Background: Glucocorticoids (GCs) are commonly used for long-term medication in immunosuppressive and anti-inflammatory therapy. However, the data describing gluco-and mineralo-corticoid (MC) properties of widely applied synthetic GCs are often based on diverse clinical observations and on a variety of in vitro tests under various conditions, which makes a quantitative comparison questionable. Method: We compared MC and GC properties of different steroids, often used in clinical practice, in the same in vitro test system (luciferase transactivation assay in CV-1 cells transfected with either hMR or hGRa expression vectors) complemented by a system to test the steroid binding affinities at the hMR (protein expression in T7-coupled rabbit reticulocyte lysate). Results and Conclusions: While the potency of a GC is increased by an 11-hydroxy group, both its potency and its selectivity are increased by the D1-dehydro-configuration and a hydrophobic residue in position 16 (16-methylene, 16a-methyl or 16b-methyl group). Almost ideal GCs in terms of missing MC effects, as defined by our in vitro assay, are therefore prednylidene, budesonide, beclomethasone and betamethasone. The MC potency of a steroid is increased by a 9a-or a 6a-fluoro substituent. A hydrophilic substituent in position 16 (like 16-hydroxylation in triamcinolone) decreases both MC and GC properties. As no substituent that leads to an isolated reduction of GC activity could be characterized in our experiments, 9a-fluorocortisol, the most frequently used steroid for MC substitution, seems to be the best choice of available steroids for this purpose.