Ghosal hematodiaphyseal dysplasia (GHD) is a rare autosomal recessive disease characterized by severe anemia and painful long-bone diaphyseal cortical endosteal hypertrophy.1 Some 40 cases have been reported, mostly in childhood. The disease results from biallelic deleterious variants in the TBXAS1 gene, disrupting TXAS protein function, causing thromboxane A2 pathway blockage and arachidonic acid metabolite accumulation.2,3 Standard care involves systemic corticosteroids, effectively alleviating hematopoietic and bone disorders. However, patients often require long-term corticosteroids, leading to iatrogenic complications.
Based on the thromboxane A2 pathway dysregulation, Brown et al.4 reported for the first time the good efficacy of cyclo-oxygenase (COX) 1 and 2 inhibition with non-steroidal anti-inflammatory drugs (NSAID) in two patients. Based on a similar hypothesis, four other non-related patients with GHD were treated in France from June 2019 with low-dose NSAID.
The French bone-marrow failure observatory (study’s ethic committee approval CLEA-2023-#312) was consulted to identify patients with biallelic pathogenic TBXAS1 variants, who underwent treatment with low doses of aspirin or NSAID.
During screening, four unrelated patients were identified among 1,857 patients included in the observatory. The characteristics of these patients and their disease are detailed in Table 1. Our four patients are male, with no medical history other than GHD. Patients #1, #3 and #4 had severe transfusion-dependent anemia in early childhood. Patients #1 and #3 achieved spontaneous red blood cell (RBC) transfusion independence without treatment for over 20 years. Patient #4 required long-lasting steroid treatment, to which he had a partial response and required occasional RBC transfusions. Patient #2 had no hematologic issues until the age of 24 years.
In the months preceding treatment with NSAID, all patients had severe anemia (median hemoglobin level 6.8 g/dL; range, 5.6-8.0 g/dL) and thrombocytopenia (median platelet count 85x109/L; range, 67-110x109/L), with normal neutrophil count, except for patient #4 (neutrophil count 0.7x109/L) and presented mild inflammation (median C-reactive protein 31 mg/L; range, 12.8-65 mg/L). Clinically, three were reported to have moderate obesity (patients #1, #2 and #3) and one had hepatosplenomegaly (patient #4). All had increased cortical density with diaphyseal involvement on bone X-rays, with a history of associated long-bone pain without skeletal deformity. Bone marrow aspirations failed, and biopsies revealed hypoplasia of the three myeloid lineages with dysmyelopoietic features, edema and fibrosis (Online Supplementary Figure S1). Investigations to rule out a myeloid malignancy were normal and genetic analyses for classical inherited bone marrow failure syndromes (Table 1) reported only the TBXAS1 variants. The genetic diagnosis was made at a median time of 15.5 years (range, 4-34 years) after the first occurrence of symptomatic anemia. Patients #3 and #4 had received high-dose steroids with a good response, but treatment was discontinued, due to relapse, after suspension and high-dose steroid requirement, respectively. Patient #1 had received hydrocortisone therapy at a dose of 100 mg/day with a weak response.
Before treatment with aspirin, patients #1, #2 and #4 required monthly RBC transfusions for over a year, while patient #3 was promptly treated with aspirin after the discontinuation of corticosteroids, due to worsening anemia, before requiring RBC transfusions. None of the patients required platelet transfusions.
The median age of the patients at the start of treatment with NSAID was 26 years (range, 23-40 years), corresponding to a median delay of 21 months (range, 1-48 months) from the first transfused anemia episode for patient #2, or from the relapse of anemia for the others. Aspirin was initiated at a dose of 75 mg/day, except for patient #3, who received 3 g/day (rheumatological dosage for rapid relief of bone pain).
All patients experienced a rapid hematologic response (Figure 1), with the hemoglobin level reaching >100 g/L without transfusion at a median of 43 days (range, 10-121 days), and were transfusion-independent at 1 month with resolution of bone pain. Their platelet counts exceeded 100x109/L in less than 4 months. The neutrophil count of patient #4 improved to over 1.0x109/L. The biological inflammatory syndrome initially resolved in all patients, concomitantly with improvement of the hematologic parameters.
