Journal of Forensic and Legal Medicine 60 (2018) 9–14

Contents lists available at ScienceDirect

Journal of Forensic and Legal Medicine

journal homepage: www.elsevier.com/locate/yjflm

The detection of significant fractures in suspected infant abuse T

a,∗ b c
Emma Raynor , Praveen Konala , Anthony Freemont
a
University of Manchester, School of Medical Sciences, Stopford Building, Oxford Road, Manchester, M13 9PT, United Kingdom
b
Manchester Royal Infirmary, Oxford Road, Manchester, M13 9WL, United Kingdom
c
Manchester Molecular Pathology Innovation Centre, 3rd floor Citylabs, Nelson Street, Manchester, M13 9NQ, United Kingdom

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
Since Caffey first established a relationship between unexplained
subdural haematoma and long bone fractures in 1946,1 studies ex-
amining links between skeletal injury and physical abuse in infants
have grown in number. Most of these are radiological. Fractures are the
second most common presentation of physical abuse, with classic me-
taphyseal lesion (CML) and posterior rib fractures being particularly
specific.2–16 The consequence of overlooking such injuries could prove
fatal4,5; fractures indicative of inflicted injury that antedate death by
days or weeks are, in our experience, often first recognised only after
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.
Imaging techniques are the mainstay of fracture recognition in live
and dead infants. The skeletal survey (SS) is the first-line investigation
for suspected physical abuse,5,8 the Royal College of Radiologists re-
commends 22 different views17 to adequately visualise the entire ske-
leton. For instance, SS identified 1029 fractures in 313 of 567 infants
aged < 12 months.9 789 fractures (77%) were clinically unsuspected,
thus highlighting the role of skeletal survey in detecting occult inflicted
fractures, particularly ribs,18 that would otherwise go undetected.
In addition to identifying fractures, SSs can provide information on
the age of skeletal injuries and the presence of any underlying bone
disease17 which may prove particularly useful where an explanation
provided by the care-giver does not match the type or extent of injuries
sustained.19
CT scanning7 has proven useful in confirming skeletal injuries
otherwise subtle on skeletal survey in both living and dead children.
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
outweighing its value as a screening test in live infants.
MRI scanning has also been used to supplement SS. When used to
assess non-cerebral injuries (including CMLs and rib fractures), whole-
body MRI has a ‘high specificity but low sensitivity’ when compared to
SS.23 For instance when whole-body MRI was compared to SS in 16
infants, 5.4% of CMLs were identified by MRI vs 64.8% by SS alone.23
MRI identified only 54% of rib fractures.
In live infants bone scintigraphy has a high sensitivity for radio-
graphically subtle lesions, such as rib fractures, with 50% in one study
being identified by bone scintigraphy alone.24,25 One drawback is the
low sensitivity of detection of CMLs because of high uptake of scinti-
graphic medium by normal growth plates.24
There is broad agreement that outside the skull, certain skeletal
injuries carry a high likelihood of being abusive. The most significant in
this respect are metaphyseal and posterior rib fractures26 particularly in
non-ambulatory infants.3–5,7,10,12 Grasping by the chest, squeezing,
violent shaking and pulling and twisting of the limbs are implicated in
these injuries.5,7,8 Caution is required in neonates as obstetric trauma,
including traumatic vaginal births and uncomplicated Caesarean sec-
tions,27–29 has been causally implicated in these types of fracture.
CMLs, particularly when acute, can be difficult to detect with con-
ventional radiography.5,7 One study reported 15 healing metaphyseal
lesions visible on follow-up SS as ‘indeterminate’ initially.4
Whilst posterior rib fractures5–7,9,11 are highly specific for non-
accidental injury,2,4–6,9–12 carrying a 95% positive predictive value for
physical abuse in children aged < 3 years,18 because of the anatomy,
with bone and soft tissue overlay, and a frequent lack of displacement,
they may be overlooked by conventional radiography, particularly if
acute.5,6,9,17,30 In life, follow-up SS increases detection rates. In one
study,30 94% of fractures first diagnosed at follow-up were either me-
taphyseal or rib fractures, and in another,9 64% of 98 new definite
fractures identified at follow-up in 41 infants were rib fractures.
In fatally abused infants follow-up SS is not possible6; necessitating
a different approach for detecting skeletal injuries after death.
Skeletal survey, palpation at autopsy, and visualised peri-osseous
bruising alert pathologists to the possibility of fractures. In addition,
