the dose, thus allowing the possibility of a different toxicity profile. Data from recent controlled trials are quite reassuring, overall. Certainly, risks and benefits of GCs need to be carefully weighed in every patient. But we need to make a clear distinction between established risks and unchecked fears while trying to get the best result for our patient. Clearly, there is a need for studies that are appropriately designed to address the toxicity of GCs and to avoid the risk of "throwing out the baby with the bath water."