Comprehensive Platelet Phenotypic Laboratory Testing and Bleeding History Scoring for Diagnosis of Suspected Hereditary Platelet Disorders: A Single-Institution Experience

Abstract

Objectives:

Patients with hereditary/congenital platelet disorders (HPDs) have a broad range of clinical manifestations and laboratory phenotypes. We assessed the performance characteristics of the International Society on Thrombosis and Haemostasis bleeding assessment tool (ISTH-BAT) and clinically validated platelet laboratory tests for diagnosis of HPDs.

Methods:

The records of 61 patients with suspected HPDs were reviewed and ISTH-BAT scores calculated.

Results:

Nineteen (31%) patients had thrombocytopenia, and 46 (75%) had positive ISTH-BAT scores. Thirteen and 17 patients had prolonged PFA-100 (Dade Behring, Miami, FL) adenosine diphosphate and epinephrine closure times, respectively. Twenty-two had abnormal platelet light transmission aggregation. Twenty-four had platelet transmission electron microscopy (PTEM) abnormalities (10 dense granule deficiency, 14 other ultrastructural abnormalities). Positive ISTH-BAT scores were associated with thrombocytopenia (P < .0001) and abnormal PTEM (P = .002). Twenty-three patients had normal results.

Conclusions:

ISTH-BAT identified patients with suspected HPDs but lacked a robust association with laboratory abnormalities. Despite comprehensive laboratory testing, some patients may have normal results.

Materials and Methods

Definitive diagnosis of hereditary/congenital platelet disorders (HPDs) relies on a constellation of clinical evaluation and extensive platelet laboratory testing. Patients with HPDs usually have a wide range of clinical manifestations, from relatively severe bleeding at an early age to mild or absent bleeding, and may remain undiagnosed until late adulthood, especially if they have not had hemostatic challenges.¹ These patients’ platelet laboratory phenotypes are also broad.² Therefore, the diagnosis of HPDs remains a clinical challenge.

Clinical investigation of suspected HPDs initially includes examination of patient bleeding history and blood platelet counts. The former can be assessed by bleeding assessment questionnaires, which are important tools for standardized evaluation and communication in the treatment of patients with inherited bleeding disorders. Various historical bleeding assessment questionnaires and scales were recently incorporated and further standardized as the International Society for Thrombosis and Haemostasis bleeding assessment tool (ISTH-BAT), which has been validated for von Willebrand disease (VWD)³ and studied in suspected HPDs.⁴ This questionnaire, based on an assessment of bleeding in different organ systems and sites, results in a summative score of bleeding symptoms based on the need for medical attention and treatment required. The final score ranges from 0 to 56 points in women and 0 to 48 in men. Various cutoffs were established for both adult and pediatric populations.^5-7 Patient personal lifelong history and/or a family history of thrombocytopenia are also suggestive for an HPD.⁸

Platelet laboratory testing includes platelet counting and sizing indices, peripheral blood smear light microscopy, standardized platelet functional analyses, and esoteric tests, which include platelet flow cytometry (PFL) and platelet transmission electron microscopy (PTEM).² Standardized platelet functional analyses include PFA-100 (platelet function analyzer) testing and platelet light transmission aggregometry (LTA). The former employs collagen–adenosine diphosphate (CADP) and collagen-epinephrine (CEPI) cartridges whose blood flow closure time results are especially sensitive for detecting more severe HPDs such as Bernard-Soulier syndrome (BSS) and Glanzmann thrombasthenia. Platelet LTA,⁹ performed using platelet-rich plasma (PRP), has been prospectively studied in patients with HPDs.¹⁰ Abnormal aggregation with at least two agonists is considered to distinguish patients with HPDs from healthy controls.¹¹ This technique is widely available, and standardization has been important in achieving reliable results, reflecting effects of preanalytical variables and variations in technique.^12-14 More esoteric platelet testing has gradually been standardized in some clinical laboratories. Flow cytometry can be used to evaluate platelet surface glycoproteins (GPs) and activation in vitro.² PTEM can detect ultrastructural abnormalities such as absent dense granules in Hermansky-Pudlak syndrome,¹⁵ deficiency of α granules in gray platelet syndrome (GPS),¹⁶ giant α granules characteristic of Paris-Trousseau and Jacobsen syndromes,¹⁷ and giant opaque inclusions in York platelet syndrome.^18,19

