Screening and diagnosis of inherited platelet disorders

Abstract

Inherited platelet disorders are important conditions that often manifest with bleeding. These disorders have heterogeneous underlying pathologies. Some are syndromic disorders with non-blood phenotypic features, and others are associated with an increased predisposition to developing myelodysplasia and leukemia. Platelet disorders can present with thrombocytopenia, defects in platelet function, or both. As the underlying pathogenesis of inherited thrombocytopenias and platelet function disorders are quite diverse, their evaluation requires a thorough clinical assessment and specialized diagnostic tests, that often challenge diagnostic laboratories. At present, many of the commonly encountered, non-syndromic platelet disorders do not have a defined molecular cause. Nonetheless, significant progress has been made over the past few decades to improve the diagnostic evaluation of inherited platelet disorders, from the assessment of the bleeding history to improved standardization of light transmission aggregometry, which remains a “gold standard” test of platelet function. Some platelet disorder test findings are highly predictive of a bleeding disorder and some show association to symptoms of prolonged bleeding, surgical bleeding, and wound healing problems. Multiple assays can be required to diagnose common and rare platelet disorders, each requiring control of preanalytical, analytical, and post-analytical variables. The laboratory investigations of platelet disorders include evaluations of platelet counts, size, and morphology by light microscopy; assessments for aggregation defects; tests for dense granule deficiency; analyses of granule constituents and their release; platelet protein analysis by immunofluorescent staining or flow cytometry; tests of platelet procoagulant function; evaluations of platelet ultrastructure; high-throughput sequencing and other molecular diagnostic tests. The focus of this article is to review current methods for the diagnostic assessment of platelet function, with a focus on contemporary, best diagnostic laboratory practices, and relationships between clinical and laboratory findings.

Keywords:

• Light transmission aggregometry
• electron microscopy
• flow cytometry
• next-generation sequencing
• inherited platelet disorders

Introduction

Inherited platelet disorders are important causes of abnormal bleeding. Among the known inherited causes of bleeding and thrombotic disorders, many are conditions that alter platelet numbers and/or function [1]. The pathogenesis of inherited platelet function disorders (PFDs) and inherited thrombocytopenias (ITs) has emerged to be quite complex and heterogeneous. New causes of PFDs and ITs continue to emerge, and presently, many commonly encountered PFDs have an unknown molecular cause, even when new tools, such as next-generation exome sequencing are applied [2–7]. The complex and heterogeneous pathogenesis of these disorders continues to challenge clinicians and the diagnostic laboratories that assess patients for potential platelet disorders.

Over the last few decades, there have been major efforts to improve and standardize the laboratory assessment for PFDs, including platelet aggregation studies. This had led to the publication of many guidelines, including a recent guidance document from the International Society on Thrombosis and Haemostasis (ISTH) on stepwise investigations for diagnosing inherited PFDs [8–11]. The last decades have also seen the launch of external quality assessment (EQA) programs to assess components of PFD diagnosis, including the post-analytical interpretation of light transmittance aggregometry (LTA) tests, a “gold standard” test for PFD assessment [12]. EQA programs have also been established for the evaluation of platelet dense (δ)-granule deficiency (DGD) by transmission electron microscopy (TEM), using air-dried, whole-mount (WM) preparations of platelet-rich plasma (PRP) to examine and quantify the average number of δ-granules per platelet [12].

The focus of this article is to review current methods for the diagnostic assessment of suspected inherited defects in platelet function, with a focus on contemporary, best diagnostic laboratory practices. The use of platelet function tests for drug monitoring and other purposes will not be discussed in this review.

Current concepts on the causes of inherited platelet disorders

New causes of inherited PFDs continue to emerge and some of the genes that are mutated in these disorders can also be mutated in acquired PFDs (e.g. pathogenic RUNX1 mutations are seen in both congenital and acquired PFDs). Figure 1 summarizes current information on the genes underlying characterized, inherited platelet disorders, including those that cause PFDs and/or ITs [13]. The genes that cause ITs are summarized separately in Table 1 [6,7,14–70,72]. Table 2 summarizes PFDs associated with a normal platelet count [43,73–93,96–98]. There are excellent recent reviews on the causes and features of inherited PFDs and ITs [1,99–102].

Figure 1. Causes of platelet function disorders and inherited thrombocytopenias. The figure summarizes the genes and pathways that contribute to the pathogenesis of platelet disorders, which includes regulatory genes that have impacts on the early or later stages of megakaryopoiesis, as summarized in Lentaigne et al. [13].

Table 1. Overview of different inherited thrombocytopenias.

Pathology	Implicated gene(s)	Inheritance pattern	Platelet size	Other hematologic features	Other syndromic features
Tangier disease [14]	ABCA1	AR or AD	Normal	Giant α-granules Hepatosplenomegaly Reduction in δ-granules	Low HDL cholesterol Peripheral neuropathy Premature atherosclerosis
Sitosterolemia [15,16]	ABCG5, ABCG8	AR	↑	Hemolytic anemia Splenomegaly Stomatocytosis	Hypercholesterolemia Premature atherosclerosis Xanthomas
Baraitser-Winter syndrome with macrothrombocyto-penia [17]	ACTB	AD	↑	Eosinophilia Leukocytosis	Cardiac and renal defects Craniofacial dysmorphism Hearing loss Infections Intellectual disability
ACTN1-related thrombocytopenia [18,19]	ACTN1	AD	↑	Platelet anisocytosis	None
ANKRD26-related thrombocytopenia [20,21]	ANKRD26	AD	Normal	Leukocytosis Polycythemia Predisposition to leukemia Reduction in α-granules	None
ARPC1B deficiency [22]	ARPC1B	AR	↓	Bleeding diathesis Eosinophilia Immunodeficiency	Allergies Autoimmunity Eczema Vasculitis
Takenouchi-Kosaki syndrome [23]	CDC42	AD	↑	None	Cardiac, genitourinary, and lymphatic defects Facial dysmorphism Intellectual disability
Thrombocytopenia Cargeeg; CYCS-related thrombocytopenia [24]	CYCS	AD	Normal	None	None
DIAPH1-related thrombocytopenia [6]	DIAPH1	AD	↑	None	Deafness
ETV6-related thrombocytopenia and leukemia predisposition [7,25]	ETV6	AD	Normal, rarely ↓	Bleeding diathesis Macrocytosis Predisposition to leukemia	None
Paris-Trousseau and Jacobsen syndrome [26–28]	FLI1, del11q23	AD	Slightly ↑	Bleeding diathesis Giant α-granules	Cardiac defects Developmental delay Skull dysmorphism
FLNA-related thrombocytopenia [29]	FLNA	X-linked	↑	Bleeding diathesis	Periventricular nodular heterotopia Otopalatodigital syndromes
Congenital autosomal recessive small-platelet thrombocytopenia [30]	FYB	AR	↓	Bleeding diathesis	None
GALE-related thrombocytopenia [31]	GALE	AR	↑	Bleeding diathesis Neutropenia	None
GATA1 X-linked thrombocytopenia with dyserythropoietic anemia [32,33]	GATA1	X-linked	Slightly ↑	Platelet dysfunction Hemolytic anemia with clinical features similar to β-thalassemia	Congenital erythropoietic porphyria
GFI1b-related thrombocytopenia [34]	GFI1B	AD	↑	Bleeding diathesis Marked reduction in α-granules	None
GNE myopathy with congenital thrombocytopenia [35,36]	GNE	AR	↑	Bleeding diathesis	Glomerulonephritis Myopathy
Bernard-Soulier syndrome [37]	GP1BA, GPIBB, GP9	AR	↑	Bleeding diathesis	None
Platelet-type von Willebrand disease [38]	GP1BA	AD	Normal to slightly ↑	Bleeding diathesis Reduced von Willebrand factor activity with a reduced ratio of activity to antigen	None
Radioulnar synostosis with amegakaryocytic thrombocytopenia [39,40]	HOXA1/MECOM	AD/AR	Normal	B-cell deficiency Progression to bone marrow aplasia	Bilateral radioulnar synostosis Cardiac and renal malformations Clinodactyly Hearing loss
IKZF5-related thrombocytopenia [41]	IKZF5	AD	Normal	Reduction in α-granules	None
ITGA2B/ITGB3-related thrombocytopenia [42,43]	ITGA2B, ITGB3	AD	↑	Bleeding diathesis Platelet αIIbβ3 deficiency Platelet anisocytosis	None
Erythrokeratosis with thrombocytopenia syndrome [44,45]	KDSR	AR	Normal	Bleeding diathesis	Harlequin ichtyosis Keratoderma
Thrombocytopenia-2 [46]	MASTL	AD	Normal	Bleeding diathesis	None
Thrombocytopenia, anemia and myelofibrosis syndrome [47,48]	MPIG6B	AR	↑	Anemia Bleeding diathesis Hepatosplenomegaly Myelofibrosis	None
Congenital amegakaryocytic thrombocytopenia [49,50]	MPL	AR	Normal to slightly ↓	Progression to bone marrow aplasia	None
MYH9-related disease [51]	MYH9	AD	↑	Bleeding diathesis Döhle-like inclusion bodies	Cataract Glomerulopathy Hearing loss Liver enzymes elevation
Gray platelet syndrome [52]	NBEAL2	AR	↑	Bleeding diathesis Marked reduction in α-granules Progressive myelofibrosis with splenomegaly	None
Stormorken syndrome [53]	ORAI1	AD	Normal	Bleeding diathesis Immunodeficiency	Ectodermal dysplasia Muscular hypotonia
Quebec platelet disorder [54]	PLAU	AD	N	Delayed-onset bleeding Platelet-dependent gain -of-function defect in fibrinolysis due to megakaryocyte-specific overexpression of PLAU	Arthropathy secondary to joint bleeds
PRKACG-related thrombocytopenia [55]	PRKACG	AR	↑	Bleeding diathesis	None
Noonan syndrome [56]	PTPN11	AD	Normal	Bleeding diathesis Predisposition to leukemia	Cardiac defects Facial dysmorphism Short stature Skeletal malformations
PTPRJ-related thrombocytopenia [57]	PTPRJ	AR	↓	Bleeding diathesis	None
Thrombocytopenia with absent radius [58]	RBM8A	AR	Normal	Bleeding diathesis Platelet count generally improves during first 2 years of life	Bilateral radial aplasia Cardiac, renal, and CNS malformations
Roifman syndrome [59]	RNU4ATAC	AR	Normal	Increased δ-granules Hypogammaglobulinemia	Growth retardation Infections Intellectual disability Retinal dystrophy Skeletal dysplasia
Familial platelet disorder with predisposition to myeloid malignancy [60]	RUNX1	AD	Normal to slightly ↑	Bleeding diathesis Predisposition to myelodysplastic syndrome and leukemia	None
SLC35A1-related thrombocytopenia [61]	SLC35A1	AR	↑	Bleeding diathesis	Ataxia Choreiform movements Delayed psychomotor development Epilepsy Microcephaly
SLFN14-related thrombocytopenia [62]	SLFN14	AD	↑	Bleeding diathesis Reduction in δ-granules	None
SRC-related thrombocytopenia [63]	SRC	AD	↑	Myelofibrosis Splenomegaly	Facial dysmorphism Osteoporosis Premature edentulism
Stormorken syndrome/York platelet syndrome [53,64,65]	STIM1	AD/AD	Normal	Anemia Bleeding diathesis Immunodeficiency Reduction in α-granules	Asplenia Facial dysmorphism Ichtyosis Intellectual disability Myopathy
Inherited thrombocytopenia from monoallelic THPO mutation [66]	THPO	AD	Normal to slightly ↑	Hypogammaglobulinemia	Developmental delay Facial dysmorphism
TPM4-related thrombocytopenia [67]	TPM4	AD	↑	Bleeding diathesis	None
TRPM7-related thrombocytopenia [68]	TRPM7	AD	↑	Bleeding diathesis	None
TUBB1-related thrombocytopenia [69]	TUBB1	AD	↑	Bleeding diathesis	None
Wiskott-Aldrich Syndrome/X-linked thrombocytopenia [70]	WAS	X-linked	↓	Immunodeficiency	Autoimmunity Eczema Increased risk of malignancies
Wiskott-Aldrich syndrome 2 [71]	WIPF1	AR	Normal	Immunodeficiency	Autoimmunity Eczema Increased risk of malignancies

AD: autosomal dominant; AR: autosomal recessive.

