Introduction : A higher risk of thrombosis in immune thrombocytopenia (ITP) as compared with the general population has been demonstrated.However, little is known about the risk factors for thrombosis in ITP, except splenectomy. The objective of this study was to assess the incidence of venous and arterial thrombosis in non-splenectomized ITP adults and to describe the baseline risk factors for thrombosis in nationwide populations of adult ITP patients in Sweden and in France.

Methods: The Swedish National Patient Register (NPR) was used to identify patients with a diagnosis of ITP from 2010 to 2014 in out- or inpatient care in Sweden. Code D69.3 from the tenth revision of the International Classification of Diseases (ICD-10) was used to identify adult patients (aged ≥18 years) with ITP. The NPR includes nationwide data for all hospital healthcare contacts covering details of hospitalizations such as date and duration of care, hospital/department name, surgical procedures and discharge diagnoses. Information on dispensed drugs was obtained from the Swedish Prescribed Drug Register, which includes information on all filled prescriptions from community pharmacies for the entire population. In France, the study was conducted in the database of the French national health insurance system (SNIIRAM). Its structure is very similar to the Swedish databases used in this study. It contains hospital and out-hospital health data for the entire French population, including drug exposure. We identified incident adult ITP patients from 2009 to 2012 using an algorithm combining long-term disease ITP code, hospital ITP code (D69.3 of the ICD-10) and exposure to ITP drugs. We assessed the incidence of venous and arterial thrombosis in non-splenectomized adult patients in both populations. Incident thrombosis was identified using a selection of ICD-10 codes as main diagnosis in hospital databases. The same codes were used in the two countries. Follow-up started from ITP diagnosis and ended in case of death, thrombosis, splenectomy or last data available (December 2014 in Sweden and December 2012 in France), whichever occurred first. The codes of ITP and of thrombosis used in this study have been validated, including the algorithm used to identify incident patients in the SNIIRAM. We also assessed the baseline frequency of thrombosis risk factors, including exposure to ITP drugs, to antiplatelet and antithrombotic drugs as well as to anti-hypertensive drugs and to statins.

Results: The Swedish cohort consisted of 3159 ITP adult patients and the French cohort of 3594 incident ITP adults. The incidence of venous thrombosis was 20.9/1000 person-years (95% confidence interval - CI, 18.16-24.0) in Sweden and 12.1/1000 person-years (95% CI, 9.5-15.4) in France. The incidence of arterial thrombosis was 22.5/1000 person-years (95% CI, 19.6-25.9) and 20.3/1000 person-years (95% CI 16.8-24.5), respectively. Mean age at ITP diagnosis was 57.9 years (standard deviation, 20.8) in Sweden and 57.4 (standard deviation, 21.7) in France; 53.9% and 56.7% were females, respectively; 5.0% and 0.3% had a history of venous thrombosis identified in the database, and 12.5% and 4.8% of arterial thrombosis, respectively. Exposures to drugs of interest during follow-up and before thrombosis or end of follow-up in the two databases were, respectively: statins, 29.3% and 22.0%; anti-hypertensive drugs, 44.3% and 48.1%; antiplatelet drugs, 28.6% and 20.9%; antithrombotic drugs, 27.2% and 22.5%; thrombopoietin receptor agonists (TPO-RAs), 4.8% and 10.4%. In France, 22.9% of the patients were exposed to intravenous immunoglobulin and 13.7% to rituximab (data not available in Sweden).

Conclusion: Incident arterial and venous thrombosis were not rare events in adult ITP patients in Sweden and in France. Risk factors for thrombosis were frequent in ITP patients with similar frequencies in both countries.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.