Is my patient a bleeder? A diagnostic framework for mild bleeding disorders

Hematologists and hemostasis laboratories are often challenged by physicians of almost every specialty to define whether the bleeding pattern of a patient is convincingly abnormal and, if so, they ask for the “name” of the bleeding disorder. Because thrombocytopenia and acquired causes of bleeding associated with diseases such as renal failure or myelodysplasia are less often diagnostic quandaries, this review will address the so-called congenital mild bleeding disorders (MBDs), which are characterized by disproportionate bleeding after trauma, surgery, and minor injuries; easy bruising; mucosal bleeding; and menorrhagia, and most patients have a positive family history. This bleeding profile, which is generically recognized as an increased tendency to suffer from mucocutaneous hemorrhages, is the hallmark of disorders involving platelet-vessel wall interactions, eg, primary hemostasis disorders such as VWD and platelet functional disorders (PFDs). However, MBD patients also include those with mild or moderate clotting factor deficiencies, those with infrequent hyperfibrinolytic disorders, and a large population of genuine bleeders of as-yet-undefined cause. This review highlights the concept that the term MBDs is not restricted to VWD and PFDs, but also refers to patients with increased bleeding tendency regardless of whether they end up with a positive laboratory diagnosis of MBD.

Historically, research interest was centered in the earlier recognized and most severe diseases Glanzmann thrombasthenia (1918) and Bernard-Soulier syndrome (1948). These classical diseases of primary hemostasis, severe hemophilias, and type 3 and some type 2 VWD variants, are at the top of a virtual pyramid that classifies congenital bleeding disorders with regard to severity, prevalence, and diagnostic difficulty (Figure 1). These diseases have in common their low frequency, well-defined hemorrhage mechanisms, more severe bleeding, and relatively easy and straightforward diagnosis. Accordingly, these are not classified as MBDs. For the purposes of this review, we will focus our attention on the lower stages of the pyramid, where we find progressively more frequent and usually less severe disorders, but at the same time, those which entail most of the clinical and laboratory diagnostic challenges. In these patients, we find type 1 VWD, with plasma VWF levels ≤ 30%, and type 1 VWD, with VWF levels between 30% and the cutoff normal values. Also, there are miscellaneous groups of patients with mild PFDs (mainly secretion defects), mild to moderate clotting factor deficiencies, and other anecdotal defects (eg, plasma PAI-1 and α-2 antiplasmin). At the bottom of the pyramid, there is a large proportion of unequivocal bleeders, in whom all of the current diagnostic tests are repeatedly within normal ranges, precluding the diagnosis of any known hemostatic disorder. Provisionally, we designate this group as bleeder of undefined cause (BUC).

To summarize, severe bleeding disorders are less frequent, which means an easier diagnosis. MBDs are more frequent, which means increasingly major diagnostic challenges. This review describes some of these clinical and laboratory challenges confronting the diagnosis of patients with MBDs.

A first problem we deal with is the high prevalence of bleeding symptoms in healthy subjects. The lowest values reported for bleeding symptoms in a healthy population were 25% and 46% in men and women, respectively.¹  A recent report compiling the results of 10 different studies comprising healthy subjects of both sexes and both children and adults calculated mean frequencies of 23%, 20%, 28%, and 35% for epistaxis, easy bruising, gum bleeding, and menorrhagia, respectively.²  In more than 1000 young healthy women, other studies found that 73% had one bleeding symptom, 43% had 2 symptoms, and 23% had more than 2 symptoms.³  In 299 young healthy subjects “self-classified” as nonbleeders, we estimated that 19%, 25%, and 12.7% had abnormal epistaxis, ecchymoses, and gum bleeding, respectively.⁴ 