Results

Clinical and Laboratory Findings of a Patient Cohort

Given the complexity of laboratory platelet testing,²⁰ algorithmic approaches have been proposed.^21-23 However, prospective studies on operational characteristics of these standard and esoteric laboratory tests and their relationship to the recently established ISTH-BAT are lacking in the literature. The goal of this study was to assess the performance characteristics of both clinically validated ISTH-BAT and various clinically validated platelet laboratory tests in the diagnosis of HPDs.

Results of PFA-100 and Platelet LTA Tests

Patients who were evaluated at Mayo Clinic-Rochester (2012-2014) for known or suspected HPDs were recruited for this study. Patient clinical features, including age, sex, and date of first medical encounter for bleeding diathesis, were collected. The ISTH-BAT was retrospectively calculated based on documentation from the medical record. A positive bleeding score was considered to be 4 or more for adult men, 5 or more for adult women, and 3 or more for children (aged <18 years) per previously published validation studies.^2-5 Standard laboratory testing results included blood platelet count, peripheral blood smear microscopic review, plasma coagulation testing results (prothrombin time, activated partial thromboplastin time, thrombin time, coagulation factor activities including factor VIII and selected other factor activities, plasminogen activator inhibitor antigen, α-2 antiplasmin activity, von Willebrand factor antigen and activity, and factor XIII activity by clot solubility assay), and platelet functional testing results (PFA-100 and LTA). Platelet esoteric testing included GP expression by PFL and dense granule and ultrastructural assessment by PTEM. The selection of platelet laboratory tests was per the discretion of the treating clinician. This study was approved by the Mayo Clinic Institutional Review Board.

Closure times measurements for the CADP and CEPI cartridges of the platelet function analyzer (PFA-100) were performed according to the manufacturer’s instructions and as described.⁹ LTA testing using PRP was performed on a PAP-8E aggregometer (Bio/Data, Horsham, PA) under standard clinical testing conditions (based on Clinical and Laboratory Standards Institute guideline H58-A) using the following agonists: arachidonic acid (AA, 1.5 mmol/L), ADP (5 mmol/L), collagen (COL, 25 µg/mL), epinephrine (EPI, 5 µmol/L), and ristocetin (RIS, 0.6 and 1.2 mg/mL). If responses to initial low doses of ADP, EPI, and/or COL were below expected values, testing was repeated with selected higher agonist doses. Percentages of maximum aggregation and the aggregometry tracing patterns were both used for interpretation in comparison to results from normal donor controls.²⁴

Correlation Between Platelet GP Expression and Other Platelet Testing Results

Discussion

Platelet Electron Microscopy Abnormalities and Their Association With Other Variables

From 2012 to 2014 (2 years total), 61 patients who had clinical features suggestive of HPDs (median age, 35 years [range, 2 months to 76 years]; 46 females) were included in this study. None of the patients had plasmatic hemostatic deficiencies, including VWD, or had any taken antiplatelet medications. A total of 19 (31%) patients had thrombocytopenia, and 46 (75%) patients had positive ISTH-BAT scores (range, 3-16). In total, 53 (87%) patients were evaluated with PFA CEPI, 52 (85%) with PFA CAPD, 39 (64%) with flow cytometry, and all 61 (100%) patients with PTEM.

These patients can be categorized into four groups: group I included four patients who had normal platelet counts and negative ISTH-BAT scores, group II had 11 patients who had thrombocytopenia (range, 13-149 × 10⁹/L) but negative ISTH-BAT scores, group III had 38 patients with normal platelet counts but positive ISTH-BAT scores, and group IV had eight patients who had both thrombocytopenia and positive ISTH-BAT scores ❚Table 1❚.

