Article Text
Abstract
A mistaken diagnosis of child abuse can occur in a number of medical conditions, many of which can be readily diagnosed by experienced paediatricians. Bleeding disorders offer a greater challenge, especially when court proceedings may demand their exclusion. Some of these disorders are rare but more prevalent in areas which have a high incidence of consanguinity. We advocate two stages of laboratory investigations but the limitations of some of these tests and their inability to exclude a bleeding disorder with absolute certainty should be recognised. However, if personal and family histories are absent and both first-stage and second-stage investigations are normal, it is highly unlikely that a bleeding disorder will be missed.
Statistics from Altmetric.com
Determining the cause of bruising in the context of child abuse is often challenging; the paediatrician has to differentiate between accidental bruising, inflicted bruising and bruising due to a medical condition, especially a haematological disorder. Furthermore, the presence of a bleeding disorder does not exclude abuse and places the abused child at a significantly greater risk. A number of publications have documented a mistaken diagnosis of non-accidental injury in children with haemophilia, von Willebrand disease, idiopathic thrombocytopenic purpura, Glanzmann thrombasthenia, platelet aggregation disorders and factor XIII deficiency.1,–,5 One of the authors, whose practice is in an area with a high incidence of consanguinity and autosomal recessive disorders has, over a 15-year period, had six patients with inherited platelet disorders and one patient with factor XIII deficiency whose initial referral was for suspected non-accidental injury. A mistaken diagnosis is likely to cause the family huge distress, deny the child adequate treatment and even have life-threatening consequences.4
Non-Haematological Conditions Which May be Mistaken for Non-Accidental Bruising
A number of diverse conditions have been mistaken for non-accidental bruising (box 1), many of which have been described by Wheeler and Hobbs.3 Of the infection-related conditions, meningococcal sepsis should not often cause difficulty but occasionally viral infections presenting with fever and petechiae (most often on the legs) can cause difficulty in infants. Most of the conditions listed in box 1, for example, blue spot, erythema nodosum and Henoch–Schönlein purpura will be readily diagnosed by experienced paediatricians, although the initial referral may have been made by someone less experienced. Blue spot, which is common in non-Caucasian babies, may be quite extensive and has at times been mistaken for bruising. Cultural practices such as the Vietnamese lay practice of coin-rubbing may result in dramatic bruising. As suggested by Wheeler and Hobbs, this should be seen as a culturally sanctioned form of abuse.3 Blue ink or paint marks on the face of a toddler can look remarkably like bruising until the mother, who knows what the marks are, wipes them off!
Box 1 Non-haematological conditions, which may be mistaken for non-accidental bruising
Vasculitic
Infection-related – meningococcal, streptococcal, viral
Drug-related
Erythema nodosum
Henoch–Schönlein purpura
Vascular
Blue spot
Capillary haemangioma
Prominent facial veins
Connective tissue disorders
Ehlers–Danlos syndrome
Hypermobility syndrome
Scurvy
Traumatic
Subconjunctival haemorrhage secondary to coughing or vomiting
Cultural practices such as coining or cupping
Self-harm
Artefactual
Ink, paint or dye marks
Factitious bruising
Haematological Disorders Which may Present with Bruising
The process leading to clot formation can be divided into the initial phase, primary haemostasis, and a second phase, secondary haemostasis. Primary haemostasis describes the interaction of platelets and the damaged blood vessel wall leading to platelet adhesion. The components of coagulation required for this phase are the endothelium, von Willebrand factor and functioning platelets. Secondary haemostasis describes the process by which the aggregated platelets are bound together by the formation of fibrin. This requires a functional phospholipid surface on the platelet, the coagulation enzymes of the intrinsic, extrinsic and final common pathways and factor XIII which cross-links the fibrin molecules. The process of fibrinolysis will lyse previously formed fibrin clots so that the presence of inhibitors of this process including plasminogen activator and alpha-2-antiplasmin is important for the stability of the clot. Defects of both phases of haemostasis will lead to bleeding symptoms including bruising. The principle symptoms of failure of primary haemostasis are bruises, petechiae and bleeding from mucosal membranes including the oral cavity, the nose and the uterus. Failure of secondary haemostasis generally leads to bleeding in deeper tissues such as muscles, joints and internal cavities. The distinction in patterns of bleeding is, however, not complete.