Note. In some inherited platelet disorders that are associated with low platelet counts, including Quebec platelet disorder and familial platelet disorder with predisposition to myeloid malignancy, platelet counts can be normal in some affected persons.

Table 2. Overview of inherited platelet function disorders associated with a normal platelet count. The conditions listed have a normal platelet size.

Pathology	Implicated gene(s)	Inheritance pattern	Key hematologic features	Other syndromic features
Scott syndrome [73–75]	ANO6	AR	Bleeding diathesis with impaired platelet procoagulant activity	None
Ephrin type-B receptor 2-related disease [76]	EPBH2	AR	Bleeding diathesis Impaired platelet aggregation with multiple agonists, including ADP, collagen, arachidonic acid, and thromboxane analogue	None
Leukocyte adhesion deficiency-III [77–79]	FERMT3	AR	Bleeding diathesis Glanzmann-like defects in aggregation Leukocytosis	Infections Osteopetrosis
GPVI-related disease [80,81]	GP6	AR	Bleeding diathesis Absent collagen aggregation	None
Hermansky-Pudlak syndrome [82–84]	HPS 1,3,4,5 and 6 AP3B1 AP3D1 BLOC1S3 BLOC1S6 DTNBP1	AR	Bleeding diathesis δ-granule deficiency Neutropenia	Granulomatous colitis Infections Oculocutaneous albinism Pulmonary fibrosis
Glanzmann thrombasthenia [43]	ITGA2B ITGB3	AR	Bleeding diathesis Absent aggregation responses to all agonists (agglutination is present with ristocetin)	None
Chediak-Higashi syndrome [85]	LYST	AR	Bleeding diathesis δ-granule deficiency Lymphohystiocytosis	Infections Oculocutaneous albinism Oral ulcerations and periodontal disease Peripheral and central neurologic manifestations
Autism with platelet dense granule defect [86]	NBEA	AD	Abnormal δ-granule structure	Autism
P2Y12-related disease [87,88]	P2RY12	AR	Bleeding diathesis Reversible ADP aggregation even at high ADP concentrations	None
Cytosolic phospholipase A2 syndrome [89]	PLA2G4A	AR	Bleeding diathesis Impaired aggregation with collagen Impaired thromboxane A2 generation	Gastrointestinal ulcers
Aspirin-like syndrome [90,91]	PTGS1	AD/AR	Bleeding diathesis Aspirin-like defect in aggregation responses	None
RASGRP2- related disease [92,93]	RASGRP2	AR	Bleeding diathesis Aggregation defects resemble Glanzmann thrombasthenia	None
TBXA2 receptor-related disease [94,95]	TBXA2R	AD	Bleeding diathesis Impaired aggregation with arachidonic acid and thromboxane analogue	None
Ghosal syndrome [96]	TBXAS1	AR	Bleeding diathesis	Osteopetrosis
Arthrogryposis-renal dysfunction-cholestasis syndrome [97,98]	VIPAS39 VPS33B	AR	Bleeding diathesis	Arthrogryposis Cardiac defects Cholestasis Growth failure Ichtyosis Infections Renal dysfunction

AD: autosomal dominant, AR: autosomal recessive.

Some disorders, such as congenital megakaryocytic thrombocytopenia (CAMT) result from mutations in the genes encoding thrombopoietin (THPO) or the thrombopoietin receptor (MPL) [1,49,50,72,99–102]. An important group of PFDs is caused by pathogenic mutations in genes that regulate transcript in early megakaryopoiesis, and thus alter platelet numbers and/or functions; these include mutations in genes that also cause hereditary predisposition to other blood disorders, including myelodysplastic syndrome and acute leukemia (e.g. mutations in ANKRD26, ETV6, RUNX1) [1,7,20,21,60,99–102].

Pathogenic mutations in genes affecting platelet granule biogenesis and trafficking cause the α-granule protein disorders gray platelet syndrome (GPS) (e.g. NBEAL2) and arthrogryposis-renal dysfunction-cholestasis (ARC) syndrome (VPS33B, VIPAS39), and the syndromic δ-granule deficiencies of Hermansky-Pudlak syndrome (HPS) (HPS1, HPS3-6, BLOC1S3, DTNBP1, PLDN, AP3D1, and AP3B1) and Chediak-Higashi syndrome (CHS) (LYST), and several syndromic disorders that impair platelet secretion (STXBP2, NBEAL2) [1,52,82,85,97–104].

Mutations in genes for a variety of platelet cytoskeletal proteins are known causes of syndromic and non-syndromic ITs, including MYH9-related disorders, Wiskott-Aldrich syndrome (WAS) and the related, X-linked thrombocytopenia (XLT), and many other ITs, some of which are benign (e.g. ACTN1-related disorders) [1,18,19,70,71,99–102]. One interesting cause of a rather benign IT results from a mutation in CYCS that increases apoptosis during platelet formation [24,105,106].

Several severe PFDs result from severe deficiencies or dysfunction of GPIbIX (Bernard-Soulier syndrome; BSS), αIIbβ3 (Glanzmann thrombasthenia; GT), or molecules that impair αIIbβ3 function through other mechanisms (e.g. deficiencies of Kindlin 3, also known as FERMT3, and calcium and diacylglycerol-regulated guanine nucleotide exchange factor I [CalDAG-GEFI]) [1,37,42,78,92,99–102].

Several platelet disorders result from mutations of platelet receptors that increase ligand binding, which interestingly, manifest with bleeding and not thrombosis. For example, gain-of-function mutations of GPIbα that increase von Willebrand factor (VWF) binding cause platelet-type von Willebrand disease (VWD) [38,107]. Activating mutations of the αIIb or β3 chain of αIIbβ3 cause ITGA2B/ITGB3-related thrombocytopenia, also known as variant GT, a dominant PFD that impairs platelet cytoskeletal rearrangement, reduces αIIbβ3 expression on platelets, and impairs platelet aggregation [108].

A number of recessive PFDs affect the platelet transmembrane receptors (genes) for platelet-activating agonists, including the receptors for ADP (P2RY12, P2RX1) or thromboxane (TBXA2R), or impair thromboxane generation (e.g. PTGS1) [1,91,94,99–102,109]. Others are caused by mutations in genes encoding proteins involved in key signaling and activating pathways, including the collagen receptors GPVI that are important for platelet activation and α2β1 that mediates adhesion [1,81,99–102,110,111]. York and Stormorken syndromes are complex, syndromic PFDs with thrombocytopenia, non-blood phenotypic features, and defects in calcium channels caused by different mutations in the gene STIM1 [1,64,65,99–102,112].

Some PFDs cause unique impairments in hemostasis. Scott syndrome is a rare, recessive PFD caused by mutations in ANO6, which encodes the calcium-dependent phospholipid scramblase anoctamin 6 [74]. Scott syndrome severely impairs the activation-induced expression of procoagulant phospholipids on the outer surface of platelets and other cells [74]. Quebec platelet disorder (QPD) is a unique α-granule disorder caused by a duplication mutation of PLAU that repositions a downstream, megakaryocyte-specific gene enhancer element that “rewires” PLAU; the consequence is a unique, platelet-dependent, gain-of-function defect in fibrinolysis, due to >100 increased levels of urokinase plasminogen activator (uPA) in platelets and megakaryocytes [113]. QPD results in intraplatelet plasmin generation, which proteolyzes diverse α-granule proteins, and accelerates clot lysis when QPD platelets release uPA [114].

Prevalence of platelet disorders

The prevalence of platelet disorders has not been examined in population-based studies, however, several studies report that amongst patients tested for bleeding problems, the prevalence of some PFDs, including DGD, is similar to that of VWD [115–125]. A worldwide survey on the investigation of PFDs indicated that laboratories commonly find aggregation abnormalities amongst referred patients, particularly with ADP and/or epinephrine [126]. In areas of the world where consanguinity is culturally accepted, rare autosomal recessive PFDs have a much higher prevalence. A recent genome database study, that looked for predicted loss-of-function variants in the exome sequences of 58 platelet-associated genes with known phenotypic correlates, estimated the prevalence of PFD at 0.329% of the general population [127]. Accordingly, PFDs are considered common and important causes of bleeding.

Approach to the diagnosis of platelet disorders: clinical evaluation of the bleeding history

The first step in the evaluation of patients with a suspected disorder of platelet function and/or numbers is a thorough evaluation of the personal and family history of bleeding. Positive family history is a predictor of finding a bleeding disorder in index patients [128], particularly for dominantly inherited disorders. Most platelet disorders are associated with symptoms of mucocutaneous bleeding, such as extensive bruising, epistaxis, and heavy menstrual bleeding. They can also manifest with abnormal bleeding with hemostatic challenges, including childbirth, surgeries, and invasive dental procedures (e.g. tooth extractions). Challenge-related bleeding from PFDs and ITs typically occurs at the time of or soon after (i.e. on the same day) a hemostatic challenge, although in QPD, the onset is delayed by 1–3 days, reflecting increased fibrinolysis [129,130]. In contrast to severe PFDs, mild PFDs can manifest with excessive bleeding after some but not all hemostatic challenges [129,131,132].

Bleeding problems of PFDs and ITs have also been evaluated using bleeding assessment tools (BATs), which have the advantages of standardizing data collection and symptom severity grading as well as identifying when a person’s bleeding exceeds that expected for the general population [130,133–138]. Recent studies indicate that patients with inherited PFDs have a significantly higher ISTH-BAT score than healthy controls and a higher score than for patients with IT, although there is some overlap between each of these groups [131]. Within each of these groups, including healthy controls, females have higher bleeding scores than males, so sex-specific cutoffs are needed for bleeding score assessment [131]. Analyses of sex differences in bleeding symptoms indicate that both healthy females and females with inherited PFD report more mucocutaneous bleeding than males [129,139]. BAT scores for inherited PFDs are also predictive of future hemorrhages requiring treatment [131,132].

BATs have also been used to evaluate the bleeding risks for specific types of PFDs, using data for the bleeding symptoms and problems reported by: (1) persons with QPD vs. unaffected family members or (2) persons with commonly encountered PFD vs. general population controls [129,130,140]. Table 2 illustrates that for commonly encountered PFDs, there are significantly increased risks for experiencing bleeding symptoms and problems [129]. The magnitude of these risks reflects that some patients with these disorders experience excessive bleeding with some but not all hemostatic challenges [129].