Additional difficulties encountered in the objective assessment of bleeding include the following. First, symptoms of most MBDs tend to decrease with age, and many adults do not appraise correctly the magnitude of their early life symptoms. This takes into account that many patients belong to families in which some types of bleeding are just a part of their normal life (ie, epistaxis) and their childhood consultations were therefore managed by the parents, who transmitted the anamnesis to the attending physician. Moreover, common manifestations of MBDs will be absent in patients, mainly children, who have not been exposed to some risks at the time of consultation (ie, teething bleeding, frequent trauma, surgery, and menarche). Women appear to be more prone and are more apprehensive of easy bruising. Common symptoms of MBDs are menorrhagia and postpartum bleeding, and women suffer these problems at an age when symptoms in men tend to decrease. This explains the higher prevalence of MBDs in women than men in most clinical studies.⁵  Considering all of these factors, most of the time, the clinical assessment of bleeding rests on the judgment of an experienced physician, who qualifies its severity after a careful analysis of the bleeding history. Although distinction of pathological bleeders from nonbleeders is straightforward in most cases, some patients remain in the gray category of doubtful, indeterminate pathological bleeders. This last group poses most of the difficulties for physicians regarding decisions about whether to order specialized hemostatic tests or authorize surgery and invasive procedures, and also counseling and prophylactic and therapeutic measures.

Results of studies performed to identify symptoms that best predict bleeding have been published, but with dissimilar results.^6–8  Using a standardized bleeding questionnaire in 280 patients and 299 controls, we found that menorrhagia, bleeding after surgery, tooth extraction or minor injuries, and postpartum bleeding were the most distinctive symptoms. Interestingly, bleeding related to aspirin intake was a highly useful discriminator.⁴  In this patient population, we found that the results of a quantitative, standardized bleeding score were highly correlated with the estimation of the physician who interviewed the patients and classified the severity of symptoms in 5 clinical bleeding categories (r = −0.75, P < .0001).⁹  Therefore, despite the high frequency of bleeding symptoms in a healthy population, in most instances, a trained physician should be able to discern a normal bleeding response to injury from an authentic manifestation of a bleeding disorder. Two editorial comments recount the importance of both the skill and the art of the “old-fashioned” history taking and the correct exercise of clinical judgment.^10,11 

In the real world of clinical practice, a personal, structured, and brief though complete interview should identify most patients with abnormal bleeding.¹²  The physician must assess the exposure to bleeding risks during the patient's lifetime, which are determined by age and sex; the spontaneous or provoked nature of bleeding; the number of bleeding sites, including unusual sites of bruising; the frequency, duration, and episodic nature of the bleeding with estimation of blood loss; the appropriate relationship of bleeding with injuries (including trauma, surgery, and partum) or drugs (aspirin); transfusion need or iron deficiency anemia; and first- and second-grade family history of excessive bleeding. Bleeding sites that best discriminate normal from abnormal bleeding are: menorrhagia; abnormal bleeding after surgery, tooth extraction, or minor injuries; and postpartum bleeding.^4,13 

The bleeding phenotypes of patients with MBDs are similar. There are no distinctive symptoms in patients suffering from VWD, PFDs, mild to moderate clotting factor deficiencies, and bleeders without a known hemostasis defect. The design of standardized bleeding scores to better assess bleeding severity and to distinguish normal from abnormal bleeding have been advocated in recent years, and their usefulness is discussed in another chapter. In 280 patients with MBDs, we found no differences in bleeding score between patients with VWD, PFDs, and clotting factor deficiencies. Furthermore, the clinical severity of patients with BUC was not significantly different from that of patients with known disorders (Figure 2).⁴  This reminds us that clinical history allows the classification of patients as pathological bleeders or those with dubious or no abnormal bleeding, but not to infer about specific defects. Recent prospective studies^12,14,15  have shown that clinical, qualitative assessment of bleeding is of similar predictive value to a standardized bleeding score.¹⁶  Therefore, physicians must be aware that, unless there are known disorders in the family (eg, VWD or PFDs), a custom-made questionnaire to predict a specific disease will be of no help. Furthermore, such a questionnaire would be worthless in a great fraction of “true” bleeders with unknown diagnosis.^4,5  An interesting observation of our study was related to the use of aspirin: 33% of “genuine” bleeders reported an increase in spontaneous or provoked bleeding after aspirin intake, whereas this condition was not observed in the control population. Although this observation may be useful to sort out the patients in the binary categories of bleeders and nonbleeders, we do not recommend the intake of aspirin as a diagnostic tool. The family history of bleeding was of restricted value in this assessment: it was present in more than 85% of the patients, but also in 51% of the nonbleeder controls.⁴ 