An acquired bleeding tendency may be seen in liver disease, renal disease, disseminated intravascular coagulation, leukaemia, neuroblastoma and bone marrow failure but these disorders are likely to have additional features. Thrombocytopenia, which may also be related to infections or drugs, is most often due to idiopathic thrombocytopenic purpura in the child who is otherwise well. Bleeding (often severe) occurring in early infancy may be due to vitamin K deficiency. Inadequate vitamin K prophylaxis at birth and exclusive breast feeding are contributory factors. Therapeutic use and accidental ingestion of drugs including warfarin, heparin, aspirin and other non-steroidal anti-inflammatory drugs (NSAIDs) may also cause a bleeding tendency.
Inherited bleeding disorders are due either to coagulation factor deficiencies or the rarer congenital thrombocytopenias and platelet function disorders (table 1). Although inherited coagulation factor disorders generally reflect a deficiency of a particular coagulation factor, abnormal molecular structure or arrangement may lead to dysfunctional proteins despite normal protein quantity as seen in type II von Willebrand disease or dysfibrinogenaemia. Inherited platelet disorders have been categorised in a recent United Kingdom Haemophilia Centre Doctors’ Organisation guideline as defects in platelet numbers, severe disorders of platelet function such as Glanzmann thrombasthenia and Bernard–Soulier syndrome, disorders of receptors and signal transduction, disorders of platelet granules and disorders of phospholipid exposure (box 2).6
Box 2 Inherited disorders of platelet numbers and function
Disorders of platelet number
MYH9 disorders
May–Hegglin anomaly
Sebastian syndrome
Fechtner syndrome
Epstein syndrome
Congenital amegakaryocytic thrombocytopenia
Amegakaryocytic thrombocytopenia with radioulnar synostosis
Thrombocytopenia/absent radius syndrome
X-linked thrombocytopenia with dyserythropoiesis
Severe disorders of platelet function
Wiskott–Aldrich syndrome
Glanzmann thrombasthenia
Bernard–Soulier syndrome
Disorders of receptors and signal transduction
Platelet cyclo-oxygenase deficiency
Thromboxane synthase deficiency
Thromboxane A2 receptor defect
ADP receptor defect (P2Y12)
Disorders of the platelet granules
Idiopathic dense-granule disorder (δ-storage pool disease)
Hermansky–Pudlak syndrome
Chediak–Higashi syndrome
Grey platelet syndrome
Paris–Trousseau/Jacobsen syndrome
Idiopathic α- and dense-granule storage pool disease
Disorders of phospholipid exposure
Scott syndrome
Clinical History
The generic enquiry “does he or she bruise easily” may not provide the clinician with the details to accurately abnormal have constructed bleeding scores in an attempt to provide a quantitative evaluation of the severity of the bleeding history. These have been mainly evaluated and used in patients with von Willebrand disease; their application to other disorders will require further validation.7 Defects of primary haemostasis (von Willebrand disease, thrombocytopenia or a disorder of platelet function) are associated with mucosal bleeding such as epistaxis, gum bleeding and menorrhagia. Symptoms such as gum bleeding or epistaxis are more relevant if recurrent or prolonged. Frequent, prolonged and bilateral nose bleeds are more likely to be due to a bleeding disorder than the occasional, short, unilateral nose bleed. A useful definition of a significant history of epistaxis is one of more than five episodes or individual episodes of greater than 10 minutes’ duration.8 It is important to characterise the extent of bleeding after haemostatic challenges such as delivery, heel prick in the neonatal Umbilical inherited feature of factor XIII deficiency. The absence of significant bleeding after surgical procedures such as circumcision, tonsillectomy or dental extractions largely excludes a clinically significant inherited bleeding disorder. A history of poor wound healing, abnormal scars, sprains or joint dislocations would suggest a connective tissue disorder such as Ehlers–Danlos syndrome. The clinical history is increasingly useful as the child ages since the number of haemostatic challenges encountered will increase. The absence of significant clinical events limits the benefit of a negative clinical history. It is important to ask about the child’s general health and to enquire about symptoms relating to malabsorption, liver disease and kidney disease. Details of recently taken drugs are important as NSAID ingestion is the most common cause of abnormal platelet function and will affect platelet aggregation tests.