A bleeding history evaluation for a PFD should also be mindful of syndromic features that suggest a specific disorder (see reviews [1,99–102]). For example, HPS is associated with oculocutaneous albinism, lung fibrosis, and colitis; Jacobsen/Paris-Trousseau syndrome (PTS) is associated with facial dysmorphism, cardiac defects, and mental retardation; MYH9-related disorders are associated with inclusion bodies in neutrophils (Döhle bodies) with or without sensorineural hearing defects, elevation of liver enzymes, nephritis, and/or cataracts; thrombocytopenia with absent radius (TAR) syndrome is associated with absent radii; and WAS is associated with microplatelets, eczema, and susceptibility to infections.

Associations between test findings and bleeding symptoms

Some interesting associations have been found between the findings for some platelet function disorder tests and bleeding symptoms. For example, impaired platelet aggregation with collagen is associated with bruising and wound healing problems, and more severe DGD is associated with surgical bleeding [129]. Among persons with PFD, bleeding scores are similar for those with or without DGD [141]. The ability of platelets to support thrombin generation (TG) may also influence bleeding as associations have been found between some PRP thrombin generation endpoints and prolonged bleeding from minor cuts and wound healing problems [142].

Approach to diagnosis: laboratory investigations

Recent guidelines on laboratory investigations for inherited PFDs have recommended a multistep diagnostic workup, after exclusion of VWD [9,10]. Many laboratories perform only a limited number of the recommended investigations, such as LTA, a test for DGD, and glycoprotein analysis when GT or BSS is suspected [143]. When LTA is included in the panel of tests for bleeding disorder diagnosis, the test panel sensitivity for detection of a bleeding disorder is improved, without sacrificing specificity [123]. Diagnostic strategies need to consider that some common PFDs, such as DGD, may go undetected unless specific tests for DGD are included in the investigations [141]. They also need to consider that abnormal findings for some PFD diagnostic tests (e.g. abnormal LTA with multiple agonists, confirmed-granule deficiency) are highly predictive of a bleeding disorder (Tables 3 and 4). Table 5 summarizes key recommendations from the most recent ISTH guidance document on investigating suspected PFD [9,131].

Table 3. Likelihood of experiencing bleeding for commonly encountered platelet function disorders.

Bleeding symptom	Odds ratio (95% CI)
Excessive bruising
Female	207 (22–1971)
Male	87 (4.2–1777)
Transfusion for hemorrhage	100 (5.9–1800)
Minor wound bleeding > 1 h	44 (2.5–790)
Excessive oral/dental bleeding	43 (5.4–350)
Epistaxis > 15 min	28 (7.3–110)
Excessive surgical bleeding	20 (5.1–77)
Wound healing problem	15 (3.8–55)
Excessive bleeding from trauma	12 (1.4–105)
Hematuria from urinary tract infection	5.2 (1.3–21)
Gastrointestinal bleeding	4.1 (1.5–12)
Excessive bleeding from childbirth or miscarriage	17 (2.9–93)
Heavy menstrual bleeding	8.8 (2.1–37)

Risk for experiencing different bleeding symptoms and problems (estimated in reference [129]) are shown as odds ratios (with 95% confidence intervals) for persons with a platelet function disorder (diagnosed based on confirmed impaired aggregation with multiple agonists by light transmission aggregometry and/or δ-granule deficiency), compared to general population controls.

Table 4. Likelihood of a bleeding disorder based on findings for different platelet function disorder tests [9,24,141,144].

Test	Criteria	Odds ratio (95% CI)
Light transmission aggregometry using platelet count adjusted PRP	Abnormal findings with ≥2 agonists vs. 0–1	23 (5–105)
Light transmission aggregometry using native PRP	Abnormal findings with ≥2 agonists vs. 0–1	27 (4–214)
Dense granule ATP release by luminescence	Abnormal findings with ≥2 agonists vs. 0–1 on two tests	1.5 (0.6–3.9)
Whole mount TEM	δ-granule deficiency (DGD) on two tests	97 (5.4–1740)

Platelet count and morphology

An essential component of platelet disorder investigations is the assessment of peripheral blood platelet count and platelet size, which help to narrow the differential diagnosis. Among hematology analyzers, those that use a fluorescence-based method for platelet counting give more precise estimates of the platelet count [145,146]. Low platelet counts should be confirmed on a light microscopy review of the blood film to assess for clumps and pseudo-thrombocytopenia, which is caused by EDTA-dependent platelet agglutination or satellitism (i.e. a rosette distribution of platelets around neutrophils, and less commonly, around other white blood cells) [147–149]. Pseudothrombocytopenia is reported in up to 17% of patients referred to hematology clinics for thrombocytopenia [149–154] and it can be readily evaluated by a review of the blood smear [155].

When assessing for hereditary macrothrombocytopenia, platelet size assessment by optical methods appear superior to impedance methods, and a mean platelet volume (MPV) >12.4 fL helps differentiate patients with MYH9-related disorders and biallelic BSS from those with immune thrombocytopenia (ITP) [156]. Measurement of the mean platelet diameter (MPD) by optical microscopy, or software-assisted image analysis, is also helpful to distinguish inherited PFD with giant platelets from ITP as well as identify PFD associated with abnormally small platelets [157], although such assessments have not been widely adopted.

The evaluation of the blood smear is also helpful to identify morphologic anomalies that are suggestive of some specific diagnoses. For example, large platelets are associated with some disorders, such as platelet-type VWD and GATA1 mutations, giant platelets are suggestive of MYH9-related disorders, PTS and BSS, whereas microplatelets are suggestive of WAS, TAR, or CAMT [9]. Giant α-granules are suggestive of PTS, whereas absent or rare azurophilic α-granules in platelets examined by light microscopy are suggestive of GPS, which can be secondary to mutations in NBEAL2, GFI1B, or GATA1 as well as the VPS33B or VIPAS39 mutations that are associated with ARC syndrome [9,27,34,52,158,159]. δ-granules are not visible on standard Wright-Giemsa or May-Grünwald-Giemsa staining and need to be evaluated by other approaches (discussed later).

Morphologic abnormalities of other peripheral blood cells can also provide helpful clues to a PFD diagnosis. For example, Döhle-like inclusion bodies in neutrophils are typical of MYH9-related disorders, and abnormal red cell morphology is a manifestation of the dyserythropoiesis seen in PFDs caused by GATA1 mutations [27,51,158].

Immunofluorescent staining of blood smears for platelet disorder assessment

Immunofluorescence staining of blood films to evaluate for a platelet disorder has not been widely adopted. However, it can help to characterize up to 27% of suspected inherited PFDs when performed with a standardized protocol to stain fixed, permeabilized blood cells with antibodies specific to platelet proteins of interest, in combination with control sample assessment and application of an algorithm that integrates MPV and immunofluorescent data [160,161]. This method can be used to evaluate for a variety of PFDs, including GT, BSS, PFDs with cytoskeleton abnormalities (such as MYH9-related disorders and PFDs with β1-tubulin and filamin A abnormalities), DGD, GPS, and PFDs caused by GATA1 and GFI1B mutations [160,161]. It requires very little blood, which is advantageous for testing children and can be done on shipped dried blood films to provide diagnostic support to resource-poor and remote settings.

Screening tests for platelet disorders and other primary hemostatic defects

Bleeding time

The bleeding time (BT) is no longer recommended as a laboratory test as it has poor sensitivity and specificity, even when the test is standardized (e.g. with control of venous pressure, temperature, and standardization of the site and the tool used to make an incision of standardized depth and length) [162,163]. Accordingly, the BT is not recommended for investigating PFDs or other disorders [9,10].

PFA-100/200 closure times

Closure times (CTs) are considered an optional test for PFDs that requires a PFA-100 or PFA-200 analyzer and sodium citrate anticoagulant whole blood (WB) [164]. Briefly, WB samples are aspirated at high shear into disposable cartridges and encounter a membrane with an aperture that is coated with collagen and epinephrine or with collagen and ADP [164]. The agonists trigger platelet activation and accumulation that closes the aperture, leading to the cessation of blood flow [164]. A prolonged CT can reflect a defect in VWF or in platelet numbers or function [165–167] but the test is also prolonged if there is a low hematocrit [167–169]. The PFA-200 has improved ease of use compared to the PFA-100 but uses the same principle [164]. An additional cartridge that is sensitive to the detection of P2Y receptor deficiency or blockade, is available, but its clinical utility for PFD diagnosis is uncertain [170–173]. While the CT is more sensitive to VWD and PFDs than the BT, it does not have sufficient sensitivity to be used for PFD screening [119,167,174–178].

Light transmission aggregometry

LTA was developed in the 1960s by Born and O’Brien, and it remains the gold standard test for PFD assessment [179,180]. Briefly, LTA measures the platelet aggregation responses of PRP to added agonists by measuring changes in light transmission that occur as platelets aggregate, which increases light transmission and are quantified, as illustrated by the normal LTA study shown in Figure 2. Table 6 summarizes the maximal aggregation (MA) findings for patient LTA results shown in the review (Figures 3–9) and includes findings for additional agonists.

Figure 2. Light transmission platelet aggregation findings for a control subject. The aggregation tracing illustrates that there is significant aggregation with all agonists, except with the lower concentration of ristocetin, which is expected. The maximal aggregation with each agonist falls within the reference interval. See Table 6 for the reference interval for each agonist.

Figure 3. von Willebrand factor loss-of-function and gain-of-function defects, evaluated by light transmission platelet aggregometry. Panel (A) illustrates ristocetin-induced platelet aggregation findings for a person with type 2A von Willebrand disease, who had delayed aggregation with ristocetin due to their loss-of-function defect. The maximal aggregation attained was within the reference interval, despite the delayed response. Panel (B) illustrates the findings for a person with type 2B von Willebrand disease and significant thrombocytopenia, whose gain-of-function defect in von Willebrand factor resulted in increased aggregation with the lower concentration of ristocetin. The responses to other agonists were within the expected range for a platelet rich plasma sample with a very low platelet count. See Table 6 for maximal aggregation data.

Figure 4. Aggregation findings for Bernard-Soulier syndrome. There is no agglutination or aggregation evident with ristocetin, whereas the responses to other agonists are within the expected range. See Table 6 for maximal aggregation data.

Figure 5. Aggregation findings for Quebec platelet disorder. There is reduced primary aggregation and no secondary aggregation with epinephrine, which is a hallmark feature of the disorder. This subject also had reduced maximal aggregation, and deaggregation, with thromboxane analogue U46619 as an incidental finding. See Table 6 for maximal aggregation data.

Figure 6. Aggregation findings for Glanzmann thrombasthenia. A response is evident with ristocetin but there is no response, apart from shape change, with other agonists. With ristocetin, the maximal response achieved is reduced, reflecting that there is agglutination but no aggregation. See Table 6 for maximal aggregation data.