forensic and paediatric pathologists in the UK are increasingly taking
key bones (e.g. posterior ribs) that might have occult fractures for
histological examination by an experienced Osteoarticular Pathologist.
Prior to histological analysis it is common practice for the laboratory
dealing with the bone samples to undertake ‘contact’ (specimen) x-rays
using a high dose, high resolution system. These are rarely, if ever,
reported by radiologists, but guide experienced histopathologists in
sample selection.
In one study comparing contact radiography (CR) with SS, fracture
detection increased from 58% to 92%, the additional skeletal injuries
including high-specificity type injuries, of metaphyses and posterior
ribs.4 Evidence indicates it may also improve fracture aging.4,7,31,32
10
Journal of Forensic and Legal Medicine 60 (2018) 9–14
Histology is regarded as the most accurate means of demonstrating
and ageing fractures,3,4,14,33 especially those that are radiographically
subtle.3,14,33 For instance, it is reported that: histology confirmed 6 CML
identified as ‘abnormal’ on SS3; four from five rib fractures ‘suspected’,
but not diagnosable, on SS33 and metaphyseal fractures in two speci-
mens that appeared normal radiographically.3 However selecting tissue
for histological examination is key, as it is not practical to sample the
whole skeleton histologically. For this the histopathologist relies on the
radiologist and autopsy pathologist to guide sampling.
There are no large studies that have systematically described the
different fracture detection rates of post-mortem SS, CR and histo-
pathology and related them to histological age of fracture, bone, and
region of bone; and yet the evidence above suggests the right balance of
these techniques could be extremely important in optimising the de-
tection of abusive injuries. This is the rationale for our study.
2. Methods
This was a retrospective audit of anonymised data from all cases of
suspected non-accidental injuries referred to the histology department
of the Manchester Royal Infirmary for specialist histopathological
evaluation of skeletal injuries between 2011 and June 2017 (226 cases).
Inclusion criteria were:
• Child < 18 months old at the time of death.
• Recorded data from a post-mortem skeletal survey reported by a
specialist paediatric or musculoskeletal radiologist.
• Archived contact radiographs of all the bones. Only cases with CR in
2 planes for all bones examined were included.
A total of 38 infants met these criteria. The median age was 12
weeks (range: 19 days–14 months).
Fractures detected for each of the 3 tests, SS, CR and histology, were
tabulated and compared. The data for skeletal surveys was as reported
by the specialist radiologists. Contact radiographs are used by specialist
histopathologists to aid sampling for histological analysis. For this study
however, in addition to being reviewed by AF and ER, all CR radio-
graphs in which: no fracture could be seen; the appearances were felt to
be equivocal; or where a fracture was seen histologically but not
radiographically by the pathologists, were reviewed by a specialist
consultant musculoskeletal radiologist (PK). All fractures identified on
CR by AF, ER or PK were tabulated as positive. Histological recognition
of fractures was made by AF, an osteoarticular pathologist specialising
in the histological recognition and interpretation of fracture healing, in
the biopsy diagnosis of paediatric metabolic bone disease, and 27 years
medicolegal experience in recognising and aging fractures in infants.
Specimens of bone were received from Home Office and forensic
pathologists from across the UK as samples of bone in formalin. A short
clinical history and the results of SS were included in all cases used in
this paper. Contact radiography was performed on a Faxitron [Faxitron
Bioptics, LLC 3440 E Britannia Drive, Suite 150, Tucson, Arizona USA
85706]. Images were taken in 2 planes. Following this, bones were
decalcified and then widely sampled to include known and suspected
fracture sites from SS, autopsy findings, CR, and sites at which identi-
fying a fracture might be of significance in assessing abusive injuries
(mainly metaphyses and posterior elements of the ribs). The tissue was
taken into paraffin using conventional methods and then sectioned at
4 μm. Serial sections were stained with haematoxylin and eosin (H&E)
and Martius scarlet blue (MSB). Two sections were examined from most
tissue blocks but equivocal microscopic findings always resulted in
“deeper” sections being taken.
Fractures were aged histologically in a manner similar to that de-
scribed by Klotzbach et al.33 supplemented with our own data derived
from previously published studies examining the in-situ molecular
biology of fractures.34–36 Briefly:
E. Raynor et al. Journal of Forensic and Legal Medicine 60 (2018) 9–14

