Académique Documents
Professionnel Documents
Culture Documents
A R T I C LE I N FO A B S T R A C T
Keywords: Objective: Skeletal survey is a commonly used means of detecting fractures in infants, and is used as a screen in
Child abuse suspected cases of physical abuse. It is recognised that in live infants, a repeat survey some days after a suspected
Infant episode of injury will detect more fractures than one taken shortly after the suspected injury, indicating that the
Rib fracture latter lacks sensitivity. In infants who die soon after a suspected episode of physical abuse, the managing
Histopathology
clinicians do not have the option of a second survey; however there is the opportunity for the microscopic
examination of bones removed at autopsy. Increasingly Osteoarticular Pathology at the Manchester University
NHS Foundation Trust (MFT) is being sent samples of bones from infants suspected of inflicted injury for his-
tological examination, both from bones with fractures detected at autopsy or skeletal survey and from posterior
ribs and long bone metaphyses (sites of significance in assessing for abusive injury) when there is no evidence of
fracture on skeletal survey or autopsy. Here we report the results of an audit of the anonymised data from a series
of such cases, to establish the sensitivity of skeletal survey (SS) to detect fractures and to define the medico-legal
value of submitting bones for histological examination.
Methods: This was an audit of skeletal injuries in 38 infants aged < 18 months presenting to MFT for specialist
histopathological evaluation of suspected non-accidental fractures between January 2011 and June 2017.
Histopathological examination was performed on all bones submitted and compared with contact radiography of
isolated bones and post-mortem skeletal surveys undertaken by specialist paediatric or musculoskeletal radi-
ologists for the presence of fracture.
Results: A total of 318 fractures were detected histologically; of these, 178 (56%) were of the ribs, 119 (37.5%)
were of major limb long bones, 10 (3%) were of the skull, and 11 (3.5%) were recorded as ‘other’. Excluding
refractures, skeletal survey detected 54% of the fractures recorded histologically. No fractures were detected
radiologically that were not seen histologically. Generally, for skeletal survey, detection rates improved with the
age of the lesion, and rib fractures were more difficult to detect than long bone fractures. Ribs 5–8 were the most
frequently fractured ribs, and metaphyses around the knee accounted for most metaphyseal limb long bone
fractures undetected by SS.
Conclusion: In infants coming to post-mortem, histopathology is more sensitive than SS for the detection of
clinically significant fractures. In children suspected of non-accidental injuries but with negative or equivocal SS,
sampling of the anterior and posterior end of ribs 5–8 and the bones around the knee for histological ex-
amination could reveal clinically unsuspected fractures and significant evidence of physical abuse. 71% of in-
fants showed evidence of old fractures typical of non-accidental injury.
1. Introduction second most common presentation of physical abuse, with classic me-
taphyseal lesion (CML) and posterior rib fractures being particularly
Since Caffey first established a relationship between unexplained specific.2–16 The consequence of overlooking such injuries could prove
subdural haematoma and long bone fractures in 1946,1 studies ex- fatal4,5; fractures indicative of inflicted injury that antedate death by
amining links between skeletal injury and physical abuse in infants days or weeks are, in our experience, often first recognised only after
have grown in number. Most of these are radiological. Fractures are the death.
∗
Corresponding author.
E-mail addresses: emma.raynor@student.manchester.ac.uk (E. Raynor), Praveen.Konala@mft.nhs.uk (P. Konala),
tony.freemont@manchester.ac.uk (A. Freemont).
https://doi.org/10.1016/j.jflm.2018.09.002
Received 12 February 2018; Received in revised form 5 July 2018; Accepted 3 September 2018
Available online 05 September 2018
1752-928X/ © 2018 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/BY-NC-ND/4.0/).
