Immune thrombocytopenia (ITP) is a rare autoimmune disease due to platelet destruction and impaired megacaryopoiesis.1 Thrombopoietin receptor agonists (TPO-RA) are widely used as second-line treatment in ITP.2-4 Due to the presence of the thrombopoietin receptor on hematopoietic stem cells, the possibility of leukemogenic risk with TPO-RA is a concern.5 In 2013, a pharmacovigilance study in the Food and Drug Administration Adverse Event Reporting System database identified 62 spontaneous reports of acute myeloid leukemia (AML) among a total of 4,821 adverse drug reactions reported in patients with ITP between 2002 and 2011.6 An association between AML reports and the exposure to TPO-RA was found with an adjusted reporting odds ratio of 10.5 (95% confidence interval [CI]: 3.3-34.2) for romiplostim and 5.9 (95% CI: 1.9-18.3) for eltrombopag.
Because many factors influence the spontaneous reporting of adverse events, association between TPO-RA and AML must be assessed in a large cohort of ITP patients.7-9 This was the aim of the present study.
The data source was the FAITH (French Adult Immune Thrombocytopenia) cohort,10 built in the French National Health Insurance System Database. This database links socio-demographic, outpatient and hospitalization data for the entire French population (67 million individuals).11,12 We identified adult patients with ITP diagnosed between January 1, 2011, and December 31, 2018, using a validated algorithm based on the International Classification of Diseases, tenth version (ICD-10) D69.3 code recorded as hospital discharge diagnosis or long-term disease diagnosis (the latter is coded by general practitioners). This algorithm had a positive predictive value of 95.8% (95% CI: 92.8-98.8) in a previous validation study.13 Secondary ITP (Online Supplementary Table S1) and prevalent cases were excluded using a prior observation period of ≥1 year. The occurrence of AML after the diagnosis of ITP was identified using ICD-10 diagnosis codes (C92.0, C92.4-C92.9) recorded as hospital discharge diagnosis or long-term disease diagnosis up to December 31, 2018. The exposure to TPO-RA was identified using pharmacy dispensing data in the whole country. Only romiplostim and eltrombopag are available in France.
Due to the rarity of the event, we designed a nested case-control study to assess the association between AML onset and a previous exposure to TPO-RA. Cases and controls were matched (1:6) for age at ITP diagnosis (± 1 year), sex, region of residency, year of ITP diagnosis (± 2 years) and duration of ITP. Potential association was investigated through a conditional regression model adjusted for immunosuppressant exposure. In a sensitivity analysis, the patients with a diagnosis of myelodysplastic syndrome before ITP were also excluded because this condition may mimic ITP and increases the risk of AML.
This study was approved by the Institut des Données de Santé (Health Data Institute) in March 2012 (number 40) and the Commission Nationale de l’Informatique et des Libertés (French Data Protection Agency) in July 2012 (DE-2012-076). The study also received the European Network of Centers for Pharmacoepidemiology and Pharmacovigilance (ENCePP) study seal approval in October 2013.
Between 2011 and 2018, 8,172 adult patients with primary ITP were selected. Patient characteristics are described in Table 1. The median age was 63.2 years and 54.3% were women. The total follow-up time was 31,410 person-years. In total, 1,636 (20.0%) patients were exposed at least once to a TPO-RA: 1,238 (15.2%) to eltrombopag with a median cumulative exposure time of 5.7 months (up to 87.5 months), and 786 (9.6%) to romiplostim with a median cumulative exposure time of 6.6 months (up to 90.9 months). The follow-up time was similar between the 1,637 patients ever exposed to TPO-RA and the 6,535 patients never exposed to a TPO-RA (median: 3.7 and 3.8 years, respectively). In total, 265 (3.2%) patients were exposed at least once to an immunosuppressant (azathioprine: N=143, 1.8%; mycophenolate mofetil: N=90, 1.1%; ciclosporin: N=66, 0.8%) with a median cumulative exposure time of 6.9 months (up to 57.3 months).
