Clinically Silent Hemorrhage in Children with Severe Immune Thrombocytopenia

Background: Assessing clinical bleeding in patients with immune thrombocytopenia (ITP) is problematic given the marked heterogeneity in bleeding manifestations. In particular, little is known about the frequency and clinical significance of occult hemorrhage in ITP. Prior publications have suggested the potential importance of assessing sites of clinically silent hemorrhage in ITP (Flores and Buchanan, Semin Hematol 2013), yet no such studies have been reported. Our aims were to prospectively explore the sites and frequency of occult hemorrhage (i.e., subclinical microbleeding) in children with severe ITP at diagnosis or upon symptomatic relapse.

Methods: This was a prospective, single-institution pilot study of children with newly diagnosed ITP, or persistent/chronic ITP undergoing a symptomatic relapse. Inclusion criteria included age ≤ 18 years and platelet count ≤ 10,000/mm³ at diagnosis or symptomatic relapse. Data collected included basic demographic information, platelet count, sites of bleeding, bleeding severity assessment (Buchanan and Adix, J Pediatr 2002), urinalysis with microscopic evaluation, fecal occult blood testing, and whether drug treatment was given at enrollment. Stool and urine samples were collected within 7 days from diagnosis or symptomatic relapse. Three months after diagnosis or relapse, a non-contrast brain MRI evaluated for brain microbleeds (hemosiderin deposits resulting from prior petechial hemorrhages). Outcome data summarizing the patients’ course was also collected. Results of urinalysis and fecal occult blood tests were defined based on the American Urological Association and American Gastroenterological Association Guidelines. A pediatric neuroradiologist reviewed the brain MRIs. Results are descriptive in nature including 95% exact confidence intervals. A one-sample proportion test was performed to compare the frequency of occult hemorrhage by site (urine, stool, CNS) with the incidence of overt hemorrhage in children with ITP reported in the literature. A Fisher exact test was performed to compare the relationships between bleeding severity grade and positive sites of occult hemorrhage. An alpha level of 0.05 was used.

Results: Between December 2011 and April 2014, 52 patients (mean age 7 years, range 1-16 years) were enrolled (40 at diagnosis and 12 upon relapse). Mean platelet count at enrollment was 4,000/mm³ (range 1-10,000/mm³). Clinical bleeding severity assessment at enrollment for epistaxis, oral and skin hemorrhage were mild to moderate for 24%, 50%, and 73% of subjects respectively. Severe hemorrhage at enrollment was reported for 8% and 14% of subjects for epistaxis and skin sites respectively. None of the subjects presented with clinically apparent hemorrhage in the stool, urine or CNS. Seventeen children (33%) were given drug treatment at enrollment. Of the 52 children enrolled, stool, urine and brain imaging was assessed in 28, 41, and 27 participants respectively. Positive fecal occult blood was identified in 21% (6.2% - 36.6%) of subjects, and microscopic hematuria in 27% (3.2% - 40.4%) of patients. CNS microbleeding was identified in just one subject (3.7%, 0.2%-21%), located in the superficial cortex of the left frontal pole. One-sample proportion test identified significantly increased frequency of occult hemorrhage in the stool and urine, but not in the CNS, when compared to clinically overt bleeding symptoms in the pediatric ITP population (P < 0.001, P < 0.001, P = 0.065 respectively). Fisher exact tests did not identify any relationships between overt bleeding by site (epistaxis, oral, skin) and positive sites of occult hemorrhage. Of the 14 children with occult hemorrhage in stool or urine at enrollment, only one (7.1%, 0.4%-36%) developed a new moderate or severe hemorrhagic event within 1 month.

Conclusion: On average, nearly 20% of bleeding in children with severe ITP is clinically silent and occurs with greater frequency than overt symptomatic bleeding. No significant relationship exists between bleeding severity grade and occult hemorrhage. CNS bleeding, likely in the form of minute cerebral capillary hemorrhage, infrequently occurs in this patient population, although the clinical significance of brain microbleeds remains uncertain.