E. Raynor et al. Journal of Forensic and Legal Medicine 60 (2018) 9–14
Imaging techniques are the mainstay of fracture recognition in live Histology is regarded as the most accurate means of demonstrating
and dead infants. The skeletal survey (SS) is the first-line investigation and ageing fractures,3,4,14,33 especially those that are radiographically
for suspected physical abuse,5,8 the Royal College of Radiologists re- subtle.3,14,33 For instance, it is reported that: histology confirmed 6 CML
commends 22 different views17 to adequately visualise the entire ske- identified as ‘abnormal’ on SS3; four from five rib fractures ‘suspected’,
leton. For instance, SS identified 1029 fractures in 313 of 567 infants but not diagnosable, on SS33 and metaphyseal fractures in two speci-
aged < 12 months.9 789 fractures (77%) were clinically unsuspected, mens that appeared normal radiographically.3 However selecting tissue
thus highlighting the role of skeletal survey in detecting occult inflicted for histological examination is key, as it is not practical to sample the
fractures, particularly ribs,18 that would otherwise go undetected. whole skeleton histologically. For this the histopathologist relies on the
In addition to identifying fractures, SSs can provide information on radiologist and autopsy pathologist to guide sampling.
the age of skeletal injuries and the presence of any underlying bone There are no large studies that have systematically described the
disease17 which may prove particularly useful where an explanation different fracture detection rates of post-mortem SS, CR and histo-
provided by the care-giver does not match the type or extent of injuries pathology and related them to histological age of fracture, bone, and
sustained.19 region of bone; and yet the evidence above suggests the right balance of
CT scanning7 has proven useful in confirming skeletal injuries these techniques could be extremely important in optimising the de-
otherwise subtle on skeletal survey in both living and dead children. tection of abusive injuries. This is the rationale for our study.
High-detail CT scans can identify CMLs,14 and create 3D reconstruc-
tions that effectively demonstrate the extent of an injury to a lay jury.7
One drawback of CT scanning is the radiation dose20–22; potentially 2. Methods
outweighing its value as a screening test in live infants.
MRI scanning has also been used to supplement SS. When used to This was a retrospective audit of anonymised data from all cases of
assess non-cerebral injuries (including CMLs and rib fractures), whole- suspected non-accidental injuries referred to the histology department
body MRI has a ‘high specificity but low sensitivity’ when compared to of the Manchester Royal Infirmary for specialist histopathological
SS.23 For instance when whole-body MRI was compared to SS in 16 evaluation of skeletal injuries between 2011 and June 2017 (226 cases).
infants, 5.4% of CMLs were identified by MRI vs 64.8% by SS alone.23 Inclusion criteria were:
MRI identified only 54% of rib fractures.
In live infants bone scintigraphy has a high sensitivity for radio- • Child < 18 months old at the time of death.
graphically subtle lesions, such as rib fractures, with 50% in one study • Recorded data from a post-mortem skeletal survey reported by a
being identified by bone scintigraphy alone.24,25 One drawback is the specialist paediatric or musculoskeletal radiologist.
low sensitivity of detection of CMLs because of high uptake of scinti- • Archived contact radiographs of all the bones. Only cases with CR in
graphic medium by normal growth plates.24 2 planes for all bones examined were included.
There is broad agreement that outside the skull, certain skeletal
injuries carry a high likelihood of being abusive. The most significant in A total of 38 infants met these criteria. The median age was 12
this respect are metaphyseal and posterior rib fractures26 particularly in weeks (range: 19 days–14 months).