During follow-up, 46 (0.6%) patients developed AML (median age of 74.8 years, 60.9% men). The 8-year cumulative incidence of AML was 1.0% (95% CI: 0.6-1.4; Online Supplementary Figure S1). The median time between ITP diagnosis and AML was 1.3 years (quartile [Q]1-Q3: 0.5-2.8). Thirty-six AML occurred in patients never exposed to TPO-RA (0.6%; 95% CI: 0.4-0.8) and ten among the patients exposed to TPO-RA (0.6%; 95% CI: 0.3-1.1): all ten patients had been exposed to eltrombopag, and three patients had been exposed to eltrombopag and romiplostim sequentially. Among these ten patients, the median age was 72.3 years (Q1-Q3: 67.8-80.0) and seven were men. The median time between ITP diagnosis and AML was 1.1 years (Q1-Q3: 0.4-2.3; min-max: 0.4-5.4). The median time between the first dispensing of TPO-RA and AML diagnosis was 8.9 months (Q1-Q3: 4.2-16.0; min-max: 1.7-62.8). The median cumulative duration of exposure to TPO-RA before AML was 3.9 months (Q1-Q3: 2.1-5.7; min-max: 1.7-12.7). After excluding the patients with a diagnosis of myelodysplastic syndrome before ITP, the 8-year cumulative incidence of AML was 0.8% (95% CI: 0.5-1.3; Online Supplementary Figure S2).
In the nested case-control analysis, the adjusted odds ratio was 1.0 (95% CI: 0.5-2.4; Table 2). The analysis excluding the patients with a history of myelodysplastic syndrome yielded similar results (Table 2).
The incidence of AML in adult patients with ITP has been demonstrated to be higher than in the general population in a recent epidemiological study in France and Denmark.14 However, this real-world nationwide pharmacoepidemiological study on 8,172 adult patients with primary ITP did not confirm the signal of increased risk of AML with TPO-RA. The short duration of TPO-RA exposure time among AML patients suggests that a causal relationship is unlikely. Conversely, more than one quarter of the patients treated with TPO-RA in the whole cohort were exposed for more than 1 year (up to 7 years), with no occurrence of delayed case of AML, which is reassuring Our study has some limitations. First, we cannot exclude residual misclassification of AML patients identified wrongly as ITP. However, at the time of this study, a systematic bone marrow aspiration (results not available in the database) was recommended by French ITP management guidelines in all patients aged >60 years, making these misdiagnoses unlikely.15 Moreover, the analysis excluding the patients with a previous diagnosis of myelodysplastic syndromes, a condition with a higher risk of AML and that can mimic or be associated with ITP, led to similar results. Second, due to a clear clinical and biological phenotype when AML occurs, an underestimation of AML during follow-up is unlikely, even if repeated bone marrow examinations are not recommended in patients treated with TPO-RA in France. Eventually, as few patients developed AML, we could not compare the risk between eltrombopag versus romiplostim. Finally, the AML frequency appears to be higher among patients exposed to immunosuppressants. However, it should be taken with caution because of the limited number of cases (4 AML patients previously exposed to ciclosporin and 1 to azathioprine). Conversely, 265 patients in the cohort were exposed to immunosuppressants (including 90 patients treated with mycophenolate). The durations of exposure up to 57,3 months are reassuring. Of note, an association between the exposure to immunosuppressants and the occurrence of AML or myelodysplastic syndrome has been demonstrated in primary autoimmune diseases with azathioprine, but not with mycophenolate.16 In our cohort study, no case of AML occurred following an exposure to mycophenolate, while this drug is used widely in some countries.4
Table 1.Characteristics of the overall cohort of primary immune thromocytopenia patients.
Table 2.Association of development of acute myeloid leukemia with previous exposure to thrombopoietin receptor agonists in adult patients with primary immune thrombocytopenia.
In conclusion, this study is reassuring regarding the possible risk of AML with TPO-RA in adult patients with primary ITP.
Footnotes
- Received November 10, 2025
- Accepted February 19, 2026
Correspondence
Disclosures
GM received meeting attendance grants from Amgen, Grifols and Novartis; he is a coordinator of research studies granted by Amgen, Argenx, CSL Behring, Grifols, Novartis and Sanofi; he participated in educational sessions funded by Amgen, Grifols and Novartis; and is on the advisory boards for Alpine, Amgen, Argenx, Grifols, Novartis, Recordati, Sanofi, Sobi and UCB. All other authors have no conflicts of interest to disclose.