Figure 7. Example of aggregation findings for δ-granule deficiency. Findings are shown for representative subject with confirmed platelet δ-granule deficiency (average: 1.2 δ-granules/platelet; reference interval: 4.9–10.0 δ-granules/platelet). Aggregation findings in this disorder can range from non-diagnostic to showing multiple abnormalities, particularly with weak agonists. This person’s maximal aggregation responses were within the reference interval with all agonists except there was a reduced response to the lower concentration of collagen and reduced secondary aggregation with epinephrine. See Table 6 for maximal aggregation values compared to the reference interval.

Figure 8. Examples of aggregation abnormalities with multiple agonists due to RUNX1 haploinsufficiency and unknown causes. Panel (A) shows the findings for a person with RUNX1 haploinsufficiency and mild thrombocytopenia, who had reduced aggregation with collagen and thromboxane analogue U46619, shared with other affected family members (not shown), who also had reduced maximal aggregation with arachidonic acid. Panel B shows the findings for an individual from a family with an autosomal dominant platelet disorder of unknown molecular cause, who had abnormal responses to multiple agonists that were particularly evident with arachidonic acid and thromboxane analogue U46619 as well as lower concentrations of collagen. See Table 6 for the maximal aggregation data of these subjects.

Figure 9. Aggregation abnormalities associated with aspirin or an aspirin-like defect. There is no aggregation with arachidonic acid whereas the aggregation response to thromboxane analogue U46619 is completely normal. Additionally, there is absent secondary aggregation with epinephrine (not shown) and reduced maximal aggregation with the lower concentration of collagen. See Table 6 for maximal aggregation data.

Table 5. International Society on Thrombosis and Haemostasis recommended investigations for suspected inherited platelet function disorders.

Line of investigation	Recommended tests
Confirmation of a bleeding diathesis	Personal and family history of bleeding and assessment of associated syndromic findings Bleeding assessment tool
Initial investigation	Complete blood count Blood smear Exclusion of a coagulopathy and von Willebrand disease (with prothrombin time, activated partial thromboplastin time, von Willebrand disease screen) Light transmission platelet aggregometry Granule release by luminometry or ELISA Flow cytometry
Second line investigations	Extended agonist panel for LTA Extended panel flow cytometry Platelet procoagulant activity Transmission electron microscopy evaluations of granules and ultrastructure Clot retraction Mixing tests (e.g. for distinguishing platelet type- from type 2B von Willebrand disease)
Optional investigations	Genetic testing Others (e.g. serum thromboxane)

Recommendations are from reference [9].

Standardization of LTA has been a challenge as it requires control of many pre-analytical and analytical variables (Tables 7 and 8). Several organizations have recommendations to standardize LTA, and the North American guideline includes recommendations on how to interpret the findings [8,10,11,181].

Table 6. Maximal aggregation findings for representative cases with platelet function disorders, evaluated by light transmission platelet aggregometry.

Agonist	Final agonist concentration	Maximal aggregation (%) reference interval	Figure 2. Normal control	Figure 3(A). Type 2A VWD	Figure 3(B). Type 2B VWD*	Figure 4. BSS*	Figure 5. QPD	Figure 6. GT	Figure 7. δ-granule deficiency	Figure 8(A). PFD due to RUNX1 haploinsufficiency*	Figure 8(B). PFD of unknown molecular cause	Figure 9. Aspirin-like defect
ADP	2.5 μM	>24	96	ND	23	85	32	1	65	37	16	29
ADP	5.0 μM	>43	97	ND	26	85	68	2	76	46	32	61
Collagen	1.25 μg/mL	>51	95	ND	12	ND	90	1	47	ND	5	32
Collagen	5 μg/mL	>85	93	ND	71	87	91	10	90	57^†	63	85
Epinephrine	6 μM	>9	95	ND	37	ND	4	0	53	ND	13	24
AA	1.6 mM	>77	95	ND	72	ND	94	5	89	61	0	1
U46619	1 μM	>70	96	ND	7	86	12	1	74	14^†	3	81
Ristocetin	0.5 mg/mL	˂7	2	1	40^†	0	2	1	1	1	2	1
Ristocetin	1.25 mg/mL	>75	85	86	62	0^†	92	69	82	59	79	81

ND: not done.

Patient and control subject findings are compared, including data for the aggregation tracings shown in Figures 3–9 and additional agonists responses. Abnormal maximal aggregation results are indicated with bold font.

*Denotes a thrombocytopenic sample.

^†Denotes abnormal maximal aggregation when compared to the authors’ reference interval for low platelet count samples.

Table 7. International recommendations on preanalytical variables for platelet aggregation studies by light transmission aggregometry.

Preanalytical variable	CLSI [181]	ISTH [11]
Pre-collection recommendations	Assess medication and dietary history; Avoid any substance affecting platelet function for at least 14 days Patient should ideally be fasting and avoid fatty meals prior to collection	Collection after a short rest Refrain from smoking at least 30 min before and from caffeine at least 2 h before testing Drug questionnaire must be administered prior to blood collection; NSAIDs should be stopped at least 3 days and aspirin at least 10 days before testing
Blood collection	Blood should be drawn with minimized stasis; with a needle of 19–21 gauge; in a 105–109 mM sodium citrate tube with an anticoagulant to blood ratio of 1:9	Blood should be drawn with minimal or no venostasis; with a needle of at least 21 gauge; in polypropylene or siliconized glass tubes; in a buffer anticoagulant (sodium citrate 109 or 129 mM)
Transportation	Hand carried at room temperature without vibration or shaking	No recommendation
Storage temperature	Samples should be stored at room temperature	Samples should be stored at room temperature
Preparation of PRP	Centrifugation at 170 g for 15 min at room temperature	Centrifugation at 200 G for 10 min at room temperature
Preparation of PPP	Centrifugation at 1500 G for at least 15 min	Centrifugation at 1500 G for 15 min
Assessment of PRP quality	Check for interfering lipemia, icteria and red cell contamination Platelet count should be <10 × 10^9/L PRP with a platelet count above 400 × 10^9/L should be diluted with PPP to standardized PRP count of 200–250 × 10^9/L	Hemolyzed samples should be discarded Check the PRP platelet count Samples with a platelet count <150 × 10^9/L should be interpreted with caution, but may be tested to exclude GT, BSS, and type 2B VWD Platelet count should not be adjusted to a standardized value with PPP

There has been limited application of EQA to LTA because the assay requires fresh PRP. Despite the recommendations to standardize agonist panels, no commercial supplier has developed a kit with all recommended agonists. The laboratories use different agonists, at different concentrations, which makes direct comparisons challenging [182]. The THROMKID-Plus Study Group reported high variability in LTA findings between 15 centers in Germany and Austria (due to different protocols, reagents, agonist concentrations, cutoff values) until they standardized their agonists and procedures [183]. The North American Specialized Coagulation Laboratory Association (NASCOLA) has collaborated with the External quality Control of diagnostic Assays and Tests (ECAT) Foundation to provide EQA assessments of the post-analytical interpretation of LTA findings, which has highlighted that there are opportunities to improve and standardize LTA interpretation practices [12].

Method and preanalytical variables

LTA is performed using samples collected by venipuncture into collection tubes containing a calcium chelator, commonly 3.2% buffered sodium citrate, at a 9:1 ratio of blood to anticoagulant [11]. Before drawing the sample, the collection of information on the patient’s medications (including non-prescription drugs), supplements, and natural products is helpful to limit preanalytical interferences [11]. As aggregation responses are affected by exercise and smoking, a minimal period of abstinence and rest before sample collection is reasonable [11,184–187]. Hemolysis and lipemia, and potentially icterus, can interfere as LTA uses optical detection. To minimize ex vivo platelet activation, LTA samples should be hand transported and not sent by pneumatic tube transport [188,189]. As there are diurnal variations in platelet function (e.g. with epinephrine), and there are time-dependent changes in aggregation responses post-sample collection, many laboratories schedule LTA tests for the morning and complete the test within 4 h post-collection, as recommended [11,190]. Alternative anticoagulants have been explored to extend LTA sample viability but none permit long delays in shipment or testing [191]. Blood and PRP collected for LTA and other platelet function studies should not be refrigerated because cold exposure alters platelet function, and PRP requires a brief preincubation at 37 °C immediately before testing [192]. To limit pH changes that affect LTA responses, blood for aggregation tests is often collected into buffered anticoagulants, and the harvested PRP is kept in capped tubes to limit air exposure [11]. PRP for LTA is prepared by low-speed centrifugation at 200 G for 10 min [11]. Higher centrifugal forces reduce erythrocyte contamination but they also reduce platelet recovery, due to the loss of the larger platelets [193,194].

Some centers adjust the platelet count of the PRP to a standardized value with autologous platelet-poor plasma (PPP) before LTA, whereas the ISTH does not recommend an adjustment [9,11]. The adjustment predominantly affects weak agonist responses, with native non-adjusted PRP showing significantly higher maximal aggregation responses to epinephrine and ADP compared to samples diluted with PPP; these differences affect MA reference intervals (RIs) [144,195–197]. Although both native and platelet count-adjusted PRP detect abnormalities predictive of a bleeding disorder, a prospective, non-inferiority study found that platelet count-adjusted PRP was actually superior to native PRP for detecting impaired aggregation responses due to a bleeding disorder [144].

Many laboratories are uncertain about how to evaluate LTA in patients with thrombocytopenia. Some have used diluted control PRP samples to establish RIs for low platelet count PRP, using regression to define the 95% confidence intervals for MA responses for a range of sample platelet counts [198]. Laboratories should not refuse to test LTA when patients are thrombocytopenic as LTA is very helpful to diagnose several thrombocytopenic disorders, including BSS, type 2B VWD, platelet-type VWD, and ITGA2B/ITGB3-related thrombocytopenia [11,198].

RIs for LTA MA responses to agonists require data for a minimum of 40 unique general population control samples, and non-parametric statistical analysis is recommended as most agonist responses do not show a Gaussian distribution [198]. Changes in procedures (e.g. new agonist lots) should be validated with a smaller number of samples [11].

LTA is measured by monitoring turbidity changes in stirred PRP samples, tested at 37 °C, with the PRP and PPP used to set the limits for 0 and 100% MA, respectively. The LTA agonist panel should include weak and strong agonists, with consideration of guideline recommendations on the agonists and agonist concentrations to test (Table 9) [11]. To ensure that the testing can evaluate for common PFDs, and important rare defects in specific receptors and pathways, the testing needs to include: ADP to test for P2Y12 and P2Y1 defects; collagen to test for a variety of defects, including those specific to GPVI and α2β1, impaired MA due to COX-1 inhibitor drugs, and other PFDs (which can require a low concentration of collagen for optimal detection of many functional impairments); arachidonic acid (AA) to test for defects in thromboxane generation and impaired aggregation from common PFDs; thromboxane analog U46619 to help detect common PFDs and distinguish these conditions from aspirin-like defects; epinephrine, as it helps detect abnormalities from common PFDs; and ristocetin, which is tested at high and low concentrations to respectively assess for loss and gain of functions defects that affect VWF binding to platelets. After adding agonists, LTA is monitored for 5 min with most agonists and for 10 min with epinephrine to ensure that MA is reached and that deaggregation can be evaluated.

Table 8. International recommendations on analytical variables for platelet aggregation studies by light transmission aggregometry.