non-ambulatory infants.3–5,7,10,12 Grasping by the chest, squeezing, Fractures detected for each of the 3 tests, SS, CR and histology, were
violent shaking and pulling and twisting of the limbs are implicated in tabulated and compared. The data for skeletal surveys was as reported
these injuries.5,7,8 Caution is required in neonates as obstetric trauma, by the specialist radiologists. Contact radiographs are used by specialist
including traumatic vaginal births and uncomplicated Caesarean sec- histopathologists to aid sampling for histological analysis. For this study
tions,27–29 has been causally implicated in these types of fracture. however, in addition to being reviewed by AF and ER, all CR radio-
CMLs, particularly when acute, can be difficult to detect with con- graphs in which: no fracture could be seen; the appearances were felt to
ventional radiography.5,7 One study reported 15 healing metaphyseal be equivocal; or where a fracture was seen histologically but not
lesions visible on follow-up SS as ‘indeterminate’ initially.4 radiographically by the pathologists, were reviewed by a specialist
Whilst posterior rib fractures5–7,9,11 are highly specific for non- consultant musculoskeletal radiologist (PK). All fractures identified on
accidental injury,2,4–6,9–12 carrying a 95% positive predictive value for CR by AF, ER or PK were tabulated as positive. Histological recognition
physical abuse in children aged < 3 years,18 because of the anatomy, of fractures was made by AF, an osteoarticular pathologist specialising
with bone and soft tissue overlay, and a frequent lack of displacement, in the histological recognition and interpretation of fracture healing, in
they may be overlooked by conventional radiography, particularly if the biopsy diagnosis of paediatric metabolic bone disease, and 27 years
acute.5,6,9,17,30 In life, follow-up SS increases detection rates. In one medicolegal experience in recognising and aging fractures in infants.
study,30 94% of fractures first diagnosed at follow-up were either me- Specimens of bone were received from Home Office and forensic
taphyseal or rib fractures, and in another,9 64% of 98 new definite pathologists from across the UK as samples of bone in formalin. A short
fractures identified at follow-up in 41 infants were rib fractures. clinical history and the results of SS were included in all cases used in
In fatally abused infants follow-up SS is not possible6; necessitating this paper. Contact radiography was performed on a Faxitron [Faxitron
a different approach for detecting skeletal injuries after death. Bioptics, LLC 3440 E Britannia Drive, Suite 150, Tucson, Arizona USA
Skeletal survey, palpation at autopsy, and visualised peri-osseous 85706]. Images were taken in 2 planes. Following this, bones were
bruising alert pathologists to the possibility of fractures. In addition, decalcified and then widely sampled to include known and suspected
forensic and paediatric pathologists in the UK are increasingly taking fracture sites from SS, autopsy findings, CR, and sites at which identi-
key bones (e.g. posterior ribs) that might have occult fractures for fying a fracture might be of significance in assessing abusive injuries
histological examination by an experienced Osteoarticular Pathologist. (mainly metaphyses and posterior elements of the ribs). The tissue was
Prior to histological analysis it is common practice for the laboratory taken into paraffin using conventional methods and then sectioned at
dealing with the bone samples to undertake ‘contact’ (specimen) x-rays 4 μm. Serial sections were stained with haematoxylin and eosin (H&E)
using a high dose, high resolution system. These are rarely, if ever, and Martius scarlet blue (MSB). Two sections were examined from most
reported by radiologists, but guide experienced histopathologists in tissue blocks but equivocal microscopic findings always resulted in
sample selection. “deeper” sections being taken.
In one study comparing contact radiography (CR) with SS, fracture Fractures were aged histologically in a manner similar to that de-
detection increased from 58% to 92%, the additional skeletal injuries scribed by Klotzbach et al.33 supplemented with our own data derived
including high-specificity type injuries, of metaphyses and posterior from previously published studies examining the in-situ molecular
ribs.4 Evidence indicates it may also improve fracture aging.4,7,31,32 biology of fractures.34–36 Briefly:
10
E. Raynor et al. Journal of Forensic and Legal Medicine 60 (2018) 9–14
Table 1
Limb Long Bone Fracture Distribution. The detailed distribution of long bone Table 3
fractures by bone and by site within the bone is given in Table 2. Detection of metaphyseal lesions in limb long bones by age.