The median age was slightly younger in groups I and II, although not statistically significant, and a female preponderance was present in groups I and III. Bleeding manifestations (of any degree of severity) included mucocutaneous bleeding present in 31 (51%) patients, menorrhagia in 23 (50%) women, postoperative bleeding in 24 (39%), and epistaxis in 23 (38%). Other bleeding sites—including joint, muscle, and central nervous system hemorrhage—were the least frequent (8%, 7%, and 2% of patients, respectively).

Thirteen (21%) and 17 (28%) patients had prolonged PFA-100 CADP or CEPI, respectively. Thirteen patients had abnormal platelet aggregometry results for at least one agonist, of which nine patients had abnormal platelet aggregometry responses for at least two agonists. Functional platelet studies (PFA-100 and LTA) were not performed in eight and six patients, respectively, due to severe thrombocytopenia. A total of 24 patients had PTEM abnormalities, which included 10 patients with dense granule deficiency, and 14 had other ultrastructural abnormalities. A small number of patients with HPDs were definitively diagnosed based on pathognomonic laboratory features, molecular testing results, or both. These included MYH9-associated platelet disorder (c.3118_3120delGAG; p.Glu1040del and 219.4-kb amplification on 22q involving exon 1; n = 1), GPS (n = 1), York platelet syndrome (n = 2, previously reported),²⁷ BSS (n = 2), and RUNX1 mutation (c.710G>A; p.Arg237Lys; n = 1)–associated thrombocytopenia (n = 1). In this cohort, positive ISTH-BAT scores were associated with thrombocytopenia (χ², P < .0001) and abnormal PTEM results (χ², P = .002) but were not associated with age, sex, blood type, abnormal PFA-100, LTA, or PFL results.

Of the 53 patients who had both PFA-100 and LTA testing performed, 17 patients had prolonged CEPI, and 13 patients had prolonged CADP closure times. Twenty-one patients had at least one prolonged closure time, while nine patients had both closure times prolonged. The remaining 32 (60%) patients had normal PFA-100 closure times. For LTA testing, 24 patients had abnormal aggregation responses to AA, 17 to ADP, 24 to EPI, 14 to COL, and 12 to RIS. In patients with abnormal LTA results, four had abnormal results for one agonist, four for two agonists, three for three agonists, one for four agonists, and one for five agonists. Results of PFA-100 and LTA are compared and summarized in ❚Table 2❚. Of the 36 patients who had normal CEPI, only three had abnormal LTA results—exclusively due to decreased AA-induced platelet aggregation. None of these patients had abnormal LTA results to two or more agonists. Of the 17 patients who had prolonged CEPI, eight patients had abnormal LTA results for at least two agonists, and one patient had abnormal LTA for only one agonist. There was a strong correlation between CEPI and LTA results (χ², P < .0001). Of the 39 patients who had normal CADP, five patients had abnormal LTA results for at least two agonists and three patients had abnormal LTA results for one agonist. Of the 13 patients who had prolonged CADP, only three patients also had abnormal LTA results for at least two agonists and one patient had abnormal LTA result for one agonist. The association between CADP and LTA results was insignificant (P = .4).

Flow cytometry studies were performed for 39 patients, and 33 (85%) had normal results. Two patients had markedly decreased GP-Ib-α and GP-IX expression. Both patients had markedly decreased RIS-induced platelet aggregation. These results were consistent with BSS. One patient who had an MYH9 mutation had globally elevated GPs, which were consistent with macrocytic platelets. This patient had normal PFA-100 and LTA results. The remaining three cases showed mild GP-Ia deficiencies. Interestingly, they all had thrombocytopenia and decreased mean dense granule counts (range, 0.7-1.0) by PTEM. The first patient was a 54-year-old man (GP-Ia = 43%) who had a negative ISTH-BAT score, absent LTA-EPI and decreased LTA-COL, and prolonged PFA-100 CEPI and CADP. The second patient was a 22-year-old woman (GP-Ia = 44%) who had a negative ISTH-BAT score and normal PFA-100 and LTA results. The last patient was a 29-year-old woman (GP-Ia = 53%) who had a positive ISTH-BAT score and prolonged PFA-100 CEPI and CADP but normal LTA results.