Family History
It is insufficient only to ask if there is a family history of a bleeding disorder. One should enquire specifically about epistaxis, menorrhagia, bleeding following surgery, dental extractions or delivery or bleeding in the perinatal period in first-degree relatives. A positive history is more likely in von Willebrand disease which, apart from the rare type 3 variant, is inherited as an autosomal dominant. Haemophilia A and B are X-linked conditions and the boy’s grandfather or uncles may have been affected although 30% of cases result from a spontaneous mutation. The rarer inherited bleeding disorders are inherited as autosomal recessive disorders and a family history may be absent. These disorders are more common in areas which have a high incidence of consanguinity and parents should be asked if they are related. If there is consanguinity, important diagnostic clues may occasionally come to light by asking about bleeding disorders in the extended family, sometimes living in another part of the country or abroad.
Examination
Assessment of growth and a full general examination should always be part of the assessment. Occasionally, growth faltering may be due to chronic illness such as malabsorption, renal disease or liver disease. In the context of a haematological disorder, signs of anaemia, lymphadenopathy, hepatomegaly and splenomegaly should be sought. Joints should be examined for evidence of hypermobility and scars examined. Thin, paper tissue scars, most often seen on the knees, together with hypermobile joints would suggest Ehlers–Danlos syndrome, a condition associated with a tendency to bruise easily.
In some cases of suspected abuse, it may not be possible on clinical grounds to distinguish bruising due to a haematological disorder from nonaccidental bruising. Some bruises strongly indicate abuse irrespective of the presence or absence of a bleeding disorder; for example, bruising caused by an implement or slap but multiple bruises, bruises of different common to found in normal children on the shins, knees, forearms and, in toddlers, on the forehead. The following features are more suggestive of abuse: a large number of bruises or clustering, the presence of bruises in unusual sites, especially the ear, face, neck, trunk, buttocks, genitalia or on soft parts of the body, and bruises in a child who is not yet mobile.9 10 Petechiae occur typically in idiopathic thrombocytopenic purpura or the severe disorders of platelet function. When confined to the area of superior vena cava drainage, they may be due to coughing or vomiting. Petechiae of the head and neck also occur which case they vasculitic lesions may be mistaken for bruising but the characteristic distribution involving the ankles, extensor surfaces and buttocks should lead to the correct diagnosis.
Investigations
A number of factors may determine the extent of investigations in cases of suspected abuse. In children with fractures which are typical of abuse, the exclusion of a bleeding disorder as a cause of accompanying bruises is less relevant and extensive tests unwarranted. We do believe that blood tests are generally not indicated when the only bruising is clearly the result of a slap or a blow with an implement. However, perpetrators may argue that the extent of bruising may have been increased by an underlying haemostatic defect so that the trauma involved was in fact minimal. This scenario would justify investigation of haemostasis. The presence of a suggestive past history, a strongly suggestive family history or the presence of consanguinity will call for more extensive testing to exclude rarer disorders if initial tests are normal.