Patients with VWD, PFDs, BUC, or mild clotting factor deficiencies have similar bleeding profiles. Accordingly, bleeding history does not allow the inference of specific diagnosis. A structured, simple, qualitative, oral recording of bleeding history^12,13  seems to be as informative and predictive as a standardized bleeding score.

Most mild hemophilias A and B and rare clotting factor deficiencies (eg, factor V [FV] and FXI)^17–19  also qualify as MBDs, with mucocutaneous hemorrhages as cardinal manifestations. Although the global coagulation assays activated partial thromboplastin time (aPTT), prothrombin time (PT), and thrombin time (TT) are nonspecific, they have enough sensitivity to detect generic defects of intrinsic and extrinsic clotting pathways when plasma factor levels decrease below their hemostatic threshold.¹⁸ 

In contrast, global tests to detect primary hemostasis defects are not specific and their sensitivity is low in MBDs. Moreover, the routine use of bleeding time (BT) has been questioned for being invasive, operator dependent, and of low predictive value.²⁰  Two recent guidelines do not recommend the use of BT in the investigation of VWD¹³  and hereditary PFDs.²¹  In our prospective study in 280 unequivocal bleeders, the Ivy BT was prolonged in only 27% of them (41% in VWD and 39% in PFDs).⁴  Interestingly, the BT was the only abnormal test in 19% of 167 patients with BUC, suggesting that some defect(s) in platelet-vessel wall interaction could be involved in the pathogenesis of hemorrhage in this group. In an increasing number of laboratories, the BT is no longer on the list of hemostasis assays and physicians are progressively restricting their orders for the test.²² 

Among the various devices conceived to replace the BT, closure time measurement using the platelet function analyzer (PFA-100) was designed as an ex vivo exploration tool of primary hemostasis under conditions of high shear. This test has proven more sensitive than template BT for testing patients with previously diagnosed VWD,^23,24  but less sensitive in those with mild platelet secretion disorders,²⁵  the most frequent disorders found with PFDs.²⁶  Two studies have assessed the value of BT and closure times in unselected incoming patients with MBDs.^27,28  Both found a higher sensitivity of PFA-100 over BT in patients with VWD-1, and in one of these reports this advantage was extended to patients with PFDs.²⁸  A recent study found even lower sensitivity of the test in patients with known platelet storage pool deficiency.²⁹  In our prospective study,²⁷  the overall sensitivity of closure time in VWD and PFDs was 50% and 18%, respectively. In the larger group of bleeders with undefined diagnosis, the abnormality of the test was approximately 15%.²⁷ 

Patients with hereditary increase in fibrinolytic activity, such as those with Quebec platelet disorder³⁰  and deficiencies of fibrinolytic inhibitors,^31,32  suffer mucous membrane and skin bleeding as their main clinical manifestations. The global fibrinolysis tests, such as clot lysis time in saline solution or euglobulin clot lysis time, have low sensitivity in disclosing mild and moderate hyperfibrinolytic activity in plasma. A recent report found lower mean values of plasminogen activator inhibitor-1 (PAI-1) in patients with MBDs than in controls.³³  Had these findings been confirmed, measurement of PAI-1 activity, and perhaps other components of the fibrinolytic system, could help in the diagnosis of these patients. Patients with severe deficiency of subunit A of FXIII present with severe, life-threatening bleeding, usually late after minor trauma. A simple clot lysis time assay in 5M urea solution usually detects this rare defect. Mild or moderate FXIII deficiencies (< 31 U/dL)¹⁸  appear to be more common than previously thought, but of weak clinical significance.³⁴ 

