Agranulocytosis associated with aminoglutethimide: pharmacological and marrow studies.

Aminoglutethimide (AG) inhibits steroid biosynthesis and the peripheral conversion of androgens to oestrogens (Santen et al., 1978; Dexter et al., 1967). Aminoglutethimide has proved to be an effective therapy in advanced breast cancer with response rates of 37.5-50% (Wells et al., 1978; Harris et al., 1983) and duration of response similar to adrenalectomy. However, blood dyscrasias have been reported in -1% of patients. We have seen two patients who developed severe agranulocytosis while taking aminoglutethimide and we describe possible mechanisms and predisposing factors.

A 62-year old woman presented with local recurrence of breast cancer and bone pain 2 years after primary treatment.
She started treatment with aminoglutethimide 250mg three times a day and hydrocortisone 20mg twice a day and 10 days afterwards she developed a skin rash. The rash faded after 5 days. A full blood count on the 28th day after starting showed agranulocytosis and no granulocytes were visible on a peripheral blood film. She developed a sore mouth and mouth ulcers which improved after 2 weeks. She continued aminoglutethimide and hydrocortisone and 3 weeks after the episode of agranulocytosis her peripheral blood film was normal. Marrow aspirated at the same time showed normal haemopoiesis and marrow infiltration with malignant cells. An abnormal alkaline phosphatase and yGT became transiently worse during the episode of agranulocytosis.
She continued on aminoglutethimide and hydrocortisone and had a complete regression of skin nodules and sclerosis of her lytic bone secondaries. Her remission lasted for 18 months.
A 50-year old woman was treated by radical mastectomy and adjuvant radiotherapy to the right chest wall and right supraclavicular fossa. She was started on Correspondence: A.L. Harris. *Present address: Dept. of Clinical Oncology, Regional Radiotherapy Centre, Newcastle General Hospital, Newcastle-upon-Tyne NE4 6BE. Received 10 January 1986; and in revised form 11 March 1986. adjuvant endocrine therapy with aminoglutethimide 250mg three times a day and hydrocortisone 20mg twice a day. Seven weeks later she had a fever, sore throat, felt generally unwell and had mild nausea. She was treated with cephalexin by her general practitioner with no improvement. A week later she had a low white cell count, total 0.8 x 1091-1; 34% granulocytes. She was seen in clinic after a further week and aminoglutethimide was stopped. The white cell count had started to improve while on aminoglutethimide (total count, 1.1 x 1091-1; 40% granulocytes). Marrow aspirate showed a hypocellular marrow with normal erythroid cells and normal megakaryocytes. There were some large early granulocytic cells present. Repeat marrow aspiration 3 weeks after recovery showed a cellular marrow with normal development of all cell lines.
Marrow was assayed for granulocyte/macrophage precursors (CFUc) in a semi solid colony assay (Barrett et al., 1976). Normal marrow and marrow from patient 2 was preincubated with plasma from patient 2 taken before starting aminoglutethimide, during and after the episode of agranulocytosis. The effects of a final concentration of 10% and 50% patient's plasma were studied on normal marrow and 50% patient's plasma on autologous marrow. The preincubation was for 1.5 h at 370C and cells were washed and then plated. Colonies and clusters were read after 10 days incubation. All assays were performed in triplicate.
Colony formation in marrow aspirated from patient 2 after recovery from agranulocytosis was very poor. Plasma containing aminoglutethimide suppressed colony formation further (Table I). There was a suppressive effect also on normal marrow, with 50% patient's plasma, while on aminoglutethimide.
Plasma levels were measured by reverse phase high pressure liquid chromatography after dichloromethane extraction, using 2 internal standards.
The levels for patient 1 were 0.4 Mg ml-1 aminoglutethimide and 4.2 ug ml -I N acetyl aminoglutethimide and for patient 2, 3.2 jug ml-1 Oestrone, oestradiol, testosterone and dehydroepiandrosterone sulphate (DHAS) were measured by radioimmunoassay using reagents in the WHO matched reagents scheme. The methods and assays have been described in detail (Harris et al., 1982;. Plasma hormones were similar to those observed in 45 postmenopausal patients and 17 premenopausal patients taking aminoglutethimide and hydrocortisone (data not shown).
Both patients had severe agranulocytosis, which recovered rapidly, and had normal platelet counts and haemoglobin. The only drugs they were receiving were aminoglutethimide and hydrocortisone.
An immune mechanism is unlikely because of the mild effects of the patient's plasma containing aminoglutethimide on normal marrow. Others have shown much more marked inhibition of CFUc formation in normal and autologous marrow in amidopyrine (Barnett et al., 1976), quinidine (Keltan et al., 1979) and phenytoin (Taetle et al., 1979) induced agranulocytosis, and in those cases the patient's serum was necessary for the effect. In other cases with quinine, amiodaquine and phenytoin, there was an increased sensitivity of the marrow to normal therapeutic plasma levels (Young & Vincent, 1980;Lind et al., 1973;Sutherland et al., 1977;Smith et al., 1977).
The most likely reason for the agranulocytosis from aminoglutethimide is a direct toxic effect on marrow with poor stem cell reserve. In a normal marrow, the effect could easily be compensated (Table I).
Pharmacokinetic differences were not evident in our patients, since plasma levels of aminoglutethimide and its acetylated metabolite were similar to levels in other patients, although the ratio of acetylated to parent compound was high in patient 2.
An unusual feature of both patients is that the granulocyte count had started to rise again while on the drug. This again suggests a direct effect that could be compensated by an increase in stem cell numbers. Gez and Sulkes (1984) reinstituted aminoglutethimide after agranulocytosis recovered and there was no repeated suppression of granulocyte count.
Another possible site of the adverse effect of aminoglutethimide may be the marrow fat cell. The growth of mammalian marrow in long-term continuous culture requires the presence of fat cells (Dexter et al., 1977). One of the effects of amino-glutethimide is to inhibit the aromatase enzymes that convert androgens to oestrogens in peripheral fat . Aromatisation activity is present in normal human marrow fat cells and is inhibited by aminoglutethimide in vitro (Frisch et al., 1980). The incidence of blood dyscrasia due to aminoglutethimide is 1%, since we have treated 228 patients with aminoglutethimide and only observed agranulocytosis in 2 patients. Lawrence et al. (1978) described one case of pancytopenia and they have treated 153 patients. Ragaz et al. (1984) found one severe case of thrombocytopenia and 2 mild cases (platelets >60 x 1091 1) in 141 patients (2%). However, compared to chemotherapy this risk is small and recovery is very rapid. The onset of agranulocytosis has been within 10 weeks of starting therapy, and it seems unlikely that routine blood counts would detect a trend in falling white count. Patients should be advised to report to their doctors if they develop sore throats, mouth ulcers or influenza-like symptoms.