In comparison to normal donor platelets ❚Image 1A❚, ❚Image 1B❚, and ❚Image 1C❚, 24 patients from all four patient groups had abnormal PTEM findings (Table 1) ❚Table 3❚. Ten patients had dense granule deficiencies by WM-PTEM ❚Image 1D❚. The range of mean dense granules/platelet was 0 to 1.1 (median, 0.75). Of the 10 patients who had dense granule deficiencies, six and five patients had prolonged PFA-100 CEPI and CADP, respectively. Two patients had abnormal LTA-AA, two had decreased LTA-EPI, one had low LTA-COL, and seven had completely negative LTA results.

Clinical Features of Patients Who Had Positive ISTH-BAT Scores but Completely Negative Platelet Laboratory Testing Results

The remaining 14 patients had other ultrastructural abnormalities by TS- and WM-PTEM. The abnormalities include decreased α granules, increased canaliculi ❚Image 1E❚, absent α granules (GPS) ❚Image 1F❚, abnormal dense inclusions in platelets from a patient with York platelet syndrome ❚Image 1G❚, large and round-shaped platelets ❚Image 1H❚, and cytoplasmic inclusions in WBCs of a patient with MYH9 mutation-associated platelet disorder ❚Image 1I❚.

Thrombocytopenia is more common in patients who had platelet-dense granule deficiencies or other ultrastructural abnormalities (Table 3). Of these 14 patients, 11 had thrombocytopenia. Seven of 12 had prolonged PFA-100 CEPI, and six of 11 had prolonged CADP closure times. For LTA testing, six of 12 patients had abnormal platelet aggregometry results for at least one agonist, and four had abnormal LTA results for at least two agonists. When both types of PTEM-abnormal cases are combined, approximately 30% to 50% of patients who had abnormal PTEM findings had abnormal PFA-100 or LTA results. Abnormal PTEM results were statistically associated with abnormal PFA-CEPI (P = .007), PFA-CADP (P = .03), and LTA–one abnormal agonist (P = .02), including abnormal LTA with ristocetin (P = .003). They were also strongly associated with positive ISTH-BAT scores (χ², P = .007) and thrombocytopenia (P < .0001).

Platelet transmission electron microscopy (PTEM) of a healthy donor and patients with suspected hereditary platelet disorders. Low (A, ×12,000; B, x8,000) and high (C, x20,000) magnification of thin-section and whole-mount (WM) PTEM of platelets from a normal healthy donor. The arrow points to a dense granule (C). A platelet (WM-PTEM) without any visible dense granules is observed in a patient who has a dense granule deficiency (D, x15,000). E, shows markedly decreased α granules and increased cytoplasmic vacuoles and canaliculi in a subset of platelets (arrows) (x15,000). A complete lack of α granules is observed in a patient with gray platelet syndrome (F, x30,000). WM-PTEM shows that abnormal cytoplasmic dense inclusions (arrow) are present in a platelet from a patient with York platelet syndrome (G, x25,000). Large and round-shaped platelets (H, x8,000) are observed in a patient who has MYH9 mutation-associated thrombocytopenia. WBC transmission electron microscopy shows an abnormal cytoplasmic inclusion (I, x15,000, arrow and inset).

In this cohort, 23 (34%) patients (median age, 50 years; 20 females and three males) had completely normal laboratory testing results, which included platelet count, PFA-100, LTA, flow cytometry, and PTEM ❚Table 4 ❚, with a median ISTH-BAT score of 5 (range, 3-13). These patients appear to have higher ISTH-BAT scores (median, 5; range, 3-13; P = .06) and more positive ISTH-BAT scores than those who had at least one abnormal platelet laboratory results (Table 4). Age, sex, and blood type O (other blood types not shown) were not significantly different between the two groups (Table 4).

Patients with HPDs may have thrombocytopenia, bleeding diathesis, or both. Thus, such patients can be categorized into four groups as reflected in this cohort. All patients⁴ in group I (normal testing results and negative ISTH-BAT scores) were female and had borderline bleeding symptoms with ISTH-BAT scores of 3 to 4. One patient had a dense granule deficiency confirmed by PTEM. In groups II and IV (thrombocytopenia without or with abnormal ISTH-BAT scores, respectively), a total of 19 patients had thrombocytopenia, and 11 (group II) had no apparent bleeding diathesis. Similar observations have been reported in various small cohort studies.^27,28 Many of the group II patients’ thrombocytopenias were detected incidentally. None of the patients who had the York platelet syndrome or MYH9 or RUNX1 mutations with associated thrombocytopenias had a bleeding diathesis.