Analytical variable	CLSI [181]	ISTH [11]
Temperature	Testing should be done at 37 °C	Testing should be done at 37 °C
Use of control	A known normal drug-free control must be run in parallel	A known normal subject must be run in parallel
Preparation of the minimal and maximal values	0 and 100% values must be set with autologous PRP and PPP, respectively	0 and 100% values must be set with autologous PRP and PPP, respectively
Stirring	According to the manufacturer	Sample should be stirred at 1000 rpm
Monitoring time	Aggregation must be monitored for 5 min, up to an additional 5 min to detect delayed aggregation	Aggregation must be monitored for 3 min, up to 10 min for agonists that do not reach maximal aggregation by 3–5 min
Time to completion	Testing should be completed within 4 h of the blood sampling	Testing should be completed within 4 h of the blood sampling

Interpretation

Adequate interpretation of LTAs requires careful assessment of multiple quantitative and qualitative parameters:

1. Shape change, which may not be evident with all agonists. With collagen, the “shape change” reflects the adhesion of platelets to collagen fibrils that reduces the light transmission before aggregation starts. P2Y1 defects are associated with a loss of ADP-induced shape change.

2. Latency of the reaction, which is usually longer with epinephrine.

3. The slope of the aggregation curve, which is often reduced when MA is reduced.

4. The presence of a secondary aggregation wave, which is particularly distinct with epinephrine, unless there is absent secondary aggregation. The primary and secondary aggregation responses are often fused with higher concentrations of ADP and with strong agonists.

5. The absence or presence of deaggregation, which may accompany reduced MA, particularly with some agonists (e.g. thromboxane analog U46619). Deaggregation with ADP is particularly striking with inherited P2Y12 defects, which are rare.

6. With ristocetin, the initial response reflects agglutination, which triggers platelet activation and aggregation. The agglutination and aggregation components may be fused, particularly when the MA response to ristocetin is normal. Figure 3 illustrates typical findings for loss of function and gain of function defects in VWF. In the loss of function defects in VWF (e.g. type 2A, type 2M, or type 1 VWD with very low VWF levels), the agglutination is often reduced and slow (as illustrated in Figure 3(A)), which delays and often reduces the aggregation that follows agglutination. With a gain of function defects, due to type 2B VWD (Figure 3(B)), and platelet-type VWD, there is an increased aggregation response to lower concentrations of ristocetin, even if there is significant thrombocytopenia as was the case in this example (note: the responses to other agonists in Figure 3(B) are within the expected range for low platelet count samples). In BSS, platelet counts are low, the blood film shows megathrombocytes, and RIPA is absent (Figure 4).

The North American guideline on LTA provides helpful guidance on how to interpret LTA findings [8]. Test records should be examined for abnormalities in the blood count and platelet morphology, to verify that the PRP contained more platelet than the peripheral blood and had minimal red cell and leukocyte contamination, and to ensure that the testing was completed within the appropriate time limits (i.e. within 4 h of sample collection). Simultaneously tested control samples are important quality controls. MA responses are very important to the overall interpretation, and it is good practice to repeat the tests for any agonists that showed abnormal findings to confirm the observation. Next, tracings should be evaluated for abnormalities (e.g. very delayed aggregation, deaggregation, comparison of the response to different concentrations of the agonist that is abnormal), which are typically only present if MA is impaired. Finally, the pattern of abnormal responses should be carefully examined to determine which agonists are affected and if the pattern suggests specific types of defects, as outlined in our customized interpretation guide (Table 10), which considers recommendations in the North American guidelines [8].

Table 9. Summary of recommended agonists for platelet aggregation studies by different international guidelines.

Agonist	Targeted receptor	CLSI [181]	North American guideline [8]	ISTH [11]
ADP Higher concentrations should be used if abnormal	P2Y1 and P2Y12	0.5–10 μM	2–10 μM	2 μM, to start
Collagen (type I fibrillary^a or Horm^b) Higher concentrations should be used if abnormal	GPVI and α2β1 Also evaluates signaling pathways	1–5 μg/mL	Start at a low concentration validated to detect NSAIDs	2 μg/ml, to start
Epinephrine Higher concentrations should be used if abnormal^c	α2-adrenergic receptors and some signaling pathways	0.5–10 μM	5–10 μM	5 μM
Thromboxane analogue (U46619) Higher concentrations should be used if abnormal^c	Thromboxane receptors (TPα and TPβ) and some signaling pathways	1–2 μM	1 μM	1 μM
Arachidonic acid Higher concentrations should be used if abnormal^d	Thromboxane pathway (TPα and TPβ) and some signaling pathways	0.5–1.6 mM	0.5–1.65 mM	1 mM
Thrombin receptor activating peptide Higher concentrations should be used if abnormal^d	PAR-1 and PAR-4	No recommendation	No recommendation	10 μM
Ristocetin Low concentration High concentration	Agglutination mediated by GP1bIXV and von Willebrand factor, followed by aggregation (αIIbβ3)	≤0.6 mg/mL 0.8–1.5 mg/mL	0.5–0.6 mg/mL 1.2–1.5 mg/mL	0.5–0.7 mg/mL 1.2–2 mg/mL

^aRecommended by the CLSI.

^bRecommended by the ISTH.

^cThe North American guideline recommends only one concentration.

^dOnly the ISTH recommends using a higher concentration.

Table 10. Approach to interpreting platelet aggregation studies by light transmission aggregometry. The approach has been updated from published guidance [8].

Finding on aggregation	Interpretation
Absent or reduced with arachidonic acid Normal with thromboxane analogue Reduced with collagen Absent secondary aggregation with epinephrine Normal to reduced aggregation with ristocetin	Suggestive of an aspirin-like defect, which is often acquired from drugs that inhibit COX1 but can be congenital
Absent aggregation with all agonists except ristocetin, which is reduced as aggregation does not occur after agglutination Thrombocytopenia with significantly impaired to absent aggregation with all agonists except ristocetin	Suggestive of inherited GT, but can also be seen in acquired GT Potential case is variant GT, which should be further evaluated by flow cytometry analysis of αIIbβ3
Absent aggregation with high concentrations of ristocetin	Suggestive of BSS, which is associated with varying degrees of thrombocytopenia If the finding is associated with macrothrombocytopenia, it may affect the amount of aggregation with other agonists Suggestive of VWD if normal platelets
Increased aggregation with low concentrations of ristocetin	Suggestive of type 2B VWD or platelet-type VWD To evaluate for a potential false positive finding, consider repeating and checking VWF levels
Markedly impaired aggregation and disaggregation with ADP Reduced aggregation with a number of other agonists	Suggestive of an ADP receptor defect affecting P2Y12 Note: P2Y1 affects shape change with ADP
Isolated, abnormal aggregation with collagen	Possible defect in collagen receptors but a false positive abnormality is much more common than a true positive
Abnormal aggregation with at least 2 agonists but does not fit the pattern of other problems in this table	Suggestive of a platelet function disorder not otherwise specified Both arachidonic acid and thromboxane analogue U46619 are impaired in many of these disorders, including those cause by pathogenic RUNX1 mutations If the response to ADP is abnormal, the cause is unlikely to be δ-granule deficiency (DGD)
Abnormal aggregation with 1 agonist excluding collagen and ristocetin	Finding is non-diagnostic False positives are common with thromboxane analogue U46619
Absent secondary aggregation with epinephrine, and a reduced primary wave with epinephrine	Might suggest QPD, which is sometimes also associated with reduced MA with ADP, collagen, thromboxane analogue U46619 Epinephrine aggregation is normalized in QPD if the testing is performed with added luciferin/luciferase reagent

Figures 3–9 illustrate the typical abnormal LTA findings from different PFDs. Isolated, reduced aggregation with thromboxane analog U46619 is a fairly common, non-diagnostic finding, but it can also be seen in persons with PFD, as illustrated in Figure 5, which shows both reduced MA and deaggregation.

Isolated, absent secondary aggregation with epinephrine is also a fairly common, non-diagnostic finding for platelet count adjusted PRP but it is an abnormal finding for native PRP [144,197,199]. In QPD, epinephrine aggregation responses are reduced with absent secondary aggregation and often reduced primary aggregation, as shown in Figure 5. QPD aggregation abnormalities with epinephrine are evident with both platelet count adjusted and native PRP samples, but the epinephrine aggregation response can be artifactually normalized if the reagent to simultaneously measure ATP release is added [200]. Some but not all QPD aggregation tests show additional abnormalities, such as reduced aggregation with thromboxane (as in the example shown in Figure 5), ADP, and/or collagen [201].

Conditions that dramatically impair MA responses to multiple agonists include acquired and inherited GT (Figure 6), and ITGA2B/ITGB3-related thrombocytopenia. In GT and ITGA2B/ITGB3-related thrombocytopenia, shape change is normal, but aggregation is either absent or markedly reduced due to severe deficiency/dysfunction of the platelet fibrinogen receptor αIIbβ3. The response to ristocetin in GT and ITGA2B/ITGB3-related thrombocytopenia is quantitatively reduced, as agglutination occurs but it is not followed by aggregation (Figure 6).

With DGD, platelet aggregation responses can be normal, non-diagnostic (e.g. absent secondary aggregation with epinephrine), or show abnormalities with multiple agonists (e.g. reduced MA with collagen, AA, and thromboxane analog U46619) [141]. The aggregation findings for an illustrative DGD case are shown in Figure 7. MA with ADP is normal in DGD, as the exogenous ADP overcomes the loss of platelet δ-granule ADP [141].

Some PFDs, including conditions associated with pathogenic RUNX1 mutations (Figure 8(A)), impair responses to multiple agonists but there is more aggregation in response to some agonists than seen with GT (Figure 6) [129,140]. Some inherited PFDs of unknown cause have aggregation abnormalities that resemble those seen with pathogenic RUNX1 mutations (Figure 8(B) vs. Figure 8(A)) [129].

A comparison of the findings with AA and thromboxane analog U46619 is important as it often helps to distinguish common PFDs with multiple abnormal agonist responses (e.g. from DGD, pathogenic RUNX1 mutations, and PFD due to other causes) from PFDs caused by aspirin or aspirin-like defects (Figure 9). With aspirin or aspirin-like defect (Figure 9), there is impaired aggregation with AA but normal aggregation with thromboxane analog U46619, which are typically accompanied by reduced aggregation with collagen (particularly with lower collagen concentrations), absent secondary aggregation with epinephrine, and sometimes, reduced aggregation with ristocetin.

Impedance aggregometry

Aggregation responses can also be evaluated by measuring changes to electrical impedance, either with PRP or WB. Impedance aggregometry tracings for PRP and WB look similar to PRP LTA tracings, although shape change is not detected by impedance changes. In impedance aggregometry, platelets in the sample form a coating on two submerged electrodes, with additional platelets attaching to the electrode after adding an agonist, to proportionately increase the electrical impedance. With WB aggregometry (WBA) with collagen, leukocytes also attach to the electrode, which contributes to the changes in electrical impedance [202,203]. Recent guidelines on testing for PFDs recommend assessing aggregation responses by LTA rather than WBA, as there is more extensive literature on LTA as a diagnostic test for a bleeding disorder [9].

WBA can be measured using several devices, including a lumiaggregometer (CHRONO-LOG Corporation, Havertown, PA, USA) and a Multiplate analyzer (Roche Diagnostics, Mannheim, Germany). While WBA with the Multiplate analyzer detects severe aggregation defects, such as GT, it is inferior to LTA for detection of a PFD [178,204–208].