It is usual practice in the investigation of a possible bleeding disorder to initially request screening investigations followed if necessary by investigations for the rarer disorders. It is important to realise that investigations have a limited ability to exclude with certainty the presence of all bleeding disorders. It is relevant to ask what degree of certainty is required. The more extensive the range of tests performed, the higher the degree of certainty is achieved in excluding a bleeding disorder. However, the more obscure the test, the less the incremental value in the exclusion of a bleeding disorder. In the case of suspected child abuse, a high degree of certainty may be sought before a court reaches a decision. However, the investigation of rare bleeding disorders may be of only limited relevance when addressing evidence of trauma.
The usefulness of individual tests depends on the accuracy of the result, the quality of normal ranges and the degree with which a clinically relevant disorder can be predicted by that test. Some bleeding disorders can be identified with relatively straightforward other disorders between laboratory result and clinical bleeding phenotype in rare disorders may be poor. The diagnosis of a bleeding disorder is often not straightforward and discussion between the paediatrician and an experienced paediatric haematologist is important.
Coagulation investigations have variable precision and are subject to pre-analytical and analytical variables. Pre-analytical variables include the stress and technique of venepuncture and drug ingestion. A difficult, traumatic venepuncture may distort results due to contamination with tissue fluid. Correct filling of the specimen bottle is essential and blood samples should not be taken from a heparinised cannula. Stressful venepuncture may elevate the plasma concentration of certain factors including factor VIII and von Willebrand factor. Consequently, multiple repeat testing of cases of suspected von Willebrand disease may be needed if discrepant results are obtained. Prior to taking the sample, one should check that the child has not been taking warfarin, aspirin or other NSAIDs during the previous 2 weeks. It is important that investigations are carried out in a laboratory experienced both in working with smaller blood samples and in carrying out the individual tests or assays. The majority of routine investigations and factor assays are assurance schemes laboratory variation and a reliably accurate result. This is not universal, however; a report on factor XIII assays from the UK National External Quality Assurance Scheme (NEQAS) showed that although severe deficiency was reliably detected, moderate deficiency was subject to greater error.11 Assays that require assessment of platelet reactivity have significant potential for poor precision. These include the ristocetin cofactor (RiCof), which is used to determine the activity of von Willebrand factor and assessment of platelet function. A recent report from the NEQAS indicated that there was very poor agreement among centres where a standard sample was distributed for RiCof analysis.12
Assessment of platelet function can be challenging and there is no completely effective screening test. Platelet aggregation in response to a number of agonists such as adenosine diphosphate (ADP), collagen, arachidonic acid and ristocetin is still the initial choice in assessing platelet function. Abnormalities in platelet aggregation are consistently present in the two most severe disorders of inherited platelet function, Glanzmann thrombasthenia and Bernard–Soulier syndrome, but are more variable in less severe disorders. Analysis by the NEQAS indicated that the methodology used to assess platelet function is very variable. In many cases, no form of quality control is employed. It is not uncommon to see variable results in the same patient and repeat testing is often indicated before a conclusion can be reached. The measurement of platelet nucleotides adenosine triphosphate and ADP content and release is useful in the diagnosis of storage pool and release defects. Normal platelet aggregation does not exclude the diagnosis of storage pool disease.6 Use of the Platelet Function Analyser-100 (Dade Behring, Marburg, Germany) has increased over the last decade, largely to replace the bleeding time. This test can reliably detect von Willebrand disease and major platelet function disorders such as Glanzmann thrombasthenia and Bernard–Soulier syndrome but is insufficiently sensitive to reliably detect minor platelet function disorders. It should be appreciated that although a very abnormal result false-negative results, storage pool disorder are not uncommon. If the only need is to exclude Bernard–Soulier syndrome or Glanzmann thrombasthenia, flow cytometry of platelets has the advantage of requiring a smaller volume of blood for testing, an important consideration in infants. This technique utilises the binding of fluorescent antibodies with high specificity to antigens expressed on the cell surface. The fluorescence of individual cells is established by analysis of single cells in single droplets by a laser of fixed wavelength; positive fluorescence establishes the presence of the antigen. Flow cytometry permits the rapid analysis of thousands of cells from a small sample volume.