References

Since ISTH-BAT scores reflect the severity of patient bleeding symptoms, it might be assumed that the bleeding score is associated with the degree of laboratory abnormalities. In this study, we examined whether ISTH-BAT scores or the qualitative positive/negative ISTH-BAT interpretations were associated with various laboratory results. Interestingly, of all laboratory results, only the presence of thrombocytopenia and abnormal PTEM findings were associated with the likelihood of a positive BAT score. In addition, 23 patients with positive BAT scores had completely normal laboratory results. The lack of a strong association between ISTH-BAT and laboratory abnormal results was consistent with a recent study by the UK Genotyping and Phenotyping of Platelets Study Group.⁴ Potential explanations include that some of the patients may not have platelet disorders, their bleeding diatheses could have been caused by other unidentified mechanisms, and/or the current panel of platelet function laboratory tests does not completely reflect all pathologic mechanisms of platelet disorders in vivo.

PFA-100 and LTA are the two most commonly performed and available clinical diagnostic laboratory tests for the study of platelet dysfunction. PFA-100 is a point-of-care platelet function device using citrated whole blood. It has high sensitivity to detect moderate to severe platelet dysfunction and VWD^29,30 and has been considered a good platelet function screening test.³⁰ However, a direct comparison of PFA-100 and LTA results in patients with suspected platelet disorders has not been well documented in the literature. Using LTA as the reference (gold standard) method, PFA-100 has a 92% NPV and 53% PPV for at least one abnormal LTA test result in this cohort. The NPV increases to 100% with at least two abnormal LTA results. PFA-CADP results have a much lower NPV and PPV in comparison to PFA-CEPI. These observations suggest that PFA-CEPI could potentially serve as a screening test for decisions about performing LTA, although it is not necessarily a good screening test for detecting platelet disorders.³¹

Platelet GP assessment by flow cytometry is considered a confirmatory test for diagnosing quantitative platelet GP deficiencies.³² In this study, only six cases showed abnormal PFL results. Both cases of BSS were confirmed by PFL. In addition, one case showed a global increment of GPs due to macrocytosis in a patient with MYH9 mutation-associated thrombocytopenia, and three showed mildly decreased GP-Ia. Interestingly, all of the latter three cases had dense granule deficiency. GP-Ia is a collagen receptor and has been shown to participate in platelet adhesion. It has not been shown to be a key player in dense granule genesis.

PTEM is a useful tool to evaluate platelet ultrastructural abnormalities, including platelet storage pool deficiencies. It was first used to visualize fibrin-platelet clot formation in 1955³³ and later used to evaluate platelet ultrastructural abnormalities.³⁴ PTEM employs two main methods, WM- and TS-TEM. The former is a quick and simple way to examine platelet electron opaque structures such as dense granules^35,36 and abnormal inclusions in rare HPDs such as York platelet syndrome^17,18 by applying PRP onto a grid.³⁷ WM-PTEM is considered the reference (gold standard) method for diagnosing platelet-dense granule deficiencies of patients with Hermansky-Pudlak syndrome,¹⁵ α-δ platelet storage pool deficiency,³⁸ Paris-Trousseau-Jacobsen syndrome,³⁹ Wiskott-Aldrich syndrome,⁴⁰ thrombocytopenia with absent radii syndrome,⁴¹ and Chédiak-Higashi syndrome.⁴² The TS-PTEM can visualize platelet α granules, canalicular networks, glycogen, other normal and aberrant cytoplasmic structures, and inclusions.³⁴ Distinct and sometimes pathognomonic ultrastructural abnormalities have been recognized in GPS with virtually absent α granules,^16,43 White platelet syndrome,⁴⁴ Medich giant platelet disorder,⁴⁵ X-linked GATA-1 macrothrombocytopenia,⁴⁶ and the recently described York platelet syndrome.^17,18 Nevertheless, wider clinical laboratory implementation of PTEM has been mainly hampered by the lack of standardized procedures, unified PTEM image interpretation guidelines, and carefully validated normal ranges for mean dense granule count/platelet and other baseline ultrastructures. We recently clinically validated the method and established a normal range of dense granules and baseline ultrastructures.²⁵

