In a paper published in 1989 we reported children with osteogenesis imperfecta frequently were perceived as having been physically abused.21 At that time, we reviewed records of 802 patients from the United Kingdom with a confirmed diagnosis of osteogenesis imperfecta. Nonacci- dental injury never was alleged in 691 patients. In 111 patients parents were accused of nonaccidental injury on at least one occasion. In 15 of these cases parents had to contend with formal case conferences, care proceedings, and repeated interviews from social workers and the police. In one case, the parents were prosecuted but acquitted.
Other authors also have drawn attention to the potential for diagnostic difficulty. Gahagan and Rimsza12 reported three patients with biochemical evidence of collagen abnormalities who initially were thought to have been abused. Only one of these three patients had blue sclerae.12 Augarten et al4 described a patient in whom the diagnosis of nonaccidental injury was eliminated only after the child had clear evidence of dentinogenesis imperfecta develop. Ojima et al19 described a 3-year-old child provisionally diagnosed with child abuse after dying suddenly and having multiple untreated fractures. The typical features of osteogenesis imperfecta Type III were not recognized until later.19 Minnis et al18 described a child eventually recognized to have osteogenesis imperfecta Type III who was placed in foster care for 5 months because of an initial diagnosis of nonaccidental injury. However, other authors have suggested that there seldom are difficulties in differentiating osteogenesis imperfecta from a nonaccidental injury.1,2,7,28
We report 12 patients evaluated from 1990 to 2000 initially believed to have sustained nonaccidental injuries, but later diagnosed with osteogenesis imperfecta after more complete investigation. We will identify the clinical and radiographic factors that led to the initial incorrect diagnosis.
MATERIALS AND METHODS
We identified 12 patients involved in formal enquiries or proceedings. Of the 12, five had been referred for diagnostic evaluation by the parents' legal representatives, six by family physicians, and one by a consultant pediatrician. Eight patients were from the United Kingdom and four were from the United States.
We reviewed the medical records of each patient (Table 1). In addition, the parents and other available relatives were interviewed to obtain independent medical and family histories. Eight patients were examined by one of the authors (CRP). The families were contacted by telephone or in person to obtain the subsequent history. We examined each file to identify any comments on the likelihood of bone disease in the initial radio- graphic evaluations. This information was available for 10 patients.
Seven patients had unrecognized fractures at the time of the initial investigation. Three patients had additional fractures occur soon after the initial fracture while they were still in the hospital or in foster care. In three patients osteogenesis imperfecta was not considered when the initial fractures were being investigated. Blue or gray sclerae were observed in nine patients, and four patients had substantially abnormal anterior fontanels.
Five infants had more than 10 substantial wormian bones. However, few patients had Towne's view radio- graphs taken. Seven patients had no skull radiographs taken at the time of the initial evaluation. In one patient (Patient 12), skull radiographs taken when the patient was 2 months old showed no good evidence of wormian bones but subsequent radiographs at 8 months, when another fracture had occurred while the infant was in foster care, showed clear evidence of multiple wormian bones.
The bones in question were regarded as normal on presentation in all but two of the 10 patients for whom initial radiological reports were available (Table 2). The 12 patients had 49 fractures before the investigation; 19 were rib fractures, three were clavicle fractures, and 27 were limb fractures (Table 3).
Four patients had been evaluated using skin biopsies and fibroblast cultures. Three patients had decreased Type I collagen.15 One of the patients had a mutation detected by polymerase chain reaction and conformation-sensitive gel electrophoresis sequencing. However, the abnormal findings did not prevent legal proceedings. In one patient (Patient 3) with negative collagen findings, there is now no doubt regarding the accuracy of the diagnosis of osteogenesis imperfecta.
As a result of formal investigations, including legal care proceedings in nine patients, 10 of the children were returned to their parents or remained at home. Two children were removed from the parents but both eventually were returned. These patients were followed for a minimum of 3 years (mean, 4.8 years; range, 3-13 years). Although seven patients have had additional fractures, there have been no additional allegations of nonaccidental injury.
The case that concerned us most in terms of failure of the physicians to respond to diagnostic information was that of a twin boy, Patient 1, who had his first symptomatic fracture, an oblique fracture of the shaft of the right femur, at 1 month of age. However, he also had an earlier clavicle fracture. The appearance of the callus strongly suggested that it had occurred at birth. The child also had asymptomatic metaphyseal fractures. He continued to experience fractures while in the hospital and in foster care. He had distinctly blue/gray sclerae and a large anterior fontanel. His mother also had blue sclerae and a history of fractures during childhood. Abnormalities were found with collagen chemistry after a skin biopsy and in the DNA evaluation. Despite these findings, it was asserted by the pediatricians concerned that the fractures were caused by child abuse. The child and his brother were placed in foster care and the parents were denied access for 3 months. During the first 4 months of life, the boy had approximately 128 radio- graphs, most of which could not be justified. After formal court proceedings, he was returned to his parents. Since then he has thrived but still sustains occasional fractures.