During follow-up, patient #2 suddenly stopped the treatment for 2 months, resulting in a relapse of his anemia and inflammatory syndrome. A new complete hematologic response occurred 1 month after resumption of treatment, indicating an aspirin-dependence profile.
Patient #4 also relapsed on aspirin 75 mg/day, with the reappearance of an inflammatory syndrome followed by anemia. As suggested by Brown et al.,4 patients may have different residual TBXAS1 enzyme activity, and in some cases require a higher dose of aspirin. Therefore, a dose of 500 mg/day was introduced; the current follow-up is too short to assess response. Patient #4 was the only patient with a chronic disease evolution since childhood, associated with the emergence of steroid resistance and clinical hepatosplenomegaly, probably related to extensive bone marrow fibrosis. Another consideration, if improvement stalls despite dose escalation, is that the prolonged illness may have severely affected the bone marrow, hindering effective hematopoiesis through treatment.
Table 1.Patients’ characteristics and main outcomes.
With a median follow-up of 15.5 months (range, 9.5-52 months), patients #1, #2 and #3 sustained a complete hematologic response (on aspirin 75 mg/day for patients #1 and #2, and on a reduced dose of 2 g/month for patient #3, due to persistent remission).
Considering the spontaneous hematologic improvement that occurred during childhood in patients #1 and #3, we looked for a correlation between the phenotype and TBXAS1 genotype. Analysis of our patients and genetically confirmed cases in the literature3-10 (Online Supplementary Table S1) revealed three phenotypes, summarized in Table 2.
Firstly, there was a group of hematologically asymptomatic carriers, or with at worse moderate anemia, with no reported history of RBC transfusions (group 1, n=5). These subjects were diagnosed at a median age of 16 years, during family screening performed because of a symptomatic relative. Although they all had abnormal bone X-rays consistent with GHD, no bone symptoms were reported. Interestingly, most were female (80%).
The second phenotypic group (group 2, n=4) comprised patients with childhood onset of severe anemia, necessitating RBC transfusions (discovered at a median of 25 months), who achieved a spontaneous hematologic improvement after a mean time of 3.5 years (range, 1-6.5), allowing RBC transfusions to be stopped. Three patients relapsed after a prolonged hematologically asymptomatic period (12.5 to 28 years), presenting with severe transfused anemia with clinical and radiological bone manifestations and a biological inflammatory syndrome.
The last group (group 3, n=14) included patients with early severe and persistent/long-lasting transfusion-dependent anemia (median age at onset of 2 years), associated with thrombocytopenia and leukopenia in most cases. They all had radiological bone involvement, with half of them experiencing bone pain. Twelve patients were treated with high-dose corticosteroids as first-line therapy (the other 2 received NSAID as first-line treatment), with a good hematologic response in each. However, only 25% of these patients were able to stop steroids without relapse (data available for 8 patients with a short follow-up).
Figure 1.The main hematologic and inflammatory biological parameters during treatment of the four patients with Ghosal hematodiaphyseal dysplasia. (A-D) Hematological results. (E-H) Inflammatory results. The standards for a positive determination were set at >5 mg/L for C-reactive protein, represented by a dashed red line, and for fibrinogen at >4g/L, represented by the dashed blue line. The shadowed background represents the evolution of results under aspirin treatment. The green color indicates the initial aspirin dose, while the change in color corresponds to a change in dose. For patient #4, inflammatory results were missing at relapse and at the end of follow-up. RBC: red blood cell; CRP: C-reactive protein.
No genotype/phenotype correlations were identified (Online Supplementary Table S2). Inter- and intra-familial heterogeneity suggest unknown factors affecting phenotype, such as, possibly, female sex, which was more frequent in asymptomatic patients (group 1: 80% vs. groups 2+3: 20%). In groups 2 and 3 the appearance of anemia coincided with a biological inflammatory syndrome, which may reflect disease activity. External factors, such as infections, might also trigger disease worsening by generating inflammation and stimulating the arachidonic acid pathway, exacerbating metabolic problems caused by TBXAS1 mutations, and contributing to clinical heterogeneity. However, grouping patients is challenging, due to the heterogeneity of the population and the reported data, but also due to the lack of information on follow-up for asymptomatic patients and the impact of treatment on spontaneous outcome of the disease.