In this study, the most common PTEM abnormalities were storage granule deficiencies of either dense or α granules, suggesting that storage granule/pool deficiencies are likely the most commonly encountered platelet ultrastructural abnormality. Of the 10 cases of dense granule deficiency by PTEM, only about 56% and 44% cases had abnormal PFA-100 CEPI or CADP results, respectively, and 33% cases had an abnormal LTA result. Similar findings were also observed in the 14 cases with other platelet ultrastructural abnormalities. Therefore, none of the standard platelet laboratory tests provide sufficient sensitivity to serve as a screening test(s) for supplementally performing PTEM, an observation that is consistent with findings from other studies.⁴⁷

A technical limitation of this study was that we did not measure dense granule adenosine triphosphate (ATP) release by luciferin/luciferase luminometry. However, a recent study demonstrated that an ATP release assay may have limitations for reliably detecting platelet dense granule deficiency, reflecting significant variation among healthy donors⁴⁸ and patients with suspected HPD.⁴⁹ Therefore, PTEM remains an invaluable test for identifying platelet-dense granule storage pool deficiency and other HPDs with ultrastructural abnormalities. In addition, a proportion of patients with normal comprehensive platelet testing may have a vascular explanation for their bleeding. Evaluation for vascular hypofunction is primarily a clinical assessment and not included in the extensive laboratory studies performed.

This is a 2-year cohort study of our institutional experience for diagnosing suspected hereditary/congenital platelet disorders. There are several limitations. Some patients were referred to us because of discrepancies observed between bleeding symptoms and unremarkable laboratory findings—and this may help explain the relatively high percentage of patients with completely normal laboratory results in comparison to the reported prevalence.⁵⁰ Also, there is a significant female preponderance in this study cohort, which, although it may reflect the sex distribution for bleeding patients, it could obscure and confound the true findings for male bleeders. The ISTH-BAT was calculated retrospectively and not prospectively as in the validation studies and hence limited by the amount and quality of information recorded in the medical record, but most patients had been evaluated at the Hemophilia Treatment Center by providers with experience in obtaining and documenting a thorough bleeding evaluation. The cutoffs used for the ISTH-BAT in our study are not specific to patients with HPDs and have not been validated in our specific patient population. Since this is a retrospective observational cohort study, selection of various platelet laboratory tests was clinician dependent, which introduced variability in the evaluation, and not all patients in this cohort were evaluated with the same testing modalities.

Author notes

Overall, this study examined both clinical practice and laboratory performance of routine and esoteric platelet testing and underscores the importance of an integrated clinical and laboratory approach to diagnosing HPDs. It also accentuates the clinical utilities of PFL and PTEM and uncovers limitations of various platelet laboratory tests. The highlights of our study include the following: (1) patients with suspected HPDs have broad clinical and laboratory phenotypes; (2) patients’ ISTH-BAT scores are not necessarily associated with the severity of platelet laboratory testing abnormalities, although a positive score seems associated with abnormal PTEM findings and thrombocytopenia; (3) the current platelet laboratory tests appear complementary to each other since none of the tests by themselves provide sufficient sensitivity and specificity of detecting a platelet disorder; and (4) the standard platelet function tests such as PFA-100 and LTA by themselves are insufficient to detected platelet storage pool deficiencies, but PFA-100 CEPI could be used as a screening assay for LTA due to its high NPV of LTA abnormalities (two agonists), which may be useful for efficient testing utilization. Last, it is worth of noticing that about 30% of the patients who have evident bleeding diathesis may have completely normal platelet laboratory testing results. Our findings imply the necessity of genetic studies in HPDs and support development and availability of more comprehensive platelet functional testing to improve the laboratory diagnosis of HPDs, along with supporting the value of ISTH-BAT scoring.