Screening of MBDs must include the global clotting tests PT and aPTT (and, optionally, TT or fibrinogen),¹³  for detecting clinically significant deficiencies of intrinsic and extrinsic clotting pathways. After ruling out acquired inhibitors, identification of the abnormally low clotting factor(s) is indicated. In contrast, template BT and PFA-100 closure time have unacceptably low sensitivity for screening mild hemostatic defects, and their routine use is therefore not recommended. Moreover, because these tests lack specificity, additional assays are required to positively diagnose VWD, PFDs, or other diseases.

The state of the art for diagnosing VWD is provided in a separate chapter. After a long-time discussion about criteria for diagnosing type 1 VWD and widely circulated guidelines, it is surprising that the results of a recent survey³⁵  detected that only a minority of North-American laboratories followed the National Heart, Lung and Blood Institute (NHLBI) recommendations¹³  (available at (http://www.nhlbi.nih.gov/guidelines/vwd). Difficulties in the diagnosis of type 1 VWD include the lack of distinctive genetic markers.³⁶  Moreover, the statistical distribution of laboratory analytes sorts a fraction of the healthy population, usually 2.5%, below the cutoff values for each variable. Therefore it is unavoidable that approximately 2.5% of the population will fall below the cutoff levels, with a high chance of coincidental association of low VWF with bleeding symptoms in the general, nonselected population.¹  In addition, normal levels of plasma VWF have a wide distribution range, with lower cutoff values overlapping with the plasma levels of patients with genuine VWD (Figure 3), adding further difficulty in the diagnosis. Plasma VWF behaves as an acute-phase reactant protein precluding or retarding the diagnosis of VWD in patients with inflammation, infection, pregnancy, cancer, and acute stress. Moreover, plasma levels of VWF increase with age, which makes diagnosis more difficult as the patient gets older. Individuals with blood type O have approximately 25 IU/dL lower levels of plasma VWF than those of the A, B, and AB types. Because type 1 VWD is defined by low plasma VWF and its deficiency is the bleeding risk factor, current guidelines recommend using a single cutoff value independent of blood ABO type. Therefore, diagnosis of VWD is made more frequently in patients of blood type O than in those with A and B phenotypes.¹³  Finally, from a clinical point of view, plasma VWF levels and bleeding severity are weakly correlated.¹  In our study of 50 patients with VWD, the bleeding scores of those with VWF ≤ 15 IU/dL and those with VWF ranging between 15 and 40 IU/dL were not significantly different.⁴  A recent study also showed that levels of VWF:RCo ≤ 30 IU/dL (definite type 1 VWD) were not better predictors of surgical bleeding than levels of VWF:RCo 31-49 IU/dL (possible VWD).³⁷ 

We diagnose type 1 VWD when patients have a history of mucosal and skin bleeding, usually with positive family history and with reduction below 2.5 percentile in at least 2 VWF variables: VWF:Ag, VWF:RCo, or VWF:CBA. Confidence in the diagnosis increases in patients with VWF levels < 30 IU/dL. Physicians must be aware that lower levels of plasma VWF reflect better the steady-state, resting concentration than higher VWF levels. Many concomitant, acquired physiological (eg, aging or pregnancy) or pathological (eg, infections or inflammation) acute or chronic conditions increase plasma VWF levels, whereas lifetime decreases in plasma VWF are exceptional (ie, in autoantibody-mediated acquired VWD). Therefore, convincingly symptomatic patients with normal or elevated plasma VWF almost always require confirmation testing, whereas low, diagnostic VWF levels almost always constitute evidence of VWD.