Some point-of-care devices test WBA by modified methods that are not validated for the investigation of an underlying PFD. The VerifyNow device (Werfen, Barcelona, Spain) tests aggregation with capture beads to assess for impaired platelet function due to drugs, such as P2Y12 inhibitors. The PlateletWorks test (Helena Laboratories, Beaumont, TX, USA) assesses the percentage of “functional platelets” by comparing differences in platelet counts for blood collected into anticoagulant, with or without an agonist to induce aggregation; it has been used, in combination with viscoelastic testing, to guide transfusion therapy after cardiac surgery [209].

When performing WBA for PFD assessment, a smaller volume of sodium citrate anticoagulant blood is needed compared to LTA (typically 6–9 vs. 20–30 ml) and the agonist concentrations tested differ from LTA [210]. WBA should be completed within 3 h from collection [211]. WBA is not affected by icterus or lipemia, but hemolyzed specimens should not be tested as hemolyzed samples often show platelet activation [210]. In WBA, the contribution of platelets of all sizes is evaluated because macrothrombocytes are not lost as there is no sample centrifugation [181,212]. Like LTA samples, WBA samples should be maintained at room temperature and not refrigerated. WB samples are diluted 1:1 with saline solution before WBA, unless the platelet count is <100 × 10⁹/L [181,211]. However, procedures have been established and validated for the assessment of thrombocytopenic patients, using impedance aggregometry, that may encourage more laboratories to perform aggregation for thrombocytopenic patients [213,214].

Lumiaggregometry

Lumiaggregometry (Chrono-log, Havertown, PA USA) is a bioluminescent method that assesses platelet δ-granule ATP release, typically at the same time as performing LTA or WBA [210,215–217]. It is commonly done to assess for platelet secretion defects [126]. The test uses a reagent and the luciferin-luciferase reaction principle to quantify secretion [218]: briefly, when platelets are activated, δ-granules release their stored ATP and ADP. The released ATP combines with D-luciferin, which in the presence of luciferase, produces luciferyl adenylate that reacts with oxygen to produce light. The light emission can be quantified, relative to an ATP standard, to determine the amount of ATP secreted by the platelets. Figure 10(A,B) shows normal lumiaggregometry findings whereas Figure 10(C) shows abnormal findings. With strong agonists (e.g. collagen, Figure 10(A)), ATP release begins at the same time as aggregation whereas, with weak agonists (e.g. epinephrine, Figure 10(B)), the release is delayed until secondary aggregation commences. In Figure 10(C), the findings with collagen are normal, whereas with AA, aggregation and ATP release are reduced, and with thromboxane analog U46619, aggregation and ATP release are completely absent.

Figure 10. Examples of normal and abnormal lumiaggregometry findings. Each tracing show simultaneous evaluation of light transmission platelet aggregation (upper part of each tracing) and ATP release (lower part of each tracing), evaluated by lumiaggregometry. In panel (A), the ATP release with strong agonists (e.g. collagen), begins at the same time as aggregation. In panel (B), the ATP release with epinephrine, which is a weak agonist, is delayed until the start of secondary aggregation. In panel (C), there are several abnormal findings: aggregation and ATP release are absent with thromboxane analogue and aggregation and ATP release are reduced with arachidonic acid. Panels (A,C) also show the normal, rapid ATP release with thrombin (aggregation was not monitored as thrombin clots the sample).

The measurement of ATP release cannot discern between abnormal test findings due to DGD or defective δ-granule content release. The clinical utility of testing ATP release by lumiaggregometry for diagnosing PFDs has been questioned as the test endpoints for individual agonists show significant intra- and inter-individual variability that negatively affect the test reproducibility and diagnostic usefulness [217,219,220].

Transmission electron microscopy

TEM has been used to evaluate platelets, with a resolution of up to 0.2 nm [221]. TEM can be done with thin sections of fixed platelets (with or without immunogold labeling) and with WM preparations of dried, unfixed platelets to visualize δ-granules (e.g. Figure 11).

Figure 11. Transmission electron microscopy images of whole platelets in whole mount preparations of control and δ-granule deficient (DGD) samples. The image from the control platelet (left) shows numerous electron dense structures and the majority of these structures have smooth contours and a uniform electron density, which is typical of δ-granules. The platelet from the δ-granule deficient subject contains only one electron dense granule, consistent with their low δ-granule count and confirmed δ-granule deficiency.

TEM evaluation of platelet thin sections permits visualization of many platelet structures, including internal and external membranes, the surface-connected canalicular system, α-granules, mitochondria, the cytoskeleton, and glycogen stores. Recently, a standardized approach to platelet TEM was proposed to improve evaluations of platelet ultrastructure [222]. Platelet TEM can be helpful to confirm some diagnoses, such as the α-granule deficiency of GPS when light microscopy suggests this diagnosis.

Platelet WM TEM is one of the accepted tests for quantifying platelet δ-granule numbers [9,141]. WM TEM is included in the initial workup for primary hemostatic defects as DGD has a prevalence similar to VWD, and LTA and lumiaggregometry are not sensitive enough to be used as a screening test for DGD [124,141,220,223–225]. Associations between DGD (diagnosed by WM TEM), bleeding, and elevated ISTH-BAT bleeding scores indicate that DGD diagnosed by WM TEM is associated with a bleeding disorder [141].

WM TEM is a relatively simple method, described by James White, to assess for DGD [226]. Briefly, blood is collected into acid-citrate-dextrose (ACD) or sodium citrate (as for LTA), spun to harvest PRP, and then kept at room temperature until a drop of the PRP is placed on a Formvar-coated grid, followed by a brief rinse with distilled water, cautious blotting with a paper filter, and air-drying of the sample before TEM [141,226–228]. As there is a significant reduction in the platelet δ-granule count within a day of sample collection, WM TEM should be done on the same day as the sample is drawn [229]. Our preference is to prepare the grids within a few hours after sample collection. Because platelet δ-granules are rich in calcium and phosphorus, they appear dark (electron-dense) without any fixation or staining (Figure 11). The average platelet δ-granule count is then determined by counting DG in 30–50 platelets (our RI is 4.8–10.0 δ-granules/platelet), with comparison to a control sample [141]. Figure 11 compares δ-granules in a representative platelet from a control and DGD patient. Some suggest counting δ-granules in more than 80 platelets to reduce the coefficient of variation (CV), but in our experience, results are similar whether δ-granules in 30 or 100 platelets are evaluated [141,222].

The NASCOLA EQA program for assessment of δ-granules by WM TEM showed good agreement (>80%) between laboratories in distinguishing between normal and DGD samples as well as good consensus on the electron-dense structures within platelets that should be counted or not counted as δ-granules [12,230]. Typically, only the uniformly electron-dense structures, with smooth boundaries (including tadpole and purse-shaped granules) are counted, whereas less dense, irregular structures are not counted (both types of structures are evident in the control platelet shown in Figure 11). Some centers have published their methods to standardize what is counted as a δ-granule, but it is generally recommended that each laboratory established its own RI for platelet δ-granule counts, with a minimal number of dedicated interpreters to reduce inter-individual variability, as average δ-granule counts can vary significantly between centers [124,141,222,231,232]. When the test was set up at our center, energy dispersive spectroscopy analysis of the electron-dense structures was performed to verify which structures had calcium and phosphorus content typical of δ-granules, but few laboratories have this capability. Some have proposed that assessments for DGD include a formal evaluation of the δ-granule diameter as some patients with a bleeding diathesis have normal numbers but smaller-sized δ-granules [225]. Finally, reduced δ-granule counts should be verified on a second sample because mild reductions (4.0–4.8 DG/platelet compared to the RI of 4.8–6.0) are often false positive [141].

Thin section (TS) TEM requires more complex sample preparation than WM TEM. It is important to limit platelet activation before fixation. Typically, blood for TS TEM is collected into ACD anticoagulant, followed by preparation of PRP, then fixation of the platelets with buffered glutaraldehyde for a fixed time, followed by serial dehydration with alcohol before embedding the sample in a resin for preparation of ultrathin sections with an ultramicrotome, as outlined in protocols [233–236]. Samples should not be collected into EDTA as platelet structure is altered [237]. TS TEM can be helpful to diagnose rare disorders (e.g. platelet spherocytosis, GPS, CHS, and PTS) [28,238–241]. As TS TEM examines thin cross-sections of platelets, it has been suggested that to diagnose α-granule deficiency, these granules need to be deficient in more than 15% of 100 platelet cross-sections [222].

Other measurements of δ-granule content

Serotonin release

Virtually all circulating serotonin, also known as 5-HT, is contained within platelet δ-granules as platelets take up serotonin for storage in these granules via the serotonin transporter (SERT), and the vesicular monoamine transporter 2 (VMAT2); this makes serotonin useful for the assessment of platelet δ-granule serotonin uptake, content, and release [242,243]. When released, platelet serotonin is thought to promote platelet aggregation via binding to platelet 5-HT_2a receptors [244]. Serotonin release assays (SRAs) with radiolabeled ¹⁴C-5-HT have been the gold standard assay for evaluating δ-granule secretion for the diagnosis of heparin-induced thrombocytopenia (HIT) but are rarely used for inherited PFD assessment because the test requires a radioisotope [126,215,245,246]. SRA is performed by incubating PRP with the radioactive serotonin at 37 °C for 15 to 30 min to allow uptake, before adding an imipramine solution to stop the serotonin uptake. Next, platelets are incubated with agonists (with or without monitoring the response by LTA), before adding a formalin-EDTA solution to stop the further release. Then, total and released (sample supernatant) serotonin are quantified, using a β scintillation counter, to determine the serotonin uptake and release [247,248]. Application of the test to PFD diagnosis requires the determination of an RI for control samples [248]. Serotonin uptake can also be assessed to help differentiate DGD from secretion defects.

Other techniques have been developed to evaluate platelet δ-granule contents and their release without radioisotopes, using either fluorimetry (with o-phtalaldehyde, a reagent that reacts with serotonin), flow cytometry with mepacrine uptake, serotonin antibodies to visualize δ-granules (discussed further in the flow cytometry section), high-performance liquid chromatography (HPLC) or enzyme-linked immunosorbent assay (ELISA) of DG contents [249–256]. HPLC endpoints for serotonin can be measured either by electrochemical or fluorescent means [254,255]. The methods have been adapted to evaluate single platelets and to interpret dynamic results [253,257]. These methods offer many advantages including high sensitivity, good precision with a CV of <10%, a relatively low cost, but they require significant technical expertise [252,258]. For the evaluation of HIT, there is good concordance (92%) of the test findings to that of radioactive SRA, with samples showing lower serotonin release (e.g. estimated release of 20–49%) accounting for most of the discrepancies [259]. A recent modification that used liquid chromatography in tandem with mass spectrometry to measure serotonin release showed excellent performance [260]. However, few laboratories use such methods [126].

While antibodies are available to quantify serotonin by ELISA [261], there is limited information on their use for evaluating PFDs, and for the diagnosis of HIT, measurement of serotonin release by ELISA is inferior to HPLC [262].