In another particularly troublesome case (Patient 7), healing rib fractures were seen in a 5-month-old child being investigated for failure to thrive. Osteogenesis imperfecta was not considered at that point. However, when referred to the senior author (CRP), the patient had distinctly abnormal sclerae, a very large anterior fontanel, substantial numbers of wormian bones, and a head shape typical for osteogenesis imperfecta. However, the alternative diagnosis of child abuse was accepted by a court, and at 12 months of age, the child was placed in foster care with limited access by his parents. Failure to thrive continued, and at 16 months he sustained a femoral fracture while in foster care. The diagnosis then was reconsidered and he was returned to his parents 1 month later. During the next 4 years he had more than 30 fractures and his growth remained poor. There is little doubt he has osteo- genesis imperfecta Type III. He has continuing behavioral problems ascribed to the period of foster care during his early life.
Osteogenesis imperfecta has been recognized for many years, and in Western countries occurs in one in 10,000 to 30,000 people.13 Most cases of osteogenesis imperfecta are caused by defects in the genes coding for Type I collagen.8 Many distinct mutations have been recognized.6,8,25 The condition is very heterogeneous at a molecular level, as reflected by the variety of clinical manifestations.
A diagnosis of nonaccidental injury should be considered for children with unexplained fractures; however, unexplained fractures are also characteristic of a range of bone disorders including osteogenesis imperfecta. Our patients initially were diagnosed with nonaccidental injuries because the parents or caretakers were unable to provide an explanation for the fractures, or because the tentative explanation was regarded as inadequate. Although unexplained fractures at different stages of healing may reflect a nonaccidental injury, they are also typical findings in osteogenesis imperfecta and similar diseases (Fig 1).
It often is suggested osteogenesis imperfecta is a form of osteoporosis, and that a reduction in the apparent bone density is important for making the diagnosis.14,30 However, we think this view incorrect. Normal bone appearances are the rule rather than the exception in many patients with the milder types of osteogenesis imperfecta at the time of the first few fractures.22,26 Bone densitometry may show reduced bone density in some children with osteogenesis imperfecta,17 but many with milder types (I and IV) have results within the reference ranges, particularly when allowance is made for the size of the patients.14 All forms of densitometry are limited by the inadequacy of reference data in childhood. When bone densitometry is performed in adults, a substantial proportion of patients still have results within the reference range.11,23 The assessment of bone density from ordinary radiographs is distinctly unreliable even if osteopenia is present. Osteopenia cannot be recognized with confidence using ordinary radiographs.29 In our cases the radiologists often failed to indicate that apparently normal bone appearances did not exclude osteogenesis imperfecta (Table 2).
Although a positive family history is helpful in the diagnosis of osteogenesis imperfecta, new mutations are well recognized.5 A negative family history does not exclude osteogenesis imperfecta. We were surprised to find that six of our patients had a family history of osteogenesis imperfecta which had not been elicited.
Fractures in patients with osteogenesis imperfecta can be any type and are not confined to the diaphysis (Table 3).10,26 Metaphyseal fractures can result from nonaccidental injuries, but also occur in a wide range of bone disorders including osteogenesis imperfecta in which the other radiographic features may confirm the diagnosis.3,10,20 In one of our patients (Patient 1), the presence of metaphyseal fractures contributed to the delay of the correct diagnosis of osteogenesis imperfecta.
In four of our patients, the finding of excessive numbers of wormian bones contributed substantially to the diagnosis of osteogenesis imperfecta. However, an excess of wormian bones is not essential to the diagnosis. In our experience with osteogenesis imperfecta Types I and IV, an excess of wormian bones occurs only in a minority of patients. Wormian bones occur in excess in most patients with osteogenesis imperfecta Type III,9 but in these patients, the diagnosis should not ordinarily present difficulties. A Towne's view radiograph should be used to maximize the chance of identifying wormian bones.9 Towne's view radiographs were performed for only three of our patients.
An often neglected measurement is the size of the anterior fontanel. As with wormian bones, abnormal fontanel size may point to osteogenesis imperfecta,27 but a normal fontanel does not exclude the diagnosis. It is unfortunate that reference data are available only for patients up to the age of 12 months.24 Individual fractures may occur without clinical evidence of trauma.16 However, the larger the number of fractures with this apparent discrepancy, the more unlikely it is that the fractures represent inflicted trauma to normal bone.
We attach considerable importance to the outcome of the cases, some of which were heavily contested. Information on the subsequent history was available for all patients. None had nonaccidental injuries. Some patients had no additional fractures but had continuing good evidence of osteogenesis imperfecta.
