Bleeding phenotype and diagnostic characterization of patients with congenital platelet defects

Abstract Phenotypic characterization of congenital platelet defects (CPDs) could help physicians recognize CPD subtypes and can inform on prognostic implications. We report the analyses of the bleeding phenotype and diagnostic characteristics of a large cohort of adult patients with a confirmed CPD. A total of 96 patients were analyzed and they were classified as Glanzmann thrombasthenia, Bernard‐Soulier syndrome, dense granule deficiency, defects in the ADP or thromboxane A2 (TxA2) pathway, isolated thrombocytopenia or complex abnormalities. The median ISTH‐BAT bleeding score was nine (IQR 5‐13). Heavy menstrual bleeding (HMB) (80%), post‐partum hemorrhage (74%), post‐operative bleeds (64%) and post‐dental extraction bleeds (57%) occurred most frequently. Rare bleeding symptoms were bleeds from the urinary tract (4%) and central nervous system (CNS) bleeds (2%). Domains with a large proportion of severe bleeds were CNS bleeding, HMB and post‐dental extraction bleeding. Glanzmann thrombasthenia and female sex were associated with a more severe bleeding phenotype.


| INTRODUCTION
Congenital platelet disorders (CPDs) are rare bleeding disorders caused by congenital defects in platelet production or platelet function.
Patients typically present with a mucocutaneous bleeding tendency.
Common symptoms include epistaxis, unexplained or extensive bruising, oral cavity bleeds, heavy menstrual bleeding (HMB) and bleeding following a hemostatic challenge such as surgery, dental extraction and childbirth. 1 So, CPDs are clinically heterogeneous; the frequency and severity of symptoms vary greatly among different types of CPDs, among patients with the same disorder and within patients over time. 2 The bleeding phenotype can be evaluated with the ISTH Bleeding Assessment Tool (ISTH-BAT). The ISTH-BAT was designed to underscore the importance of repetitive minor bleeding in addition to more severe bleeds and can be used in all hemorrhagic disorders. 3 The ISTH-BAT is primarily designed as a screening tool. Since the ISTH-BAT documents large variety of bleeding symptoms, it is also used for the phenotyping of patients. 4,5 Very few studies have reported the bleeding phenotype in adult patients with CPDs 6 and most studies focused on a few specific types of CPD 7,8 or a specific mutation. 9,10 Phenotypic characterization of the whole spectrum of CPDs is necessary, since this could help physicians recognize CPD subtypes and inform new patients on prognostic implications regarding their bleeding phenotype. In this study, we aimed to evaluate the bleeding phenotype of a large cohort of adult patients with CPDs and to search for correlations between the bleeding score and different laboratory phenotypes of CPDs.

| Participant selection
Data were derived from patients included in the "Thrombocytopathy in the Netherlands" (TiN) study; a nationwide cross-sectional study to collect data on clinical features, functional assays and genetics in a population of patients with or suspected for a CPD. Patients were included in the TiN study when von Willebrand disease or a coagulation factor deficiency were excluded and when they were previously diagnosed with a CPD, they had previously abnormal platelet count or function test results, or they exhibited a predominantly mucocutaneous bleeding tendency compatible with a platelet function disorder. Within the TiN study, a CPD was confirmed when abnormal platelet count or function was found on at least two occasions, of which one was in our diagnostic laboratory. For the current evaluation, we included only TiN patients in whom a CPD diagnosis was confirmed.

| Bleeding phenotype
The ISTH-BAT was used for evaluation of the patients' bleeding symptoms and was administered by an experienced physician prior to platelet function testing. It contains questions on 14 domains: epistaxis, cutaneous bleeding, bleeding from minor wounds, urinary tract bleeding, gastrointestinal bleeding, oral cavity bleeding, post-dental extraction bleeding, post-operative bleeding, heavy menstrual bleeding (HMB), post-partum hemorrhage (PPH), muscle hematomas, hemarthrosis, central nervous system (CNS) bleeding and one final domain on other bleeding symptoms. Each domain was scored on a scale ranging from zero to four points. We classified a bleeding symptom as severe when the domain score was three or higher, since this indicates that the bleeding symptom required medical treatment. The total of all domains resulted in a bleeding score ranging from 0-56.
The cut-off values for an abnormal bleeding score are >3 for men and > 5 for women. 11

| Laboratory assessment
Laboratory tests were performed for platelet count, aggregation in response to four agonists (ADP, arachidonic acid, collagen, ristocetin), platelet ADP and ATP content, surface receptor expression with flow cytometry and whole-exome sequencing (WES) with a selected 76-gene panel (Table S1). Platelet morphology, gray platelets and leukocyte inclusion bodies were assessed in a peripheral blood smear. The cut-off value for abnormal platelet aggregation was determined for every agonist and was based on the 2.5th percentile, plus the coefficient of variation of 52 healthy donors (Table S2). The cut-off value for abnormal platelet receptor expression was determined based on the 2.5th percentile of 49 healthy donors (Table S3). Dense granule deficiency was diagnosed when the ADP content was lower than 1.4 μmol/10 11 platelets, based on the 2.5th percentile of 49 healthy donors. In line with the American College of Medical Genetics guidelines, a genetic variant was stated to be causal when a (likely) pathogenic variant (class four or five, respectively) 12 was identified in one or more of the selected genes that corresponded to the platelet phenotype.