Platelet ATP-ADP measurement

Bioluminescent assays can also be used to measure the ATP-ADP content of platelet lysate and releasate [263,264]. With this approach, the enzyme adenyl kinase in samples is inhibited by an ethanol-EDTA solution to prevent the conversion of ADP to ATP, and the content of ATP in samples is determined by the luciferin-luciferase reaction. Next, the sample ADP is converted in ATP by the phosphoenolpyruvate-pyruvate kinase enzymatic reaction so that it can be measured by the luciferin-luciferase reaction. The difference between the results is used to calculate the platelet ADP content, and the ratio of ATP:ADP as a ratio of more than 2 in platelets is suggestive of DGD [265]. The same testing can also be done with the supernatant of activated platelets, to differentiate between DGD and a secretion defect [266,267]. HPLC has also been used to quantify ADP and ATP content of δ-granules with good test performance characteristics [268–270].

Measurement of α-granule content

The contents of α-granules include both granule membrane proteins and stored, soluble proteins. While many α-granule membrane proteins are shared with the outer platelet membrane, P-selectin is a marker of platelet activation that is sequestered in both platelet α- and δ-granules until granule release occurs [271]. The soluble contents of α-granules include procoagulant and anticoagulant substances (fibrinogen, VWF, multimerin 1, coagulation factor V, PAI-1, α2-antiplasmin, uPA), growth and angiogenic factors (TGF-β, PDGF, VEGF), and chemokines, including PF4 [271]. Some soluble α-granule constituents are present at undetectable or very low levels in the plasma (e.g. PF4, β-TG, multimerin 1) and have been tested as markers of α-granule release, mainly in research settings [126,272–276]. P-selectin expression on activated platelets is often evaluated by flow cytometry [126].

Assays to assess platelet procoagulant function, including thrombin generation assays

Platelet activation leads to redistribution of negatively charged procoagulant phospholipids, with the movement of phosphatidyl serine (PS) and phosphatidyl ethanolamine (PE) from the inner to the outer platelet membrane; the release of polyphosphate from δ-granules; and the release of partially activated factor V from α-granules which enhances coagulation. Flow cytometry analysis of activated platelets suggests that there are subpopulations of platelets that express procoagulant PS and PE [277]. Platelet activation through glycoprotein VI or platelet thrombin receptors leads to increases in intracellular calcium and PS and PE externalization, which facilitate tenase and prothrombinase complex formation and subsequent TG [278]. TG is defective in Scott syndrome, a rare bleeding disorder caused by mutations in ANO6 (also named TMEM16F), which encodes the scramblase responsible for calcium-mediated PS and PE exposure; Scott syndrome results in significantly impaired TG on platelets and other cells [75].

Increases in circulating procoagulant platelets are postulated to cause a pathogenic hypercoagulable state that complicates some malignancies and ischemic strokes [279–284]. However, few laboratories offer tests to assess this.

DDAVP treatment for inherited PFDs increases platelet procoagulant activity and TG, without affecting P-selectin expression, δ-granule secretion, or platelet aggregation responses [285].

Newer methods of assessing TG, including the assessment by calibrated automated thrombogram (CAT), with PRP, have been used to evaluate platelet procoagulant function in PFDs and other bleeding disorders, including severe hemophilia [142,286–294]. CAT is done using PRP (adjusted to 150 × 10⁹ platelets/L), a fluorogenic substrate of thrombin (Z-GGR-AMC), with added calcium and recombinant tissue factor. This mixture is incubated at 37 °C in a fluorometer [295], with a calibrator, to quantify TG, with measurement of the endogenous thrombin potential, peak thrombin generation, time to peak thrombin generation, and lag time [295,296]. In severe hemophilia and other severe factor deficiencies, PRP TG is more impaired in those that have more severe bleeding, suggesting platelet-dependent TG influences the phenotype of bleeding disorders [286,287,291]. Not surprisingly, the most profound abnormalities in platelet-dependent TG have been observed in patients with Scott Syndrome and in mice lacking ANO6 [73,297].

CAT has recently been used to study TG in commonly encountered PFDs that impair the ability of platelets to support coagulation, as illustrated by Figure 12 that compares thrombograms for a control subject to a δ-granule deficient subject to subjects with a pathogenic RUNX1 mutation and QPD [142,292]. However, the overlap between PRP TG findings for controls and PFD subjects limits potential diagnostic applications of CAT for PFD assessment [142]. The defects in platelet-dependent thrombin generation in persons with pathogenic RUNX1 mutations likely reflect impaired platelet activation and the defects in DGD likely reflect impairments from the combined deficiencies of δ-granule ADP, which impairs platelets activation, and polyphosphate that enhances TG [142]. PRP, but not PPP, TG is impaired in QPD (Figure 12), proportionate to the loss of platelet factor V that results from intraplatelet plasmin generation, triggered by megakaryocyte-specific overexpression of PLAU, the urokinase plasminogen activator gene [292]. Presently, CAT and other TG assays are mainly used to assess PFDs for research purposes [298].

Figure 12. Thrombograms compare platelet rich plasma thrombin generation profiles for representative control and platelet function disorder subjects. The control sample generated more thrombin than the samples from the subjects with confirmed Quebec platelet disorder (QPD) or a familial platelet function disorder due to RUNX1 haploinsufficiency (RUNX1). The defects in these platelet function disorders include impairments in the ability of platelets to support thrombin generation.

Flow cytometry

While LTA is considered the gold standard in the investigation of PFDs, flow cytometry analysis can also provide important information. First, flow cytometry can be used to investigate: platelet glycoprotein deficiencies; abnormal platelet aggregation or secretion; and defects in platelet procoagulant function (e.g. by assessment of annexin V binding) [250,299–306]. Second, flow cytometry can be used to assess WB, even if there is significant thrombocytopenia or limited sample volume, which can be helpful for the investigation of PFDs in children [307–310]. Nonetheless, flow cytometry tests for PFDs are expensive to run, and they need to be standardized and validated for PFD evaluation, as discussed in the recent guidelines from the ISTH [306,311–313]. Although flow cytometry is recommended by the ISTH for the investigation of inherited PFDs, only a minority of clinical laboratories use it for initial tests, as more use it for second step investigations [9,126]. Flow cytometry protocols have been developed that show a good correlation with LTA and detect abnormalities associated with bleeding [306,314–318].

Mass cytometry has potential advantages over flow cytometry analyses of platelets as it allows for multiplex assessment of single cells, with simultaneous evaluation of many different components (e.g. surface and intracellular proteins, markers of platelet activation) [319], as well as the option of using barcodes to analyze results for individual cells or multiple samples simultaneously [320,321].

Protocols for performing flow cytometry analyses of platelets include a recent ISTH document [311,322–324]. Briefly, blood is collected in anticoagulant, similar to LTA, after excluding recent ingestion of potentially interfering medications that could affect platelet activation [311,325–327]. Sample centrifugation, and red cell lysis, should be avoided [327]. The sample is then diluted in buffered saline, at a physiological pH, to limit platelet activation, and incubated with appropriate fluorochrome-conjugated monoclonal antibodies [326]. Fixatives (e.g. paraformaldehyde or a commercially available preparation like PAMfix) can be used provided they do not interfere with recognizing the markers of interest [317,324,327–330]. Ideally, the test should be completed within 4 h from sample collection, although some glycoprotein and functional analyses give acceptable results up to 24 h post-collection, with fixation permitting even longer delays or transport of the sample to a central laboratory [311,317,331–333]. Samples for platelet flow cytometry are often kept at room temperature to limit activation, but there may be advantages to testing the samples at 37 °C, as this reduces the variability in findings [333]. Local protocols require validation to determine optimal sample preparation (including fixation), appropriate conjugated antibody concentrations for sample incubations, and appropriate agonist concentrations if testing platelet activation [312]. Control samples should be tested, and ideally, an RI should be established for the procedure to interpret the findings [312].

Surface glycoprotein analysis

Flow cytometry can be helpful to quantify platelet surface glycoproteins with much smaller volumes of blood than required for LTA. It has been used successfully to diagnose the glycoprotein deficiencies of GT (congenital but also acquired forms), BSS, and collagen receptor deficiencies [299–302,334,335]. The last ISTH guidelines recommend assessing αIIbβ3 (CD41/CD61) and GPIb/IX (CD42a/CD42b) in the initial investigations, and then α2β1 (CD49/CD29), GPIV (CD36), and GPVI as a second step [9], although many laboratories only perform flow cytometry to confirm the certain diagnosis (e.g. GT and BSS, as shown in Figure 13) [126]. Consensus ISTH recommendations on flow cytometry assessments of inherited and acquired PFDs and thrombocytopenias were recently published [311] (key recommendations are summarized in Table 11).

Figure 13. Examples of flow cytometry analysis to confirm the glycoprotein deficiencies for cases with Glanzmann thrombasthenia (GT) and Bernard-Soulier syndrome (BSS). The panels compare the relative mean fluorescent intensity (MFI) from flow cytometry analysis of control compared to patient samples, evaluated with fluorescent antibodies to the platelet glycoprotein (GP) receptor αIIbβ3 that is deficient in GT (upper panel) and to the GPIbα and GPIX components of the platelet GPIb/IX/V complex that is deficient in BSS (lower panel).

Table 11. Summary of key consensus recommendations from the International Society on Thrombosis and Haemostasis on using flow cytometry for the assessment of disorders of platelet number and function [311].

Variable	Recommendations
Clinical utility	Diagnosis of: Inherited or acquired deficiencies of platelet surface glycoproteins (BSS, GT, GPVI defects) Platelet α-granule secretion defects Defects in signaling pathways (RASGPR12, P2Y12, TXA2R) GFI1B macrothrombocytopenia (CD34) Procoagulant activity disorders (Scott and Stormkorken syndromes) Assessment or monitoring of: Increased platelet activation in prothrombotic states P2Y12 antagonists Immature platelet fraction
Pre-analytical variables General Analysis of surface molecules Analysis of agonist-induced platelet activation Analysis of monocyte-platelet aggregates	Medication history should be assessed Blood sampling should avoid procedures that might cause platelet activation Citrate is the first-choice anticoagulant Samples should rest for a minimum of 30 min then stored at room temperature A stabilizing solution may be used to allow delayed flow cytometry analysis of some platelet antigens Should be done on fresh whole blood samples within 24 h from collection Should be done on fresh whole blood samples within 4 h from collection Should be performed immediately after blood collection or in fixed whole blood
Instrument and reagent standardization	Must detect forward and side scatter Must detect at least two fluorescence signals at <1000 FITC molecule per cell Software for data acquisition should allow data storage in a sharable format Validated antibodies should be used as per the manufacturer recommendations Less commonly used antibodies should be validated
Platelet flow cytometry methods	Protocols should include: Instrument description User-configurable settings Description of the reagents used (source, lot number, target protein, concentration) Positive and negative controls Strategies for gating and event analysis A local reference interval for healthy individuals should be done locally
Reporting and quality control	Should be standardized Should allow for external audits

Platelet function testing after agonist addition

Platelet activation is evaluated with antibodies that detect conformational changes in platelet membrane glycoproteins, such as αIIbβ3, or the secretion of P-selectin or CD63 onto the outer platelet membrane [330,336,337]. The testing can be done with agonists that target different pathways, such as ADP, thrombin receptor-activating peptide (TRAP)-6, cross-linked collagen-related peptide (CRP-XL), AA, thromboxane analog U46619, and ristocetin [306,313,315,326,338,339].

