The child with unexplained fractures
Colin R Paterson
Published in the New Law Journal
1997 volume 147 pages 648 to 652
1A recently reported judgement has drawn attention to the frequent difficulty attending
the diagnosis of the child with unexplained fractures. To many observers the failure of parents to come up with an explanation for fractures found radiologically is ipso facto
evidence for non-accidental injury; the lack of explanation must represent a failure to tell the truth about their own or their partners’ actions. However, unexplained fractures in childhood are also the hall-mark of all forms of brittle bone disease and immense harm can be done to families by the inaccurate diagnosis of non-accidental injury.
Much of our research over the last 25 years has related to the clinical aspects of the brittle bone diseases and we hold a database with details of over 1,300 patients. The best known of these is osteogenesis imperfecta which has a prevalence of about one in 10,000 in the United Kingdom. It is caused by abnormalities in collagen, the fibrous protein essential for the mechanical strength of bone. In turn, in most cases, this is now known to be caused by defects in the genes responsible for collagen formation. Since collagen is abnormal in tissues other than bone, patients with osteogenesis imperfecta may have detectable features in addition to fractures. These include blue or grey discolouration of the sclerae (whites of the eyes), discolouration and fragility of the teeth, laxity of joints and an increased tendency to bruising or pinpoint bruises known as petechiae. The bruising is thought to reflect abnormalities in the collagen of small blood vessels. X-rays may show obvious abnormalities but in a majority of patients the appearances are normal at the time of the first few fractures; many of the abnormalities seen later reflect the fractures and the immobilisation used in their treatment. In some cases osteogenesis imperfecta is passed down from a parent to a child, but many cases are ‘sporadic’ with no known family history.
In most patients with osteogenesis imperfecta the diagnosis is made without undue difficulty on the basis of the clinical signs, the fracture history or the family history. In a retrospective survey of 802 known cases of osteogenesis imperfecta in the United 2Kingdom we found that in 691 the diagnosis had been made confidently at birth or at the time of the first fracture. In 96 cases the parents were accused of non-accidental
injury on at least one occasion. In 15 cases they had had to contend with case conferences, care proceedings or criminal proceedings.
Over the last 12 years we have identified a distinctive pattern in a minority of patients initially thought to have osteogenesis imperfecta. In this variant, known as temporary 3brittle bone disease, the fractures are limited to the first year of life and, to a large extent, the first six months of life. The fracture pattern is often distinctive with rib fractures and fractures at the ends of long bones (metaphyseal fractures) being frequent. These patients may have other features such as vomiting (often projectile) and anaemia. While there is usually no family history of fractures, there is a family history of joint laxity in about two thirds of cases. The cause of the disorder is not yet known but it appears to be more common in twins and infants born before full term.
It is not surprising that both osteogenesis imperfecta and temporary brittle bone disease are often considered in cases in which a child is found to have unexplained fractures. This article summarises a personal experience of cases in which the author prepared a report on the causes of fractures and the likelihood of an underlying bone disease. Since these cases have been studied over some 21 years it has been possible to follow up most of the children concerned for substantial periods.
A database was prepared to include details of each child with information on the mode of referral, the diagnosis reached personally, the legal outcome and the details of the follow-up. Additional clinical information was recorded in each case. The current report is restricted to 128 patients living in the United Kingdom, in whom the major problem was the fractures.
Table 1 shows the source of the referrals. Table 2 shows the diagnosis made by the author in each case. While patients with temporary brittle bone disease were not recognised as such before 1985, it was clear in retrospect that some patients seen earlier had this disorder. Two infants with an initial diagnosis of temporary brittle bone disease were later re-classified as osteogenesis imperfecta in the light of subsequent fractures.
Source of medico-legal referrals 1974-96
Parents’ representatives 102
Guardians ad litem 8
Local authority 6
Senior hospital staff 8
General practitioners 3
Diagnosis of 128 patients referred for
the diagnosis of unexplained fractures
Osteogenesis imperfecta 33
Temporary brittle bone disease 65
Vitamin D deficiency rickets 5
Scurvy (vitamin C deficiency) 1
Accidental injury 9
Unresolved/non-accidental injury 14
* Hypophosphatasia is an uncommon
heritable disorder of bone
Of the 105 patients thought to have bone disorders the author provided evidence for care proceedings in 102. Of these infants the eventual outcome was that 78 were returned to their parents (56 initially with conditions), three went to other family members and 21 were removed permanently from their families. In three of these the parents had given up before formal proceedings. In seven families the parents separated; in three because one parent was blamed.
Of the 33 children thought to have osteogenesis imperfecta 25 were returned to their parents. One died later with bronchopneumonia and multiple unexplained gastrointestinal problems. The remaining patients have been followed up for between one and 18 years (total 136 patient-years, mean 5.6 years). There was no evidence of non-accidental injury in this period.
Of the 65 children thought to have temporary brittle bone disease, 48 were returned to their parents. Two died later; one with a cot death and one with late sequelae of birth injury; in neither was non-accidental injury postulated. In 43 of the remaining patients follow-up information was available for between 1 and 11 years (total 248 patient years, mean 5.8 years). There was no evidence of non-accidental injury during this period.
In the whole group of 105 children thought to have bone disease the evidence was rejected judicially in 29 cases and formally accepted in 23. In the remaining cases there was no formal finding for a variety of reasons, most commonly because rehabilitation of the child with the family was agreed without a hearing. Among the 65 patients thought to have temporary brittle bone disease this evidence was rejected in 18 and accepted in 11; in the remaining 36 patients there was no judicial finding. An analysis of the clinical findings in these three groups did not demonstrate any differences in relation to a wide range of clinical features.