Born and O'Brien introduced light transmission platelet aggregometry (LTA) into clinical practice 50 years ago. Despite being labor demanding and expensive, LTA is still the gold standard for diagnosing PFDs. However, only recently have several guidelines on how to perform and interpret the test been updated.^21,38–40  There is expert consensus that LTA should be performed in specialized hemostasis laboratories.^21,41  Historical problems with the assay include lack of standardization, unknown reproducibility, and a lack of universally accepted diagnostic criteria. The majority of cases of inherited PFDs have in common mild impairment of platelet aggregation and secretion with various agonists due to signal transduction defects, the underlying molecular mechanisms of which are largely unknown.^42,43  This review focuses on these patients, because their diagnosis is more difficult and less accurate than those with severe dysfunctions such as Glanzmann and Bernard-Soulier diseases, rare disorders sharing distinctive and reproducible aggregation profiles and straightforward, precise methods for diagnosis. Most mild PFDs have no demonstrable abnormalities of major membrane glycoproteins (except glycoprotein VI and integrin Ia-IIa defects) and are more frequent than commonly thought in different populations. Among 280 patients with unequivocal pathologic bleeding, 65 (23.2%) had abnormal LTA vs 50 (17.9%) with VWD and 11 (3.9%) with mild clotting factor deficiencies.⁴  Gupta et al⁴⁴  also observed a higher prevalence of PFDs than VWD in an Indian population.

In another study, we showed that the diagnostic consistency of LTA is high when using the appropriate reference values and predefined diagnostic criteria, but also if careful analysis of LTA tracings by experienced operators is included.⁴⁵  In this context, in 213 patients with inherited MBDs, normal and abnormal LTA tests were reproducible in 93.3% and 90.4%, respectively. Mean intrasubject coefficients of variation for LTA with strong agonists (ie, collagen and arachidonate) were < 9%, and the mean intraclass correlation coefficients for weak agonists (ie, epinephrine and ADP) were > 0.86 (P < .0001). Remarkably, only a primary wave of aggregation with 10μM adrenaline associated with reversible aggregation with 4μM ADP was observed in 13.7% of the controls (Figure 3). For this reason, the combined defects of aggregation with epinephrine and low ADP concentrations in a patient preclude diagnosing PFDs, and repeated testing is therefore required to rule it out. The positive counterpart for this observation is that LTA with these weak agonists may be highly informative: LTA ≥ 42% with 10μM adrenaline or irreversible LTA with 4μM ADP are strong predictors of a normal platelet function, with 93% and 95% negative predictive values for diagnosing PFDs, respectively. It is remarkable that these aggregation defects are consistently associated with abnormal serotonin secretion,⁴⁵  indicating that a primary impaired secretion (either by a signal transduction defect or lack of storage granules) gives count of the impaired aggregation.^26,42 

It has been well known for more than 2 decades that some patients with secretion defects due to storage pool deficiency have normal LTA.^46,47  We found that 9 of 65 (14%) patients had normal LTA with defective serotonin secretion. This constitutes a sound basis to recommend the simultaneous measurement of both parameters.^41,48  [¹⁴C] or [³H]-serotonin (5-HT) release by platelets is the gold standard for measuring platelet secretion. The test is highly reproducible: in 136 assays performed in 77 patients with repeatedly abnormal LTA, concordant results were obtained in 75 (97.4%). Nevertheless, limitations for using radioactive compounds in clinical laboratories have replaced this assay by the simultaneous measurement of platelet aggregation and ATP secretion using either platelet-rich plasma (PRP) or whole blood lumiaggregometry. The test in PRP has been subjected to more in-depth assessment, revealing a large within- and between-subject variability, which cast some doubts on its diagnostic power.⁴⁸  Moreover, among patients with common platelet dysfunctions diagnosed by standard LTA instruments, the LTA measured in the lumiaggregometer in presence of Chrono-Lume reagent was normal in approximately 58% of them. One possible explanation for this high discordance is the presence of Mg²⁺ ions in the reaction mixture.⁴⁹  The inconvenience of radioactive serotonin assay and the pitfalls of ATP release by lumiaggregometry must stimulate alternative means to evaluate platelet secretion. Measurement of endogenous 5-HT secreted by platelets seems the most promising approach, because 5-HT is an ideal secretion marker. First, essentially all of the blood 5-HT (4000-5000 nmol) is transported inside the platelets, with no 5-HT transported by other blood cells, and negligible or absent plasma levels of the monoamine. Second, unlike ATP, all of the 5-HT in platelets is stored in dense granules with no meaningful pools in other organelles. Third, careful blood sample manipulation does not result in loss of 5-HT from platelets.⁵⁰  Therefore, measurement of 5-HT secretion by nonradioactive assays such as ELISA or HPLC could become standard methods in the near future.

