Leinung and colleagues [1] recorded evidence of adrenal suppression by megestrol acetate. We have made similar observations in a study of cortisol function in 49 patients positive for human immunodeficiency virus (HIV) in Manchester, United Kingdom [2].

We assessed cortisol responses to the short-stimulation cosyntrophin test in HIV-positive patients who were most vulnerable to adrenal insufficiency, that is, those with CD4 lymphocyte counts of 50 cells/mm³ or less. Forty-nine patients were recruited (42 men, 7 women; age range, 25 to 56 years; CD4 count range, 4 to 50 cells/mm³; median CD4 count, 16 cells/mm³). Plasma cortisol levels were measured before and 30 minutes after a single intramuscular injection (except in one hemophiliac patient who received the compound intravenously) of 250 µg of tetracosactrin (cosyntropin) [3]. Clinical data (including concurrent medications) were recorded. Patients were excluded if they were already receiving steroid therapy with hydrocortisone or prednisolone but were enrolled if they were only receiving megestrol acetate. A normal test result was defined as a poststimulation level of 450 nmol/L (16 µg/dL) or greater [2, 3]. An abnormal result was defined as a poststimulation cortisol level of less than 350 nmol/L (12.5 µg/dL). Any intermediate result was considered an impaired response.

Of the 49 patients, 35 (71%) had a normal cosyntropin stimulation test result, 6 (12%) had an abnormal result, and 8 (16%) had an impaired response. For the analysis, we combined the abnormal and impaired groups. No significant differences (P> 0.2) were seen between the normal group (n = 35) and the abnormal or impaired group (n = 14) in the use of antiretroviral medication (21 [60%] and 8 [57%]), cotrimoxazole (18 [51%] and 8 [57%]), azole antifungal therapy (20 [57%] and 10 [71%]) or known inducers of cytochrome p450 such as rifampin c (7 [20%] and 1 [7%]; P> 0.2). No patient admitted taking opiates. Patients with an abnormal or impaired response, however, were significantly more likely to be taking megestrol acetate (dose, 40 to 80 mg/d; n = 7 [50%]) than were those with a normal response (n = 8 [23%]) (P = 0.05).

Our results agree with those of Leinung and colleagues. Not only may adrenal insufficiency occur after abrupt discontinuation of prolonged megestrol acetate therapy, but treatment with this drug may also make results of the short cosyntropin test difficult to interpret. We urge caution in the use of megestrol acetate in patients with late-stage acquired immunodeficiency syndrome (AIDS).