High annualized bleeding rates in pediatric patients with inherited platelet function disorders

Platelet function disorders are common inherited bleeding disorders.^1,2 While severe inherited platelet function disorders (IPFD) such as Glanzmann thrombasthenia and Bernard Soulier syndrome are easier to diagnose, diagnosing milder phenotypes of IPFD remains complex. Up to 9% of healthy children can have epistaxis, and 12% of infants in the general pediatric population demonstrate bruising, complicating the ability to diagnose IPFD in otherwise healthy children.^3,4 Structured bleeding assessment tools (BAT), such as those developed by the International Society of Thrombosis and Hemostasis (ISTH), are useful as screening instruments to guide testing for bleeding disorders.^5,6 Our group has previously established an IPFD incidence of 4.5% in an unselected, prospective cohort of adolescents referred for heavy menstrual bleeding (HMB) using ISTH-BAT and systematic testing for IPFD.⁷ We aimed to determine the prevalence and types of IPFD in children referred for bleeding, examine bleeding events and annualized bleeding rates (ABR), and investigate the predictive ability of the ISTH-BAT for IPFD diagnosis, treatment for bleeding events, and severe bleeding events. We hypothesized that bleeding scores (BS) are higher in pediatric patients with IPFD than those without, and that the ISTH-BAT would predict bleeding events and receipt of treatment in IPFD.

Our retrospective study, approved by the institutional review board, included patients aged 0 to 18 years referred for bleeding symptoms between 2019 and 2021 to The University of Texas Southwestern Medical Center (UTSW) Pediatric Hemostasis and Thrombosis Program. A standardized hemostatic evaluation is undertaken for every patient with high suspicion for primary hemostatic defect. The minimum standardized evaluation includes a complete blood count with mean platelet volume, blood smear, prothrombin time, activated partial thromboplastin time, fibrinogen activity, von Willebrand panel, and platelet aggregation using whole blood impedance aggregometry (WBA). Additional testing is individualized and may include clotting factor assays, evaluation for disorders of hyperfibrinolysis, platelet flow cytometry, and/or genetic testing. To be eligible, we included patients who completed more than one clinic visit and, either had an ISTH-BAT documented in the medical chart or sufficient bleeding history captured to calculate the ISTH-BAT, and underwent the minimum standardized hemostatic evaluation. Only patients with reproducible (≥2) abnormal WBA findings were included. We excluded patients without documented bleeding histories, those with anticoagulant-associated bleeding, or those with additional causes of increased bleeding risk.

Demographic and clinical information from each patient’s electronic medical record was recorded. IPFD was diagnosed when there was impaired aggregation to ≥2 agonists (excluding collagen or ristocetin) and/or impaired ATP secretion on ≥2 occasions using WBA and luminescence as previously described.^7,8 von Willebrand disease (vWD) was defined as quantitative vWF levels <0.50 IU/dL (vWF antigen, vWF:RCo or vWF:GP1bM) on ≥2 occasions. All patients met Type 1 vWD (vWD-1) criteria based on previously established guidelines.⁹ Controls had a bleeding phenotype but no identifiable bleeding disorder on comprehensive laboratory evaluation. A total of 193 participants (67% female) with a median age of 13 years (interquartile range [IQR]: 7-15) were included; 18% (N=34) were diagnosed with IPFD, 24% (N=46) with vWD-1, and in 58% (N=113) the bleeding disorder evaluation did not reveal an identified hemostatic defect (Table 1). Of the IPFD participants, 12 (35%) had aspirin-like defect, one (3%) had Bernard Soulier, and in the remaining 20 (59%), the identified defect was non-specific (reduced aggregation and/or secretion to ≥1 agonist).

Table 1.Patients� characteristics.

Annualized bleeding rate was calculated by dividing the number of clinically significant episodes one year before the first evaluation by history and/or emergency department, doctor’s surgery, or hospital visits for bleeding. Median ABR was 12 (IQR: 2-13) for IPFD, 12 (IQR: 4.5-14) for vWD-1, and 12 (IQR: 1-13) for controls (P=0.13) (Figure 1A). The most frequent bleeding event in all groups was HMB (N=96, 48.2%). The subsequent most frequent bleeding event in the IPFD cohort was surgical bleeding, while in vWD-1 and controls, it was epistaxis (Online Supplementary Figure S1A).

Severe bleeding events, defined as any bleeding event requiring medical or surgical interventions including high-dose oral hormonal therapy every 6 or 8 hours/day for several days, intravenous estrogen, blood transfusion, and surgical interventions to stop bleeding, occurred in 18 (53%) IPFD, 20 (43%) vWD-1, and 50 (44%) control participants. The median number of bleeds requiring treatment was 2 (IQR: 1-12) for IPFD, 12 (IQR: 1-13) for vWD-1, and 3 (IQR 0-12) for controls (P=0.16) (Figure 1B). Receipt of treatment was defined as any bleeding requiring systemic therapy or surgical intervention for bleeding symptoms, but excluding topical agents (i.e., oxymetazoline for epistaxis). Hormonal therapy was the most utilized treatment in vWD-1 (N=26; 43%) and controls (N=56; 48%), whereas surgical therapy, including packing, suturing, and cauterization, was highest in IPFD (N=14; 33%) (Online Supplementary Figure S1B).