Our four GHD patients, treated with low-dose aspirin, confirm this treatment’s efficacy with three sustained hematologic responses. Our last patient, despite an initial response, relapsed and is currently being treated with a higher dose. Hypotheses to explain the loss of response were non-compliance, drug interaction and a more severe enzymatic defect. Considering the treatment’s excellent tolerability, especially its lower infection risk compared with corticosteroids, low-dose NSAID appear promising. However, data on the efficacy of NSAID in correcting radiological bone lesions are lacking, unlike steroid treatment, for which some pediatric studies have reported improvements. Our review of the literature highlights the clinical heterogeneity of GHD, including the possibility of long-lasting spontaneous hematologic improvement in childhood, which has been rarely reported so far. In most of these patients, anemia required RBC transfusions over a prolonged period, supporting treatment with low-dose NSAID as soon as transfusions are required. No genotypic or clinical predictive factors were identified to predict a spontaneous improvement, which would have helped us to discuss the discontinuation of treatment in some patients. For others, such as patient #2, who relapsed quickly after treatment was discontinued, suggesting NSAID dependency, lifelong treatment would seem necessary.
Lastly, our study highlights the importance of considering GHD in adults, as well as in children presenting with severe anemia and thrombocytopenia, especially in the case of abnormally tough bones, independently of myelofibrosis or radiological findings. Given the low toxicity of aspirin, TBXAS1 mutation should be added to the panel for inherited bone marrow failure screening, especially if transplantation is considered. Unlike other inherited bone marrow failure syndromes, no evolution to myeloid malignancy has been reported. While most reported cases involved children without long-term follow-up, five patients (four from our study) provided reassuring follow-up data into their fourth decade.
Table 2.Subgroup phenotype description.
Footnotes
- Received October 27, 2023
- Accepted February 28, 2024
Correspondence
Disclosures
No conflicts of interest to disclose.
References
- Ghosal SP. An unusual bone dysplasia. Indian J Pediatr. 1957; 24(12):359-368. Google Scholar
- Isidor B, Dagoneau N, Huber C. A gene responsible for Ghosal hemato-diaphyseal dysplasia maps to chromosome 7q33-34. Hum Genet. 2007; 121(2):269-273. Google Scholar
- Geneviève D, Proulle V, Isidor B. Thromboxane synthase mutations in an increased bone density disorder (Ghosal syndrome). Nat Genet. 2008; 40(3):284-286. Google Scholar
- Brown TJ, Barrett NA, Meng H. Non-steroidal anti-inflammatory drugs as a targeted therapy for bone marrow failure in Ghosal hematodiaphyseal dysplasia. Blood. 2023; 141(13):1553-1559. Google Scholar
- Jeevan A, Doyard M, Kabra M, Daire VC, Gupta N. Ghosal type hematodiaphyseal dysplasia. Indian Pediatr. 2016; 53(4):347-348. Google Scholar
- Joy P, Yoganathan S, Korula S. Ghosal hematodiaphyseal dysplasia and response to corticosteroid therapy. Am J Med Genet A. 2021; 185(2):596-599. Google Scholar
- Kim SY, Ing A, Gong S, Yap KL, Bhat R. Novel compound heterozygous variants of TBXAS1 presenting with Ghosal hematodiaphyseal dysplasia treated with steroids. Mol Genet Genomic Med. 2021; 9(3):e1494. Google Scholar
- Sharma R, Sierra Potchanant E, Schwartz JE, Nalepa G. Chronic steroid-response pancytopenia and increased bone density due to thromboxane synthase deficiency. Pediatr Blood Cancer. 2018; 65(1):e26777. Google Scholar
- Sudhakar M, Sharma M, Kandasamy S, Gummadi A, Rawat A, Vignesh P. Novel TBXAS1 variants in two Indian children with Ghosal hematodiaphyseal dysplasia: a concise report. Eur J Med Genet. 2022; 65(5):104498. Google Scholar
- Selina A, Kandagaddala M, Madhuri V. A recurrent biallelic pathogenic variant in TBXAS1 gene causing Ghosal hematodiaphyseal dysplasia. Indian J Pediatr. 2021; 88(4):381-382. Google Scholar
Data Supplements
Figures & Tables
Article Information
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.