Do Children with Immune Thrombocytopenia Have Cerebral Microbleeds?

Background Children with immune thrombocytopenia (ITP) have a low platelet count due to autoimmune destruction and reduced platelet production. ITP increases bleeding risk, with potentially devastating consequences. However, most children have mild disease that resolves spontaneously (Cooper 2014). For this reason, international guidelines recommend observation unless there is overt bleeding (Neunert et al. 2019; Provan et al. 2019). This has resulted in fewer patients exposed to treatment, but there are few data evaluating the safety and impact of this approach.

Patients with ITP often have petechiae associated with thrombocytopenia. We hypothesised that a similar process may be occurring in the brain parenchyma. Using susceptibility weighted magnetic resonance imaging (SW-MRI-a technique that is particularly sensitive to haemosiderin deposition), we have previously identified occult cerebral microbleeds (CMBs) in 43% of thrombocytopenic adults with ITP (Cooper et al. 2020). Studies in other disorders have associated CMBs with impaired cognitive function (Morrison et al. 2017). This method has not previously been used to evaluate thrombocytopenic children.

Methods 38 children aged 8-17 years old with a diagnosis of ITP and a platelet count under 30 x10⁹/L at any time during their disease course, underwent SW-MRI. Recruitment was from the paediatric immune haematology clinic at Imperial College Healthcare NHS Trust specialist ITP centre. SW-MRI were assessed by two neuroradiologists blinded to clinical details. Number, size and location of CMBs were correlated with clinical metadata including paediatric bleeding questionnaire (PBQ) score, validated cognitive testing (Cogstate), health-related quality of life (HRQoL) scores and fatigue scores. We compared imaging data to 20 age-matched healthy controls (siblings from another SW-MRI Imperial College study using the same imaging protocol).

Results The study cohort was 55% male with median age of 12.4 years (range 8-18 years). 66% had chronic ITP (lasting over 12 months) and 76% had active disease (platelet count below 100 x10⁹/L or were on treatment) at the time of MRI. Median disease duration was 33 months (range 1-129 months). Of the cohort, 73% had received treatment including: tranexamic acid, steroids, thrombopoietin-receptor agonists-TPO-RA, intravenous immunoglobulins (IVIg), rituximab and others, with a median treatment episodes of 3 and 68% are currently on TPO-RA treatment.

Of 38 patients, 5 (13%) had a detectable CMB (median 3 CMBs, range 1-10); no CMBs were detected in controls. All patients with CMBs had chronic ITP. The patient with the highest CMB burden had 10 CMBs in total (Figure 1). He had multi-refractory disease, 3002 days with a platelet count under 10 x 10⁹/L, a paediatric bleeding score of 14 and minimal responses to IVIg, prednisone, dexamethasone, rituximab, eltrombopag, romiplostim, sirolimus and mycophenolate mofetil.

There was a statistically significant correlation with higher PBQ score and the presence of CMBs (p=0.047 using the Mann-Whitney U test); no participants with a bleeding score under 4 had a CMB (PBQ median score 5, range 1-19); conversely many patients with a higher bleeding score did not have a CMB either (Figure 2). There was no correlation with disease duration and CMBs. Cognitive impairment, fatigue and reduced HRQoL were not associated with CMBs.

Discussion Although there is heterogeneity in this cohort, the high proportion of chronic ITP and treated patients reflects referrals to an ITP specialist centre. The proportion of children with a CMB is lower than in adults, however disease duration is shorter in this cohort. As none of the children with acute ITP or a bleeding score less than 4 had CMBs, this may support the practice of observing asymptomatic patients whilst awaiting spontaneous remission, although confirmation from a larger cohort is needed.

SW-MRI may also be valuable in stratifying patients with intermediate bleeding scores for therapy. In this study, CMBs were not associated with poor HRQoL, fatigue or cognitive impairment. A national study with differing centres is required to better understand the significance of this novel finding of CMBs in childhood ITP.

Laffan:Silence Therapeutics: Consultancy.