We are pleased that our article has generated such interest. We welcome the thoughtful comments of Loprinzi and colleagues and the observations of Khoo and colleagues and Drs. Stoffer and Krakauer on the effects of megestrol acetate on the pituitary-adrenal axis. We cited the data of Loprinzi and associates [1] on suppression of cortisol levels in patients with cancer but had been unaware of their trial involving megestrol acetate and combination chemotherapy in patients with small-cell lung cancer [2]. It is interesting that they saw an apparent increase in the number of deaths caused by sepsis among patients receiving megestrol acetate. We agree that adrenal suppression may have played a role and feel it is prudent to give stress doses of corticosteroids to any patient receiving megestrol acetate. In addition, patients in whom megestrol acetate is abruptly withdrawn should be observed closely for, and warned about, adrenal insufficiency. Although reports of this reaction are rare, we believe it has been under-recognized, as shown in the case reported by Drs. Stoffer and Krakauer. The data of Khoo and colleagues imply that even megestrol acetate doses well below those recommended can induce adrenal suppression in patients with late-stage AIDS. We agree with Drs. Stoffer and Krakauer that this information should be included in sources such as Physicians' Desk Reference.

We did glucose tolerance testing so that we could have a biological marker of glucocorticoid activity other than pituitary responsiveness. The effects of megestrol acetate on blood glucose are somewhat confusing and may depend on the clinical state of the patient. Alexieva-Figusch and coworkers [3] found that a worsened glucose tolerance was associated with slightly decreased fasting glucose levels. Loprinzi and colleagues [1] found patients receiving megestrol acetate to have a statistically (but not clinically) significant decrease in fasting blood glucose levels after nearly 3 months of therapy (insulin levels were not measured). Although three of our patients showed worsening glucose tolerance, one showed no apparent change [4]. It appears that, as is the case with excess endogenous and exogenous steroids, individual responses to megestrol acetate vary widely. In addition, evidence from studies in rats suggests that progesterone may act directly on islet cells to stimulate hyperinsulinemia [5]. If this were true in humans, it could explain decreased glucose levels in the fasting state, whereas the peripheral (glucocorticoid) effects of insulin antagonism would account for altered results of glucose tolerance testing.