The paper by Meijer et al 1 compares the effects of inhaled fluticasone propionate (2 mg and 0.5 mg daily) and oral steroids (prednisolone 30 mg daily) in patients with mild to moderate asthma. Many patients with severe asthma are dependent upon corticosteroids, although inhaled steroids can effectively replace oral prednisolone.2

The biological effects of oral versus inhaled steroids can be compared in terms of lung function, airway responsiveness, and blood eosinophil number and activity, but suppression of the hypothalamic-pituitary-adrenal (HPA) axis is very difficult to assess. Prednisolone is so chemically similar to cortisol that most analytical methods (particularly radioimmunoassay, as used by Meijeret al) cannot distinguish between the two steroids. Although the authors attempted to overcome this problem, the correction of the measured level of cortisol for a potential cross reaction with prednisolone is not valid. If there is any possibility of cross reaction from other steroids, cortisol assays are only specific if performed using high performance liquid chromatography or mass spectrometry.3 It is therefore incorrect for the authors to conclude that cortisol levels after 30 mg oral prednisolone were comparable to those after inhaled fluticasone in a dose of 2 mg/day. The paper did not describe the precise dosages and timings for fluticasone or prednisolone administration and these could have large effects on 08.00 hour serum cortisol concentrations. The increased systemic effect of prednisolone compared with fluticasone is strongly supported by the significantly higher serum ECP level and blood eosinophil count.

Many papers on the function of the HPA axis in the context of safety of oral and inhaled steroids fail to take account of the normal function of the axis, the way in which the axis is perturbed by exogenous steroids, and the best methods for testing the axis, particularly when patients are also taking oral corticosteroids. Care should be taken at the outset to select analytical methods appropriate to the design of such studies to avoid misinterpretation.