Current Use of Androgens in Bone Marrow Failure Disorders: A Report from the Severe Aplastic Anemia Working Party (SAAWP) of the European Society of Blood and Marrow Transplantation (EBMT)

Anabolic steroids have been used for several decades in inherited and acquired bone marrow failure (BMF), due to their pleiotropic effects on erythrogenesis, telomere regulation and immune homeostasis. In some inherited disorders, mainly Dyskeratosis Congenita (DC) and Fanconi anemia (FA), they still represent one of the main non-transplant approaches, achieving satisfactory outcomes. In acquired setting, they have been used prior to the availability of anti-thymocyte globulin (ATG) and subsequently, in combination to it, or in refractory/relapsed patients with heterogeneous results. Albeit still in the armamentarium of possible therapeutic options for BMF, their place, optimal timing, and dosage have never been formally defined in aplastic disorders. Here we assembled the so far largest cohort of patients receiving androgens for an idiopathic or an inherited BMF, in order to track their current real-life use, indications, outcomes, and long-term effects in European and EBMT-affiliated centers.

We collected data on 274 patients receiving an androgen treatment between 1990 and 2021 across 82 centers. Patients were diagnosed with a severe/very severe BMF in 74% of the cases. Etiology was acquired in 71% (N=192) and inherited in 29% (N=78, missing N=4), with, in this last group, FA (88%) and DC (11%) being predominant, followed by fewer other indications (1%). Male/Female ratio was 1.5. Median age at the time of androgen start was 32 (18-52) years for acquired and 8 (6-12) for inherited BMF, while median time from diagnosis to first androgen use was 4 (0.3-17.6) and 8.5 (0.4-34.9) months respectively. In most cases (78%), patients were transfusion-dependent upon treatment initiation; 91% of inherited versus 54% of acquired cases received androgens before other therapies. In the idiopathic setting these compounds were administered as 2^nd ,3^rd or further line in 23%, 13% and 10% of the patients, respectively. As to the type of molecules, oxymetholone was most commonly used (48%), followed by danazol (32%), norethandrolone (11%), nandrolone (2%), testosterone (4%) and others (3%). Overall, median duration of androgen treatment was 9 months (IQR:3-26). In 60% of the cases a concomitant immunosuppressive therapy (mainly cyclosporine) was administered. Median follow-up was 77.6 months (95% CI 65.3 - 96). At six months after androgen treatment, complete and partial response rates were 13% and 37% in acquired and 4.5% and 40.9% in inherited BMF patients respectively. In total, 160 patients of this cohort (58%) received an allogeneic hematopoietic cell transplantation (HCT) at the median time of 18.4 months (IQR: 7-43.9) from androgen initiation. The 5-year post-androgen cumulative incidence (CI) of clonal events (leukemic transformation or evolution to secondary PNH) was 1% (95%CI 0-2%) for PNH, 3% (1-6%) for MDS and 1% (0%-3%) for AML. Five-year failure free survival (probability of being alive after androgen treatment, without switching to further lines, including HCT), was 13% (95%CI 6-21%) for inherited and 23% (95%CI 16-30%) for acquired BMF. Five-year overall survival after androgen initiation was instead 79% (95%CI 69-88%) and 64% (95%CI 56-71%) respectively. Taking advantage of long-term follow-up data available for 89 patients, we were also able to highlight androgen-related toxicities that tended to occur with variable post-treatment intervals: while liver, renal and gastro-intestinal toxicities were encountered early after treatment initiation (at a median of 3.4, 5.4 and 6.4 months respectively), endocrinological and psychiatric effects were more delayed (occurring after a median of 15.1 and 18.1 months respectively). Liver dysfunction was the most frequent type of side effect (10%) at 1-year, followed by gastro-intestinal (3%) and renal toxicities (2%). Psychiatric and endocrine side effects were rare, reported in 1% and 2% of the cases respectively at 3 years. Solid neoplasms were also rarely observed (CI of 1% at 3 years before other treatments). In total 3 patients developed a neoplasia at any time during the follow-up.

While still in use with a manageable level of toxicities, and low rate of clonal events and secondary neoplasia, this category of therapeutics remains associated with low response rates and its use should be carefully evaluated and reserved to selected cases both in idiopathic and inherited setting.