Number of fractures Number of infants in which fracture was Metaphyseal fractures aged Metaphyseal fractures aged
detected ≤14 days (N = 46) > 14 days (N = 10)
11
E. Raynor et al. Journal of Forensic and Legal Medicine 60 (2018) 9–14
12
E. Raynor et al. Journal of Forensic and Legal Medicine 60 (2018) 9–14
our study 27 of the 38 infants (71%) had at least one fracture > 14 days Declarations of interest
old. This emphasises the proportion of cases of infants suspected as
having died an abusive death who have evidence of previous, probably Miss Emma Raynor: None
inflicted, skeletal injuries. Praveen Konala: None
We have included contact/specimen radiography here because of a Professor Anthony Freemont: None
previously reported study comparing SS, CR and histology in 3 in- Professor Freemont is supported by the MRC and EPSRC (MRC
fants.33 In that study CR was the referent and it noted that all fractures Grant MR/N00583X/1) for work on molecular biomarkers and is part of
detected by CR were also detected by histology. If histopathology was the Musculoskeletal Division of the Manchester NIHR BRC.
superior to CR in detecting fractures the study design would not have
allowed this to become apparent. Equally the study design would have Acknowledgements
meant that if there were any discrepant findings between CR and SS, SS
would inevitably have been found to be inferior to CR. Our study shows This research did not receive any specific grant from funding
that SS and CR detect similar numbers of fractures but that the 2 agencies in the public, commercial, or not-for-profit sectors.
techniques are complimentary, together recognising a greater number
of fractures that either technique alone. However, even in combination, References
they are less efficient at identifying fractures than histology. An argu-
ment could be made that the CR data appear less accurate than his- 1. Caffey J. The classic: multiple fractures in the long bones of infants suffering from
tology and different from SS because the contact radiographs are re- chronic subdural hematoma. Clin Orthop Relat Res. 2011;469(3):755–758.
2. Offiah A, van Rijn RR, Perez-Rossello JM, Kleinman PK. Skeletal imaging of child
ported by histopathologists lacking skill in interpreting radiographs. We abuse (non-accidental injury). Pediatr Radiol. 2009;39(5):461–470.
have tried to control for that by having a specialist musculoskeletal 3. Kleinman PK, Marks SC, Blackbourne B. The metaphyseal lesion in abused infants: a
radiologist review our findings. The relative influence of the detection radiologic-histopathologic study. AJR Am J Roentgenol. 1986;146(5):895–905.
4. Kleinman PK, Marks Jr SC, Richmond JM, Blackbourne BD. Inflicted skeletal injury: a
of fractures by the use of CR within the context of this paper which is postmortem radiologic-histopathologic study in 31 infants. AJR Am J Roentgenol.
designed to compare SS and histopathology is probably small as in our 1995;165(3):647–650.
experience CR is not used to directly detect previously unrecognised 5. Kleinman PK. Diagnostic imaging in infant abuse. AJR Am J Roentgenol.
1990;155(4):703–712.
fractures so much as to guide the histopathologist to areas of bone to 6. Kleinman PK, Marks SC, Adams VI, Blackbourne BD. Factors affecting visualization of
sample histologically. posterior rib fractures in abused infants. AJR Am J Roentgenol. 1988;150(3):635–638.
Although most of the cases coming to our laboratory have SS as the 7. Dwek JR. The radiographic approach to child abuse. Clin Orthop Relat Res.
2011;469(3):776–789.
only imaging study of non-cranial bones, a small, but increasing min-
8. Paddock M, Sprigg A, Offiah AC. Imaging and reporting considerations for suspected
ority of cases, have undergone post mortem CT (PMCT). The potential physical abuse (non-accidental injury) in infants and young children. Part 1: initial
of this technique is discussed in Arthurs et al.38 In one study,39 com- considerations and appendicular skeleton. Clin Radiol. 2017;72(3):179–188.
paring thoracic radiography, thoracic PMCT, and post mortem ex- 9. Barber I, Perez-Rossello JM, Wilson CR, Kleinman PK. The yield of high-detail
radiographic skeletal surveys in suspected infant abuse. Pediatr Radiol.
amination (including histology) fractures were identified by a specialist 2014;45(1):69–80.
radiologist radiographically in 7/12 patients who had fractures found at 10. Barber I, Kleinman PK. Imaging of skeletal injuries associated with abusive head
autopsy, and on CT in 11/12 patients. In total, 46% of fractures were trauma. Pediatr Radiol. 2014;44(Suppl4):613–620.