It is important to appreciate that the normal ranges for some tests depend on the age of the patient – factors II, IX, X, XI and the activated partial thromboplastin time (APTT) may not reach adult levels until 6 months post-term. It is difficult, for example, to diagnose mild factor IX deficiency during the first few months of life. Diagnosis of von Willebrand disease in the early weeks of life is also difficult because the level of von Willebrand factor is elevated at this time. Ideally, normal ranges should be obtained from normal subjects tested in each individual laboratory but this is ethically difficult in children. Therefore normal ranges are frequently based on published data.13 14 There are concerns regarding the use of published normal ranges. The laboratory methodology used to determine published normal ranges may have been different to that used in the hospital laboratory of the patient and the subjects used to determine the normal range may not be completely comparable with the subject under investigation.15 Rare disorders tend to have less gives useful information, for example, in platelet Table 2 First-stage investigations well-established normal ranges especially in the younger age range. These limitations must be considered when published normal ranges are used since potential inconsistencies in interpretation of results may follow.
First-Stage Investigations
Suggested first-stage investigations are shown in table 2. These are more extensive than those suggested in the Royal College of Paediatrics and Child Health “Child Protection Companion”.16 Screening for von Willebrand disease is included in our list of first-stage investigations as this is by far the most common inherited bleeding disorder and the APTT is often normal. Similarly, the APTT may be normal or slightly prolonged in boys with mild factor VIII or factor IX deficiency. We therefore agree with Thomas that these tests should be included in first-stage investigations.17 We have not included bleeding time in our list as this test is poorly reproducible, invasive, insensitive and time-consuming. It is frequently normal or only minimally prolonged in mild platelet function disorders. Although thrombocytopenia will be detected in our first-stage investigations, abnormal platelet function will not and we have deferred tests of platelet function to investigation because platelet require a large volume of blood. Examination of a blood film by a haematologist is important not only to detect evidence of leukaemia but to examine platelet size. Mean platelet volume should be requested for patients with an isolated thrombocytopenia. Large platelets are characteristic of Bernard–Soulier syndrome in which thrombocytopenia is accompanied by abnormal platelet function and a severe bleeding phenotype. Small platelets are characteristic of Wiskott–Aldrich syndrome.
In certain situations, it may be appropriate, in first-stage investigations, to test for certain rarer disorders which have normal standard coagulation tests such as prothrombin time (PT) and APTT. Although rare, factor XIII deficiency is an important diagnosis to make because of the high incidence of intracranial bleeding and the availability of effective prophylactic treatment with factor XIII concentrate.18 Previous umbilical bleeding, mucocutaneous bleeding and consanguinity would influence the need to exclude this disorder. Flow cytometry to detect abnormal platelet membrane glycoprotein expression in Glanzmann thrombasthenia and Bernard–Soulier syndrome should also be considered if severe bruising or consanguinity is present.
The diagnoses suggested by different patterns of abnormal first-stage coagulation tests are shown in table 3.
Second-Stage Investigations
Any abnormal results found as a result of firststage investigations should be repeated after discussion with a paediatric haematologist. If the APTT or PT is prolonged, correction studies with normal plasma may be considered. Correction of prolonged coagulation time following addition of normal plasma suggests a factor deficiency; lack of correction is more suggestive of an inhibitor, most commonly a lupus anticoagulant. Isolated prolongation of APTT would lead to assay of factors VIII, IX, XI and XII, von Willebrand screening (even if normal in first stage) and testing for lupus anticoagulant. Factor XII deficiency is not associated with a bleeding tendency; nor is the presence of lupus anticoagulant in the absence of acquired prothrombin deficiency or associated autoantibodies against other coagulation factors. Prolongation of PT or prolongation of PT and APTT may be due to vitamin K deficiency. This would be suggested by a reduction in concentrations of factors II, VII, IX and X and the presence of proteins in the absence of Vitamin K. A repeat screen for von Willebrand disease should be considered if the results are borderline or if the personal history or family history is suggestive despite the initial tests being normal.