Several point-of-care instruments to evaluate platelet function have been devised in recent years. A large number of reports have been published, but almost all of them were conducted to assess the effect of antiplatelet drugs or platelet resistance to these drugs or to have a fast answer on bleeding risks before invasive procedures. These instruments, which were not initially conceived for the diagnosis of bleeding disorders, include whole blood aggregometry and lumiaggregometry, which still lack validation for diagnosis of PFDs.²¹  Traditional whole-blood lumiaggregometry lacks sensitivity to weak platelet agonists, and a new generation of this instrument has not been clinically validated for the diagnosis of heritable PFDs.⁵¹ 

From a clinician's point of view, it may not be easy to explain the patient's symptoms by the defects in LTA and platelet secretion. This is a problem in patients with mild PFDs, but not in those with severe disorders such as Glanzmann thrombasthenia. In fact, it is remarkable that an overwhelming majority of patients taking aspirin do not complain of abnormal bleeding despite having an unquestionably more severe and extended platelet dysfunction ex vivo than most of the patients with hereditary PFDs (Figure 4). Other drugs and food products (eg, NSAIDs, β-lactam antibiotics, and curcumin)²¹  also affect or may affect these tests, most without clinical consequences. Again, this reflects the weak correlation between abnormalities of ex vivo platelet function tests and clinical symptoms, analogous to the weak correlation between bleeding severity and plasma VWF.⁵²  As mentioned previously, neither the indices of platelet dysfunction nor the antigen or activity of plasma VWF are significantly correlated with bleeding severity in patients with MBDs.⁴  Interestingly, zero of 185 healthy controls versus 33% of patients with inherited MBDs complained of excessive bleeding after aspirin. This last observation supports the notion that mild PFDs might be better considered as risk factors rather than univocal bleeding causes. This is similar to the multifactorial risk influences on thromboembolic disorders, in which an accumulation of weak congenital and/or acquired risk factors finally triggers a thrombotic event.^53,54  A recent report on more than 30 000 pregnant women and a similar number of controls showed an statistical interaction of mild hemostatic risk factors with severe postpartum bleeding.⁵⁵  This concept would be useful to ponder the relevance of subtle “defects” in LTA and secretion assays in the clinical context of the patient.

In centers with specialized hemostasis laboratories, we recommend simultaneously investigating VWD and PFDs. Both disorders have similar prevalence and clinical bleeding patterns, and are not unusually present concomitantly in the same patient. Moreover, considering that a significant fraction of patients with PFDs present with isolated secretion impairment and normal platelet aggregation, this recommendation will explore both processes simultaneously. Currently, the best assay licensed to perform both measurements is PRP-lumiaggregometry. Clinicians should be aware that the current format of this assay has limitations: it tends to potentiate the aggregation tracing,⁴⁹  and the measurement of ATP secretion shows a large coefficient of variation.⁴⁸  In case of abnormal results, a repeated study should be carried out, ideally using LTA and an independent assessment of platelet secretion.⁴¹ 

For small community hospitals where platelet function testing may not be available, our recommendation is to start ruling out VWD and then refer the patient to a specialized hemostasis laboratory.