11. van Rijn RR, Bilo RA, Robben SG. Birth-related mid-posterior rib fractures in neo-
reported on radiography, and 85% of fractures were reported on PMCT nates: a report of three cases (and a possible fourth case) and a review of the lit-
(a senior staff radiologist recorded 29% of fractures on the same erature. Pediatr Radiol. 2009;39(1):30–34.
radiographs and 51% on CT). Whilst supporting our contention that 12. Lonergan GJ, Baker AM, Morey MK, Boos SC. From the archives of the AFIP. Child
abuse: radiologic-pathologic correlation. Radiographics. 2003;23(4):811–845.
histology is essential in detecting fractures, this also emphasises the 13. Borg K, Hodes D. Guidelines for skeletal survey in young children with fractures.
potential of PMCT. Arch Dis Child Educ Pract Ed. 2015;100(5):253–256.
Increasingly, bones received in our laboratory have been in- 14. Tsai A, McDonald AG, Rosenberg AE, Gupta R, Kleinman PK. High-resolution CT with
histopathological correlates of the classic metaphyseal lesion of infant abuse. Pediatr
vestigated by PM micro-CT. This is a largely research tool, but one that Radiol. 2014;44(2):124–140.
has the potential to demonstrate the microstructure of fracture callus at 15. Maguire S. Which injuries may indicate child abuse? Arch Dis Child Educ Pract Ed.
a resolution similar to that of low power microscopy.40 It requires a 2010;95(6):170–177.
16. Leventhal JM, Martin KD, Asnes AG. Incidence of fractures attributable to abuse in
similar approach to sampling of bone tissue as histology. A careful
young hospitalized children: results from analysis of a United States database.
comparative study of PM micro-CT and histology is awaited. Pediatrics. 2008;122(3):599–604.
17. Skeletal Surveys for Non-accidental Injury. The Royal College of Radiologists;
2017http://www.rcr.ac.uk/audit/skeletal-surveys-non-accidental-injury.
18. Barsness KA, Cha ES, Bensard DD, et al. The positive predictive value of rib fractures
5. Conclusion as an indicator of nonaccidental trauma in children. J Trauma.
2003;54(6):1107–1110.
This study has shown that post-mortem bone histology is superior to 19. Lloyd-Roberts G. The diagnosis of injury of bones and joints in young babies. Proc
Roy Soc Med. 1968;61(12):1299–1300.
post mortem skeletal survey and contact radiography for detecting 20. Bajaj M, Offiah AC. Imaging in suspected child abuse: necessity or radiation hazard?
fractures in infants, particularly re-fractures, which have forensic im- Arch Dis Child. 2015;100(12):1163–1168.
plications, and acute fractures of ribs and metaphyses. However, our 21. Pearce MS, Salotti JA, Little MP, et al. Radiation exposure from CT scans in childhood
and subsequent risk of leukaemia and brain tumours: a retrospective cohort study.
experience is that skeletal survey (together with macroscopic autopsy Lancet. 2012;380(9840):499–505.
findings) is essential in directing histopathologists to sites of potential 22. Johnson JN, Hornik CP, Li JS, et al. Cumulative radiation exposure and cancer risk
fractures. Without these findings to guide the histopathologist, histo- estimation in children with heart disease. Circulation. 2014;130(2):161–167.