Over the last 20 years there has been some reduction in the number of new cases referred in which the diagnosis was osteogenesis imperfecta. Increased familiarity with the clinical features of this disorder has led to more frequent early diagnosis. In the past some of the cases referred to the courts had classical features such as abnormal sclerae 2or teeth, or had a clearly positive family history which had not been sought.
However, retrospective study of confirmed cases of osteogenesis imperfecta continues to demonstrate that, in a minority of patients, the diagnosis was extremely difficult at the time of the earlier fractures. Since there may be long fracture-free periods in known cases it is possible to be misled by the lack of subsequent fractures. In one particularly unfortunate family, in which the author was not involved legally, a child was taken into care at the age of 18 months after two fractures. A subsequent fracture did not occur for a further 18 months and the diagnosis of osteogenesis imperfecta was only made at the age of five years when she was returned to her mother. Retrospective study of the medical records and x-rays in this case revealed little evidence that would have helped to make the correct diagnosis at the time.
While such difficult cases are uncommon they occur too frequently in the United Kingdom as a whole to allow for complacency. Our experience in the current series indicates that, where a diagnosis of osteogenesis imperfecta is made and the child is returned to the parents, no evidence of subsequent non-accidental injury has been observed in 136 patient-years of follow-up. In most cases subsequent fractures occurred but mainly at ages at which the child was able to give a clear account of the events.
In recent years it has become possible to identify abnormalities in collagen formation by cells grown in culture from excised samples of skin. With one approach it was claimed that such abnormalities could be demonstrated in over 80 per cent of cases of 4osteogenesis imperfecta. Such assays are time-consuming and labour-intensive; they
are not widely available. In the past some reports have relied on such methods even in cases in which there was already ample clinical evidence of osteogenesis imperfecta. It is important that the limitations of such tests are recognised.
5,6,7Temporary brittle bone disease is a much more controversial subject. Some of its
features as reported by us are those that have been conventionally regarded as typical of 8,9non-accidental injury for the last thirty years. However, the evidence that these
features, including rib fractures and metaphyseal fractures, are linked to non-accidental injury, is limited. In addition these fractures occur in a wide range of known bone disorders. For example, rib fractures occur spontaneously in known cases of ordinary osteogenesis imperfecta and may occur in utero. Metaphyseal fractures occur not only
in osteogenesis imperfecta but also in at least five other bone disorders in the first year of life.
There are four principal types of evidence that support the view that temporary brittle bone disease exists and does not represent misdiagnosed non-accidental injury. First the patients all show striking similarities in their clinical features, the types of fractures, the ages at which they occur, the other symptoms such as vomiting, the other signs such as enlarged fontanelles, and the family history observations. Were these infants not thought to have sustained non-accidental injury they would readily have been recognised as having a distinctive syndrome.
Second, as with ordinary osteogenesis imperfecta, there is often a striking discrepancy between the fractures and other evidence of injury. In typical non-accidental injury bruises greatly outnumber fractures. In this disorder there may be over twenty fractures but reliable evidence that no superficial sign of injury was present at the time when the fractures occurred. Third, the same syndrome occurs in infants in whom non-accidental injury can be excluded with confidence, generally because the fractures occurred while the child was in hospital.
Fourth, the evidence provided in this report emphasises that when these patients were returned to their parents no subsequent evidence of non-accidental injury has been identified in 248 patient-years of follow-up. The premise underlying care proceedings is that abusive parents remain abusive and that there is substantial risk of further non-accidental injury if an abused child is returned. The follow-up findings in this report support the view that, in this small distinctive group of infants with unexplained fractures, the diagnosis was not non-accidental injury.
References 1 Wall J (1995) Re AB (child abuse: expert evidence). 1 FLR 181. 2 Paterson CR, McAllion SJ (1989) Osteogenesis imperfecta in the differential diagnosis of child abuse. BMJ 299: 1451. 3 Paterson CR, Burns J, McAllion SJ (1993) Osteogenesis imperfecta: the distinction from child abuse and the recognition of a variant form. Amer J Med Genet 45: 187. 4 Steiner RD, Pepin M, Byers PH (1996) Studies of collagen synthesis and structure in the differentiation of child abuse from osteogenesis imperfecta. J Pediatr 128: 542. 5 Smith R, Wynne JM, Hobbs CJ, Carty H (1995) Osteogenesis imperfecta, non-accidental injury and temporary brittle bone disease. Arch Dis Childh 72 : 169. 6 Shaw DG, Hall CM, Carty H (1995) Osteogenesis imperfecta: the distinction from child abuse and the recognition of a variant form. Amer J Med Genet 56: 116. 7 Paterson CR, Burns J, McAllion SJ (1995) Osteogenesis imperfecta variant v child abuse: reply. Amer J Med Genet 56: 117. 8 Carty HML (1993) Fractures caused by child abuse. J Bone Joint Surg 75-B: 849. 9 Chapman S (1993) Recent advances in the radiology of child abuse. Baill Clin
Paediatr 1: 211.
Dr Colin R Paterson, Department of Medicine, Ninewells Hospital and Medical School, Dundee DD1 9SY, Scotland. firstname.lastname@example.org
I am indebted to Mrs EA Monk for preparing the databases used in this work, to Dr SJ McAllion and Ms J Hoyal for advice on this article in draft and to the Cunningham Trustees for their support for our work on osteogenesis imperfecta.