Analyses of many clinical studies on patients with previously undiagnosed nonthrombocytopenic MBDs have consistently shown that, after wide-ranging laboratory explorations, between 47% and 69% of the patients will remain undiagnosed. Table 1 shows that this is a common finding in many reports,^{4,12,28,33,44,56–58}  confirming observations in outpatient populations with MBDs,⁵⁹  epistaxis,⁶⁰  or menorrhagia.⁶¹  Our prospective study in incoming patients revealed a testing efficacy of only 40.4%. The remaining 59.6% had no VWD, PFDs, or clotting factor deficiencies and were therefore entered into the BUC category.⁴  Moreover, the normality in hemostatic tests of these patients (Figure 3) persists after exhaustive and repeated testing. Patients with BUC are clinically indistinguishable from those with known MBDs (Figure 2). In fact, the nature, sites, and severity of bleeding are similar, including those with VWD and the lowest values of VWF.⁴  This, along with the abnormal BT in a fraction of these patients (19% in our study),⁴  suggests that patients with BUC have a primary hemostasis disorder. In these patients, empirical treatment with desmopressin and/or tranexamic acid is indicated to stop or prevent bleeding during surgery, invasive procedures, and menorrhagia. In this regard, we hypothesize that BUC represents a distinctive and likely heterogeneous group of disorders of high frequency and unknown pathogenesis. Currently, we can only speculate on the nature of BUC. For example, factors derived from the endothelium that affect platelet-vessel wall interaction, such as NO and PGI2, have been shown to prolong the BT,^62,63  but their very short half-lives preclude their detection by current ex vivo platelet testing. In addition, tissues with high fibrinolytic activity may contribute to bleeding⁶⁴  or subtle decreases in platelet procoagulant activity, previously known as platelet factor 3 availability defect, may also be involved.⁴⁴ 

Bleeding in MBDs is likely multifactorial, as is the case with thrombophilia,^53,54  and the frequency of combined hemostatic defects is greater than that predicted by chance.⁴  Therefore, the accumulation of subtle impairments in primary and/or secondary hemostasis functions, perhaps associated with a slightly increased fibrinolytic activity, would tilt the hemostatic balance toward bleeding.⁵⁵  For patients with BUC, this multifactorial approach would not require that the function of one or more hemostatic components decrease below their cutoff levels, but that several factors converge and assemble in the lower functional range of their distribution. This notion could explain, at least in part, why bleeding in some of these patients may be empirically controlled or attenuated with such diverse therapeutic measures as desmopressin, inhibitors of fibrinolysis, plasma factors, or platelet transfusions. In the same context, the quantitative and functional enhancement of hemostatic factors with aging may explain the improvement of bleeding episodes after puberty.

Better standardization of current laboratory assays, progress in the knowledge of fibrinolytic mechanisms and their laboratory evaluation, new understanding of the factors contributing to platelet-vessel wall interaction, and the corresponding development of laboratory tools should improve our capacity to diagnose a greater proportion of patients with MBDs.

We still ignore the pathophysiology of bleeding in approximately half of the patients with congenital MBDs. A normal platelet aggregation and secretion in a first work-up constitutes strong evidence that the patient does not have any of the frequent platelet defects; interfering conditions (eg, drugs or analytical errors) result in loss of ex vivo platelet function, not in platelet hyperfunction. Therefore, we feel that repeated testing should include mainly VWF and aPTT-PT, given that an initial testing may be normal by the inflammatory or acute-phase reactant nature of VWF and some clotting factors. If hemostatic tests are repeatedly normal, the patient should be classified in the BUC category. Empirical treatment or prophylaxis with desmopressin (0.3-0.4 mg/kg) and/or tranexamic acid (25-50 mg/kg) usually controls or prevents excessive bleeding in these patients.

Conflict-of-interest disclosure: The authors declare no competing financial interests. Off-label drug use: None disclosed.

Teresa Quiroga, MD, Director, Department of Clinical Laboratories, School of Medicine, Pontifical Catholic University of Chile, Av Vicuna Mackenna 4686, Santiago 7820436, Chile; Phone: 56-2-3548537; Fax: 56-2-5523464; e-mail: tquiroga@med.puc.cl.