In some of these patients, it was difficult to diagnose osteogenesis imperfecta at the time of the initial fractures. To avoid the misdiagnosis of nonaccidental injury, it is important to do a careful and unprejudiced review of the patient's clinical history, family history, physical examination findings, and radiographic findings. A mistaken diagnosis of nonaccidental injury causes substantial harm to the family and particularly to the child.
1. Ablin DS, Greenspan A, Reinhart M, Grix A. Differentiation of child abuse from osteogenesis imperfecta. AJR Am J Roentgenol
2. Ablin DS, Sane SM. Non-accidental injury: confusion with temporary brittle bone disease and mild osteogenesis imperfecta. Pediatr Radiol
3. Astley R. Metaphyseal fractures in osteogenesis imperfecta. Br J Radiol
4. Augarten A, Laufer J, Szeinberg A, Paswell J. Child abuse, osteo- genesis imperfecta and the grey zone between them. J Med
. 1993; 24:171-175.
5. Blumsohn A, McAllion SJ, Paterson CR. Excess paternal age in apparently sporadic osteogenesis imperfecta. Am J Med Genet
6. Byers PH, Wallis GA, Willing MC. Osteogenesis imperfecta: translation of mutation to phenotype. J Med Genet
7. Chapman S, Hall CM. Non-accidental injury or brittle bones. Pediatr Radiol
8. Cole WG. Advances in osteogenesis imperfecta. Clin Orthop Relat Res
9. Cremin B, Goodman H, Spranger J, Beighton P. Wormian bones in osteogenesis imperfecta and other disorders. Skeletal Radiol
. 1982; 8:35-38.
10. Dent JA, Paterson CR. Fractures in early childhood: osteogenesis imperfecta or child abuse? J Pediatr Orthop
11. Deodhar AA, Woolf AD. Bone density measurement in osteogenesis imperfecta may well be important. Postgrad Med J
. 1994;70: 463-464.
12. Gahagan S, Rimsza ME. Child abuse or osteogenesis imperfecta: how can we tell? Pediatrics
13. Kuurila K, Grenman R, Johansson R, Kaitila I. Hearing loss in children with osteogenesis imperfecta. EurJ Paediatr
. 2000;159: 515-519.
14. Lund AM, Mølgaard C, Müller J, Skovby F. Bone mineral content and collagen defects in osteogenesis imperfecta. Acta Paediatr
15. Marlowe A, Pepin MG, Byers PH. Testing for osteogenesis imperfecta in cases of suspected non-accidental injury. J Med Genet
16. Mathew MO, Ramamohan N, Bennet GC. Importance of bruising associated with paediatric fractures: prospective observational study. BMJ
17. Miller ME, Hangartner TN. Bone density measurements by computed tomography in osteogenesis imperfecta type I. Osteoporos Int
18. Minnis H, Ramsay R, Ewije P, Kumar C. Osteogenesis imperfecta and non-accidental injury. Br J Psychiatry
19. Ojima K, Matsumoto H, Hayase T, Fukui Y. An autopsy case of osteogenesis imperfecta initially suspected as child abuse. Forensic Sci Int
20. Paterson CR, Burns J, McAllion SJ. Osteogenesis imperfecta variant vs. child abuse. Am J Med Genet
21. Paterson CR, McAllion SJ. Osteogenesis imperfecta in the differential diagnosis of child abuse: reply (reply to Letter to the Editor). BMJ
22. Paterson CR, McAllion SJ, Shaw JW. Clinical and radiological features of osteogenesis imperfecta type IVA. Acta Paediatr Scand
23. Paterson CR, Mole PA. Bone density in osteogenesis imperfecta may well be normal. Postgrad Med J
24. Popich GA, Smith DW. Fontanels: range of normal size. J Pediatr
25. Prockop DJ, Kuivaniemi H, Tromp G. Molecular basis of osteogenesis imperfecta and related disorders of bone. Clin Plast Surg
. 1994; 21:407-413.
26. Smith R. Osteogenesis imperfecta, non-accidental injury, and temporary brittle bone disease. Arch Dis Child
27. Smith R, Francis MJO, Houghton GR. The Brittle Bone Syndrome: Osteogenesis Imperfecta
. London, England: Butterworths; 1983.
28. Taitz LS. Child abuse and metabolic bone disease: are they often confused? BMJ
29. Williamson MR, Boyd CM, Williamson SL. Osteoporosis: diagnosis by plain chest films versus dual photon bone densitometry. Skeletal Radiol
30. Zionts LE, Nash JP, Rude R, Ross T, Stott NS. Bone mineral density in children with mild osteogenesis imperfecta. J Bone Joint Surg Br