| Statistical analysis
Statistical analyses were performed with IBM SPSS Statistics 25, Gra-phPad Prism software version 6 and RStudio version 0.99. Descriptive results for continuous variables were presented as medians (IQR), and categorical variables were presented as frequencies (percentages). The difference in bleeding score between types of CPDs, and between men and women was evaluated with linear regression analysis. Correlations between bleeding score and number of abnormal agonists in light transmission aggregometry (LTA), and between bleeding score and ADP content were calculated with non-parametric Spearman's rank correlation.

| Glanzmann thrombasthenia
Glanzmann thrombasthenia (GT) was diagnosed in 14 patients based on decreased or absent aggregation in response to ADP, arachidonic acid and collagen and decreased αIIbβ3 expression. The median bleeding score was 21 (IQR 16-22) ( Figure 1). The median platelet count was 183 × 10 9 /L (IQR 133-226) and in 5/14 patients the platelet count was below the normal range. The median αIIbβ3 expression was 1.9% (IQR 1.4% -3.3%). Eleven patients had type 1 GT, one patient had type 2 GT and two patients had variant type GT. Genetic mutations were identified in all 14 patients.

| Dense granule deficiency
Dense granule deficiency was diagnosed in 13 patients based on platelet ADP content below 1.4 μmol/10 11 platelets. The median bleeding score was 10 (IQR 9-14). The median platelet count was 165 × 10 9 /L (IQR 69-303) and in 6/13 patients the platelet count was below the normal range. The median platelet ADP content was 0.9 μmol/10 11 platelets (IQR 0.5-1.2). The LTA results were abnormal in 6/13 patients, but did not reveal a specific pattern. Genetic mutations were identified in five patients.

| ADP pathway defect
An isolated ADP pathway defect was diagnosed in 22 patients based on decreased aggregation in response to ADP. A normal aggregation in response to arachidonic acid differentiates them from patients with a thromboxane A2 (TxA2) pathway defect ( Figure S1A). The median bleeding score was seven (IQR 5-10). The median aggregation in response to ADP was 54% (IQR 47% -61%). Platelet ADP content was normal in all patients. Genetic mutations were identified in three patients.

| Complex abnormalities
A complex abnormality was diagnosed in six patients. A complex abnormality was diagnosed when the patterns of platelet function defects in LTA did not support the diagnosis of a particular type of CPD as described above. The median bleeding score was eight (IQR 4-24). No genetic mutations were identified in this subgroup.
We classified a bleeding symptom as severe when the domain score was three or higher. Both CNS bleeds were classified as severe The bleeding score was significantly higher in women than in men (β 3.5, 95% CI 0.9 to 6.1). When the bleeding score was calculated without the domains HMB and PPH, there was no significant difference (β 0.1, 95% CI −2.1 to 2.4).

| Bleeding symptoms according to type of CPD
Adjusted for age and sex, patients with GT had a significantly higher bleeding score than patients with other types of CPDs. As compared to other types of CPDs, GT patients more frequently (P < .01) experienced epistaxis (100% vs 45%), urinary tract bleeds (21% vs 1%), gastrointestinal bleeds (50% vs 7%), oral cavity bleeds (71% vs 12%) and hemarthrosis (36% vs 5%) ( Figure 3C). For the other types of CPDs, the bleeding phenotype was not distinctive. The bleeding phenotypes of CPDs other than GT were similar.

| Association between bleeding score and laboratory measurements
Correlations between the bleeding score and ADP, arachidonic acid

| DISCUSSION
This is the first study to report the bleeding phenotype and diagnostic characteristics of a large cohort of adult patients with CPDs. Patients were classified as Glanzmann thrombasthenia, Bernard-Soulier syndrome, dense granule deficiency, defects in the ADP or thromboxane A2 (TxA2) pathway, isolated thrombocytopenia or complex abnormalities. Most frequently occurring bleeding symptoms were HMB, PPH and post-operative bleeds. Glanzmann thrombasthenia and female sex were associated with a more severe bleeding phenotype. These patients possibly had more hemostatic challenges, since they had more time to develop bleeding symptoms, resulting in a higher bleeding score. Some of these bleeds might have required treatment, which would also have led to an increase in the bleeding score.
The strength of our study is the inclusion of a large number of CPD patients. To the best of our knowledge, this is to first study to