23. Perez-Rossello JM, Connolly SA, Newton AW, Zou KH, Kleinman PK. Whole-body
logical assessment would be almost impossible. We have shown, how-
MRI in suspected infant abuse. AJR Am J Roentgenol. 2010;195(3):744–750.
ever, that there are specific sites in the body (namely ribs 5, 6, 7 and 8 24. Mandelstam SA, Cook D, Fitzgerald M, Ditchfield MR. Complementary use of radi-
and metaphyses around the knee) that should be routinely sampled and ological skeletal survey and bone scintigraphy in detection of bony injuries in sus-
pected child abuse. Arch Dis Child. 2003;88(5):387–390.
submitted for histological assessment even if no fracture is suspected on
25. Kemp AM, Butler A, Morris S, et al. Which radiological investigations should be
SS or at autopsy as these often yield vital data that would otherwise be performed to identify fractures in suspected child abuse? Clin Radiol.
missed. 2006;61(9):723–736.
A significant majority of the 38 infants studied here had evidence of 26. Child abuse. BMJ Best Practice. 2017; 2017http://bestpractice.bmj.com/topics/en-
us/846.
healing fractures not associated with the terminal event, many of which 27. Sieswerda-Hoogendoorn T, van Rijn RR, Robben SGF. Classic metaphyseal lesion
were however typical of non-accidental injury. following vaginal breech birth, a rare birth trauma. J Forensic Radiology Imaging.
2014;2(1):2–4.
13
E. Raynor et al. Journal of Forensic and Legal Medicine 60 (2018) 9–14
28. O'Connell A, Donoghue VB. Can classic metaphyseal lesions follow uncomplicated 35. Andrew JG, Hoyland JA, Freemont AJ, Marsh DR. Platelet-derived growth factor
caesarean section? Pediatr Radiol. 2007;37(5):488–491. expression in normally healing human fractures. Bone. 1995;16(4):455–460.
29. Lysack JT, Soboleski D. Classic metaphyseal lesion following external cephalic ver- 36. Andrew JG, Andrew SM, Freemont AJ, Marsh DR. Inflammatory cells in normal
sion and cesarean section. Pediatr Radiol. 2003;33(6):422–424. human fracture healing. Acta Orthop Scand. 1994;65(4):462–466.
30. Kleinman PK, Nimkin K, Spevak MR, et al. Follow-up skeletal surveys in suspected 37. Thompson NB, Kelly DM, Warner Jr WC, et al. Intraobserver and interobserver re-
child abuse. AJR Am J Roentgenol. 1996;167(4):893–896. liability and the role of fracture morphology in classifying femoral shaft fractures in
31. Geller E. Imaging in Child Abuse. 2017; 2017http://emedicine.medscape.com/article/ young children. J Pediatr Orthop. 2014;34(3):352–358.
407144- overview. 38. Arthurs OJ, Hutchinson JC, Sebire NJ. Current issues in postmortem imaging of
32. Worlock P, Stower M, Barbor P. Patterns of fractures in accidental and non-accidental perinatal and forensic childhood deaths. Forensic Sci Med Pathol. 2017;13(1):58–66.
injury in children: a comparative study. Br Med J. 1986;293(6539):100–102. 39. Hong TS, Reyes JA, Moineddin R, Chiasson DA, Berdon WE, Babyn PS. Value of
33. Klotzbach H, Delling G, Richter E, Sperhake JP, Püschel K. Post-mortem diagnosis postmortem thoracic CT over radiography in imaging of pediatric rib fractures.
and age estimation of infants' fractures. Int J Leg Med. 2003;117(2):82–89. Pediatr Radiol. 2011;41(6):736–748.
34. Kwong FN, Hoyland JA, Evans CH, Freemont AJ. Regional and cellular localisation of 40. Rutty GN, Brough A, Biggs MJ, Robinson C, Lawes SD, Hainsworth SV. The role of
BMPs and their inhibitors' expression in human fractures. Int Orthop. micro-computed tomography in forensic investigations. Forensic Sci Int.
2009;33(1):281–288. 2013;225(1–3):60–66.
14