Median BS was 3 (IQR: 2.75-4) in the IPFD group, 4 (IQR: 3-5) in vWD-1, and 3 (IQR: 3-4) in controls (Figure 1C). BS between the vWD-1 and control groups varied significantly (P=0.012), while no significant difference was found between IPFD and vWD-1 or controls. Clinically significant bleeding events were defined as bleeding involving all sites but excluding bleeding lasting <10 minutes, events not requiring intervention, and all cutaneous bleeding except for wound bleeds lasting >10 minutes. In all study groups, median BS was significantly higher in those with clinically significant bleeding events than in those without (P<0.05) (Figure 2). A BS >4 in IPFD (area under the curve [AUC]=0.85) and vWD-1 (AUC=0.85) groups predicted clinically significant bleeding events with very good accuracy (Online Supplementary Figure S2A). For bleeding events requiring treatment, a BS >3 in the IPFD group (AUC=0.95) predicted treatment with excellent accuracy, while a BS >2 in the vWD-1 group (AUC = 0.71) predicted treatment with good accuracy (Online Supplementary Figure S2B). Severe bleeding events were predicted at BS >7 in the IPFD group (AUC=0.89) with very good accuracy and BS >5 in the vWD-1 group (AUC=0.62) with sufficient accuracy (Online Supplementary Figure S2C). The likelihood of clinically significant bleeding events, bleeds requiring treatment, and severe bleeding events were significantly higher in the IPFD and vWD-1 groups when the BS was above the respective identified cut-off values, except for severe bleeding in vWD (Online Supplementary Figure S3). We show that IPFD is common among pediatric patients referred to a tertiary care center, with 18% of our participants diagnosed with IPFD. Varying incidence seen in other studies, such as 5% by Bidlingmaier et al.¹⁰ and 21% by Adler et al.,¹¹ can be explained by differences in sample size, prospective nature, and different methods of testing platelet function. We demonstrate high ABR in IPFD participants that were not significantly different than those with vWD-1. Gresele et al.¹² found higher bleeding rates in IPFD than vWD patients, but once patients with Glanzmann thrombasthenia and Bernard Soulier were removed, similar to our study population, no significant differences remained. A higher proportion (20%) of IPFD participants suffered surgical bleeding events compared to vWD-1 and required more invasive interventions, such as surgery. This is in stark contrast to the data provided by Gresele et al., which reported <5% of pediatric IPFD bleeding events to be surgical, likely explained by the prospective nature of the study where IPFD diagnosis was already established and early intervention possible.¹²

Figure 1.International Society of Thrombosis and Hemostasis Bleeding Assessment Tool Bleeding Score, annual bleeding rate, and bleeds requiring treatment among study population. (A) Annual bleeding rate (ABR), (B) bleeding events requiring treatment, and (C) International Society of Thrombosis and Hemostasis Bleeding Assessment Tools Bleeding Score between study group participants. Data are presented as medians and interquartile ranges. **P<0.01. vWD: von Willebrand disease; IPFD: inherited platelet function disorder; yr: years. Controls had a bleeding phenotype but no identifiable bleeding disorder on comprehensive laboratory evaluation.

Figure 2.International Society of Thrombosis and Hemostasis Bleeding Assessment Tool Bleeding Score between participants with and without clinically significant bleeding events in each study group. Data are presented as medians and interquartile ranges. *P<0.05; **P<0.01. ISTH-BAT: International Society of Thrombosis and Hemostasis Bleeding Assessment Tools; vWD: von Willebrand disease; IPFD: inherited platelet function disorder. Controls had a bleeding phenotype but no identifiable bleeding disorder on comprehensive laboratory evaluation.

ISTH-BAT scores were comparable for IPFD and vWD-1 participants. Gresele et al. found a significantly higher bleeding score in pediatric IPFD versus vWD-1 participants (8 vs. 4, respectively).¹² However, Bidlingmaier et al. found no difference in BS, with both pediatric IPFD and vWD-1 participants having a BS of 3.¹⁰ This is likely driven by the inclusion of severe platelet disorders in the Gresele study, whereas the Bidlingmaier study included a similar cohort of IPFD as ours, leading to comparable results.

For our IPFD participants, BS >4 predicted subsequent bleeding events, BS >3 predicted bleeding events requiring treatment, and BS >7 predicted severe bleeding events. Our study predicted bleeding events at a lower BS than Gresele et al. (BS >7), likely due to higher ABR within our study population.¹² Gresele et al. found that 75% of adult IPFD patients with a BS >2 required treatment, but did not comment on the ability of ISTH-BAT to predict treatment requirements.¹³ Gresele et al. found that a BS >6 was predictive of severe bleeding events in adult IPFD participants;¹² their inclusion of severe bleeding disorders, such as Glanzmann thrombasthenia and Type 2 and 3 vWD, likely increased their ability to predict severe bleeding events at lower bleeding scores.

In summary, IPFD is common in children referred for bleeding and characterized by high ABR. IPFD presentation overlaps with children without an identified bleeding disorder and vWD-1. The ISTH-BAT alone was not discriminatory for IPFD in our cohort, highlighting the need for platelet aggregation testing in children referred for bleeding. Patients with high ISTH-BAT scores at IPFD diagnosis are more likely to experience severe bleeding events and require treatment. While HMB was the most common cause of bleeding in all cohorts, IPFD participants received more surgical interventions compared to vWD-1 and controls, whose most common treatment was hormonal therapy; as such, HMB was undertreated in IPFD. Our results underscore the importance of high diagnostic suspicion for IPFD in children referred for a bleeding disorder evaluation. Further, we confirm the utility of the ISTH-BAT for IPFD and bleeding prognosis in pediatric patients.

Footnotes

• Received January 7, 2024
• Accepted March 19, 2024

Correspondence

A. ZIA - Ayesha.zia@utsouthwestern.edu

Disclosures

No conflicts of interest to disclose.

Contributions

SS and AZ conceptualized, designed, and performed research. SS, SZ and AZ analyzed the data. SY and RS blindly and independently adjudicated abnormal whole blood aggregation results for all participants. SS and AZ wrote, and all other authors critically edited the manuscript.