Thrombocytopenia found on initial testing may be spurious due to platelet clumping and the platelet count should be rechecked unless the clinical assessment is highly suspicious and rapid correction of thrombocytopenia is required. Although idiopathic thrombocytopenic purpura, infection, drugs and leukaemia are the most likely causes, occasionally the child may have an inherited disorder and discussion with a paediatric haematologist is appropriate.
If all the first-stage investigations are normal, then second-stage investigations listed in box 3 should be carried out, again after discussion with a paediatric haematologist. Individual factor assays should be performed even when initial PT and APTT are normal. A high factor VIII may mask other factor deficiencies such as factor XI deficiency by “normalising” the APTT. A heterozygous factor deficiency leading to a mild bleeding disorder will not be the sole cause of significant bruising but may increase the degree of bruising if there has been associated trauma. Normal haemostasis is achieved with abnormally low levels of factors II, V, VII and X, therefore a result outside the normal range will not necessarily be the cause of bruising. If not included in the first-stage investigations, factor XIII should be assayed in the second-stage investigations. Platelet function testing should be undertaken.
Box 3 Second-stage investigations
Repeat of abnormal investigations found in first stage
Consider correction studies with normal plasma
Factor assays – II, V, VII, VIII, IX, X
Factor XIII assay (if not included in first-stage investigations)
Investigation of thrombocytopenia if present
Platelet aggregation tests (or flow cytometry in infants)
Platelet nucleotide analysis
Alpha-2-antiplasmin level
Plasminogen activator inhibitor-1 activity
If all the investigations mentioned above are normal, it is probable that the patient does not have a bleeding disorder. Investigations for the extremely rare disorders alpha-2-antiplasmin deficiency and plasminogen activator inhibitor-1 (PAI-1), which are available in only a few laboratories, may be undertaken but the diagnostic yield will be very low. Alpha-2-antiplasmin deficiency is a severe disorder, similar in its phenotype to factor XIII deficiency; PAI-1 deficiency has a milder phenotype.19 20
Conclusions
An awareness of the conditions which can be mistaken for non-accidental bruising is essential in the assessment of suspected abuse. Difficulties may arise when other features typical of abuse do not accompany bruising. Non-accidental bruising and bruising due to haematological disorder may share common features and are not mutually exclusive. Some conditions such as severe haemophilia can be readily diagnosed; however, the most common bleeding disorder von Willebrand disease, as well as rarer conditions, are more difficult to diagnose with certainty. It is essential to discuss interpretation of abnormal results with a paediatric haematologist. When a mild bleeding disorder is diagnosed one then has to ask if the severity of bruising is more than one would expect in that particular disorder.
In our proposed two-stage investigations, we have advocated more tests than are usually performed in the assessment of abuse by many paediatricians. Despite the difficulties involved, we feel it is important to assess platelet function. Some of the disorders tested for are rare but in areas which have a high incidence of consanguinity they may be collectively as common as severe haemophilia. Several of these disorders have featured in reports of mistaken diagnoses of abuse and a formal legal challenge may demand their exclusion.
A full personal and family history is essential in the assessment of a bleeding disorder; however, young children frequently need assessment before significant haemostatic challenges have been encountered. Given the limitations of blood tests, it may not be possible to exclude a bleeding disorder with absolute certainty. However, the absence of a positive clinical and family history and normal results of investigations in both stage 1 and stage 2 make the presence of a bleeding disorder highly unlikely.
Acknowledgments
We are grateful to Gillian Purvis for secretarial assistance.
References
Footnotes
-
Competing interests None.
-
Provenance and peer review Not commissioned; externally peer reviewed.