Contributions
Funding
This study was academic and was funded by the Toulouse University Hospital.
Acknowledgments
We thank the Toulouse University Hospital legal team for the access to the SNDS.
References
- Cooper N, Ghanima W. Immune thrombocytopenia. N Engl J Med. 2019; 381(10):945-955. Google Scholar
- Provan D, Arnold DM, Bussel JB. Updated international consensus report on the investigation and management of primary immune thrombocytopenia. Blood Adv. 2019; 3(22):3780-3817. Google Scholar
- Neunert C, Terrell DR, Arnold DM. American Society of Hematology 2019 guidelines for immune thrombocytopenia. Blood Adv. 2019; 3:3-3866. Google Scholar
- Moulis G, Chen F, Carli G. Use of second-line and beyond maintenance therapies in adult patients with primary immune thrombocytopenia in Europe: a parallel study of six prospective multicenter national registries. Haematologica. 2025; 110(12):3094-3098. Google Scholar
- Liebman HA, Pullarkat V. Diagnosis and management of immune thrombocytopenia in the era of thrombopoietin mimetics. Hematology Am Soc Hematol Educ Program. 2011; 2011:384-390. Google Scholar
- Oshima Y, Yuji K, Tanimoto T, Hinomura Y, Tojo A. Association between acute myelogenous leukemia and thrombopoietin receptor agonists in patients with immune thrombocytopenia. Intern Med. 2013; 52(19):2193-2201. Google Scholar
- Weber J, Rainsford K, Velo G. Epidemiology of adverse reactions to nonsteroidal anti-inflammatory drugs. Adv Inflammation Res. 1984; 6:1-7. Google Scholar
- Pariente A, Gregoire F, Fourrier-Reglat A, Haramburu F, Moore N. Impact of safety alerts on measures of disproportionality in spontaneous reporting databases: the notoriety bias. Drug Saf. 2007; 30(10):891-898. Google Scholar
- Moulis G, Sommet A, Durrieu G. Trends of reporting of ’serious’vs. ‘non-serious’ adverse drug reactions over time: a study in the French PharmacoVigilance Database. Br J Clin Pharmacol. 2012; 74(1):201-204. Google Scholar
- Moulis G, Sailler L, Adoue D, Lapeyre-Mestre M. Pharmacoepidemiology of immune thrombocytopenia: protocols of FAITH and CARMEN studies. Therapie. 2014; 69(5):437-448. Google Scholar
- Bezin J, Duong M, Lassalle R. The national healthcare system claims databases in France, SNIIRAM and EGB: powerful tools for pharmacoepidemiology. Pharmacoepidemiol Drug Saf. 2017; 26(8):954-962. Google Scholar
- Moulis G, Lapeyre-Mestre M, Palmaro A, Pugnet G, Montastruc JL, Sailler L. French health insurance databases: what interest for medical research?. Rev Med Interne. 2015; 36(6):411-417. Google Scholar
- Mezaache S, Derumeaux H, Ferraro P. Validation of an algorithm identifying incident primary immune thrombocytopenia in the French National Health Insurance Database. Eur J Haematol. 2017; 99(4):344-349. Google Scholar
- Mannering N, Zadro Y, Hansen DL. Risk of cancer in adults with primary immune thrombocytopenia: a binational population-based real-world cohort study from Denmark and France. Haematologica. 2025; 110(11):2764-2773. Google Scholar
- Haute Autorité de Santé. Protocole national de diagnostic et de soins (PNDS) - Purpura thrombopénique immunologique de l’enfant et de l’adulte. 2017. Publisher Full TextGoogle Scholar
- Ertz-Archambault N, Kosiorek H, Taylor GE. Association of therapy for autoimmune disease with myelodysplastic syndromes and acute myeloid leukemia. JAMA Oncol. 2017; 3(7):936-943. Google Scholar
Data Supplements
Figures & Tables
Article Information
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.