Markers of platelet activation are typically used to assess disorders that impair platelet activation and signaling, which is outside the realm of many diagnostic laboratories [304,340–342]. P-selectin levels on the platelet surface should be low before platelet activation and a basal level of 15% is considered acceptable for proceeding with testing platelets for activation-induced changes [326,343,344]. P-selectin analyses have also been used to characterize platelets from patients with α-granule deficiency, although normal levels of P-selectin have been reported in some cases of GPS [345–348]. P-selectin also mediates the interaction of activated platelets with neutrophils and monocytes [349,350], which can be evaluated in WB samples, analyzing for events that show dual-labeling for platelet (usually with CD41, CD61 or CD42a or CD42b antibodies) and leukocyte markers [323,341] that are not normally evaluated when performing flow cytometry for PFDs [351–355].

Granulophysin is a membrane protein of the tetraspanin family that is normally sequestered in lysosomes and DG and becomes expressed on the outer surface of platelets after activation [356,357]. Granulophysin is reduced in platelets from persons with DGD [356,358,359], which may affect platelet activation as it is postulated to modulate integrin-mediated signaling and other activation-induced changes [360,361]. Platelet granulophysin has been tested to evaluate platelet secretion and DGD presence [338,347].

Other probes of platelet activation have also been used for flow cytometry. PAC-1 is a monoclonal antibody that recognizes an epitope on αIIbβ3 associated with activation-induced conformational changes [330,362,363]. Activation-induced changes in platelet αIIbβ3 can also be evaluated using fibrinogen antibodies [246,363–365], fluorescent fibrinogen, ligand-induced binding site (LIBS) antibodies that assess αIIbβ3 conformational change, and antibodies that detect conformational changes in platelet collagen receptors [366–369]. Protocols for testing these markers have been described [306,313,315,316,318,338,347,370,371].

Mepacrine uptake and release

Mepacrine is a fluorescent, antimalarial drug that mimics serotonin and accumulates in platelet DG that has been used as a reagent to assess platelet DG and platelet DG secretion defects [372,373]. Briefly, platelets are incubated at 37 °C with mepacrine (e.g. for 30 min), then its uptake (by test compared to control platelets) is measured by flow cytometry [250,374] and it can be reassessed after stimulation, to quantify DG secretion [250,375]. Mepacrine testing needs to be done within 6 h of blood collection, and while there can be some background staining of the platelet cytoplasm and RBC, mepacrine assessment shows a good negative predictive value and moderate diagnostic accuracy compared to ADP content measurement [376]. Concomitant testing of CD63 expression can be helpful to evaluate for DGD, even if lumiaggregometry findings are normal, as it can help distinguish DGD from secretion defects [377]. Nonetheless, few diagnostic laboratories use mepacrine and CD63 markers for PFD diagnosis.

Genetic investigations for platelet disorders

Genetic investigations have provided important insights on the causes of some inherited platelet disorders, and the genes that are mutated in many PFDs and ITs have now been identified [2]. With the transition from Sanger sequencing to high-throughput sequencing (HTS), either with whole-exome sequencing (WES), whole-genome sequencing (WGS), or targeted gene sequencing (TGS), genetic testing for the diagnosis of PFDs continues to improve. The cost of some HTS, such as WGS, has fallen ∼1000-fold, with WGS now costing <1000 USD [378]. These improvements have made HTS much more feasible for both research and clinical settings.

For the investigation of inherited PFDs, the yields are dependent on the HTS platforms, the genes and regions covered, and the phenotype and pretest probability of an inherited disorder (e.g. clinical suspicion of an inherited PFD based on symptoms and family history; the presence of thrombocytopenia; clinical or laboratory findings suggestive of a known disorder). One study that evaluated a 72 genes panel, comprised of genes associated with inherited PFDs or known to be critically important to platelet physiology, and interpreted the findings according to the American College of Medical Genetics (ACMG) guidelines, observing a diagnostic yield of ∼88% for patients with a clinical and laboratory phenotype that was suggestive of a particular PFD, whereas the yield was less (∼54%) when the features were not suggestive of specific etiology [4]. Similarly, the UK-GAPP study that used a WES approach to study a cohort of patients with IT had a yield of 46%, while their 30 genes panel had a 26% yield in the same population [379,380]. Both studies found a large number of variants of unknown significance (VUS), in ∼22 and 51% of subjects, respectively. The ThromboGenomics has studied a panel of 96 genes by HTS, which included 66 platelet genes, showing a diagnostic yield of ∼48% in patients with an abnormal platelet count but a yield of only 26% in patients with platelet dysfunction (defined as abnormal aggregation, reduced membrane glycoprotein levels, and/or abnormal secretion). It also detected molecular abnormalities in about 3% of patients with completely normal coagulation and platelet studies [2], some of which were VUS, suggesting that the true yield of the investigations is lower. Molecular studies of pediatric patients with platelet abnormalities have yielded a lower rate of molecular diagnosis, with about 16–24% showing pathologic variants [3,381]. Similar low yields have been seen for cohort studies of PFDs, which limit selection bias [129]. The SSC for Genetics in Thrombosis and Haemostasis of the ISTH has developed a cloud-based clinical software interface to encourage the community to submit curated genetic variants for diagnostic-grade genes causing bleeding, thrombosis, and platelet disorders, to improve diagnosis, variant classification, and regular bulk data transfer to ClinVar [382].

For some rare, severe PFDs, a molecular diagnosis can be helpful to prenatal counseling or patient care. For example, some MYH9 mutations are predictive of renal disease and of disease severity, which is helpful for planning follow-up and treatment [383]. However, access to molecular investigations is limited in some regions of the world. Additionally, some molecular defects are best evaluated by specific approaches, as some HTS methods do not detect mutations outside of the coding regions. Copy number variation (CNV) can be missed by some molecular testing strategies [384]. One example is QPD, in which the causative genetic mutation is a gain-of-copy number mutation due to tandem duplication of PLAU, for which specific diagnostic testing has been developed based on the unique breakpoint [385].

While it is intellectually satisfying to identify the cause of an inherited PFD, there are ethical issues to consider. Some disorders have other health implications that the patient who is being tested for may not be expecting. For example, the diagnosis of a hereditary PFD with a predisposition to leukemia (e.g. from pathogenic ANKRD26, RUNX1, or ETV6 mutations) may cause unintended psychological distress and increase the need for medical monitoring. Questions have been raised about whether it is appropriate to test for such disorders in young children, particularly when they are too young to independently provide informed consent. These issues have been addressed in some recent guidelines [386,387]. Some suggest a stratified approach to genetic testing, with the first priority of assessing for gene mutations with management implications, and second, to evaluate for mutations in genes with a higher risk of harm (i.e. mutations associated with an increased risk of malignancy) after obtaining informed consent [388].

It is too early to recommend molecular diagnosis as a first-line investigation, except for some rare disorders that are difficult to diagnose otherwise (e.g. QPD) or conditions for which there is a high clinical suspicion of a specific PFD. However, as knowledge on the pathogenesis of PFD improves, and database information on different variants expands, the yields of HTS may improve. Such improvements could change recommendations on which conditions to investigate for a molecular cause and the best molecular strategies to employ.

Conclusion

The investigation and diagnosis of an inherited PFD have evolved, from the recent validation of the ISTH-BAT score and its application for prediction of future bleeding to improved standardization of essential investigations, such as LTA, and improved guidance on the selection of other methods to consider, including specific tests for testing for DGD, expanded application of immunofluorescence staining for PFD diagnosis, and the use of flow cytometry panels and HTS, among others. However, some tests that are recommended for PFD diagnosis are performed in a minority of clinical laboratories and require further standardization and validation to verify utility beyond research investigations.

With the continued expansion of molecular databases that are helping to identify non-pathogenic variants and improvements in knowledge on the molecular pathogenesis of inherited PFD, HTS could have a growing role in PFD diagnosis, as long as ethical issues are properly considered.

Author contributions

AB led the manuscript preparation, with contributions from ST and CPH. CPMH supervised the manuscript preparations. ST prepared the manuscript figures.

Abbreviations
5-HT	=	serotonin
AA	=	arachidonic acid
ACD	=	acid-citrate-dextrose
ACMG	=	American College of Medical Genetics
ARC	=	arthrogryposis-renal dysfunction-cholestasis
BAT	=	bleeding assessment tool
BSS	=	Bernard-Soulier syndrome
BT	=	bleeding time
CalDAG-GEFI	=	calcium and diacylglycerol-regulated guanine nucleotide exchange factor I
CAMT	=	congenital amegakaryocytic thrombocytopenia
CAT	=	calibrated automated thrombogram
CHS	=	Chediak-Higashi syndrome
CLSI	=	Clinical and Laboratory Standards Institute
CNV	=	copy number variation
CRP-XL	=	cross linked collagen-related peptide
CT	=	closure time
CV	=	coefficient of variation
DGD	=	δ-granule deficiency
ECAT	=	External quality Control of diagnostic Assays and Tests
ELISA	=	enzyme-linked immunosorbent assay
EQA	=	external quality assessment
FERMT3	=	Kindlin 3
GPS	=	Gray platelet syndrome
GT	=	Glanzmann thrombasthenia
HIT	=	heparin-induced thrombocytopenia
HPLC	=	high-performance liquid chromatography
HPS	=	Hermansky-Pudlak syndrome
HTS	=	high throughput sequencing
ISTH	=	International Society of Thrombosis and Haemostasis
IT	=	inherited thrombocytopenia
ITP	=	immune thrombocytopenia
LIBS	=	ligand induced binding site
LTA	=	light transmittance aggregometry
MA	=	maximal aggregation
MPD	=	mean platelet diameter
MPL	=	thrombopoietin receptor
MPV	=	mean platelet volume
NASCOLA	=	North American Specialized Coagulation Laboratory Association
NSAID	=	non-steroidal anti-inflammatory drug
PE	=	phosphatidyl ethanolamine
PFD	=	platelet function disorder
PPP	=	platelet poor plasma
PRP	=	platelet rich plasma
PS	=	phosphatidyl serine
PTS	=	Jacobsen/Paris-Trousseau syndrome
QPD	=	Quebec platelet disorder
RI	=	reference interval
SERT	=	Serotonin transporter
SRA	=	serotonin release assay
TAR	=	thrombocytopenia with absent radius
TEM	=	transmission electron microscopy
TG	=	thrombin generation
TGS	=	target gene sequencing
THPO	=	thrombopoietin
TRAP	=	thrombin receptor-activating peptide
TS	=	thin section
uPA	=	urokinase plasminogen activator
VMAT2	=	vesicular monoamine transporter 2
VUS	=	variant of unknown significance
VWD	=	von Willebrand disease
VWF	=	von Willebrand factor
WAS	=	Wiskott-Aldrich syndrome
WB	=	whole blood
WBA	=	whole blood aggregometry
WES	=	whole-exome sequencing
WGS	=	whole-genome sequencing
WM	=	whole mount
XLT	=	X-linked thrombocytopenia

Disclosure statement

The authors have no conflicts of interest to declare.

Additional information

Funding

Funded in part by the Canadian Institutes for Health Research (201603PJT-364832, CPMH).