• Fractures were recognised as deficiencies in bone with a distinct Table 2

fracture gap and some associated reaction. All the metaphyseal
fractures extended into adjacent bone and were aged from the
changes within the bone.
• < 2 days – Haemorrhage in and around the fracture site with os-
teocyte necrosis in bone abutting the fracture line and, in some,
more polymorphs in areas of haemorrhage than one would expect to
have resulted from haemorrhage alone.
• < 7 days – Formation of granulation tissue, evidence of chronic
inflammation and formation of early fracture callus with con-
densation of connective tissue into primitive bone and cartilage like
structures.
• < 14 days – Obvious trabeculae of woven bone and nodules of
cartilage in the medulla and on periosteum typical of “primary
callus”.
• About 14 days – Almost all cases show widespread more intense
haematoxyphilia of the matrix indicative of mineralisation. Callus
bridges the fracture site.
• > 14 days – Onset of remodelling of primary callus.
• > 28 days – Formation of lamellar bone on the surface of existing
woven bone and cartilage nodules.
• > 2 months – Progressive matrix reorganisation.
In the context of this paper, fractures “≤14 days old” are those
showing neither widespread matrix colour change typical of miner-
alisation nor bridging callus formation. This time point was chosen
because it corresponds to defined radiological events.
Histologically detected fractures were tabulated by bone, site within
the bone and histological age.
3. Results
General findings: In no case did SS or CR detect a fracture that was
not identified histologically. Comparative analyses given below are
therefore described relative to the fractures identified histologically.
Histologically, 318 fractures were identified in the 38 infants. Of
these, 178 (56%) affected ribs, 119 (37.5%) limb long bones, 10 (3%)
skull and 11 (3.5%) were classified as ‘other’ fractures (3 clavicle, 4
cervical spine, 2 lumbar spine, 1 ankle not otherwise specified, 1 first
metacarpal).
Limb Long Bone Fractures: The 119 limb bone fractures were found
in 20/38 infants (53%); the number of fractures identified in a single
infant ranged from 1 to 19 (median: 4). The number and distribution by
bone are presented in Table 1.
The most common locations for limb long bone fractures were of, or
involved, the proximal tibial and the distal femoral metaphyses (17
fractures [14%] each). Thus 28% of all long bone fractures affected the
metaphyses around the knee.
Histologically, 19 of the limb long bone fractures were new frac-
tures through the site of an existing, older fracture (refractures). SS
detected 13 of the older fractures and CR 16, however neither SS nor CR
distinguished refracture from the older fracture. With this in mind a
Limb long bone fracture distribution within each bone.
Proximal metaphysis Shaft Distal metaphysis
Femur
Tibia
4
17
10
5
17
7
Humerus 9 8 11
Radius 5 3 6
Ulna 3 4 4
Fibula 3 0 3
decision was made to assess the detection rate of the imaging techni-
ques based on site alone, giving a total of 100 fracture sites.
46 long bone fractures were detected by both SS and CR (46%); 14
by CR only (14%); 15 by SS only (15%); and 25 were not detected by
either (25%). 48 fractures were aged ≤14 days; of these, SS failed to
identify 15 (31%), and CR 14 (29%). 52 fractures were aged > 14 days;
of these, SS failed to identify 24 lesions (46%), and CR failed to identify
26 lesions (50%).
Of the 119 long bone fractures, 46 (39%) were defined as meta-
physeal histologically. Generally, the detection rate of metaphyseal
fractures by SS and CR increased as the age of the lesion increased. 36
fractures were aged ≤14 days; of these, SS and CR each failed to
identify 9 (25%), but they were not the same 25%. 10 fractures were
aged > 14 days; of these, SS identified all, and CR 9 of the 10. This data
is presented in Table 3.
Rib Fractures: The total number of rib fractures was 178 (56% of
total number fractures). These were distributed among 21 infants, the
number of fractures per infant ranging from 2 to 26 (median: 6). Rib
fractures were identified in at least two adjacent ribs in 19 of the 21
infants.
There were generally more fractures of the left ribcage than of the
right ribcage. The seventh rib was the most frequently fractured rib of
both the left and the right ribcage (Graph 1).
25 fractures were refractures. Again SS and CR failed to identify any
of the refractures, therefore, as for the limb long bones, comparative
studies were based on 153 fracture sites rather than 178 fractures. 72
rib fractures (47%) were detected by SS and 83 (54%) by CR. 52
fractures (34%) were not detected by either SS or CR. The position in
the rib of the fractures and the number detected by SS and CR are given
in Table 4.
This analysis showed: 38 fractures of the anterior metaphyses, 14
shaft, and 18 posterior metaphyses 14 days old or less at the time of
death; and 26 anterior metaphyseal, 11 shaft, and 46 posterior meta-
physeal fractures aged > 14 days.
For those fractures dated ≤14 days, SS detected 8 (21%) of the
anterior metaphyseal fractures, 1 (7%) of the shaft fractures and 5
(28%) of the posterior metaphyseal fractures. CR increased the detec-
tion rate of anterior metaphyseal fractures to 37%, shaft fractures to
79% and posterior metaphyseal fractures to 44%.
Of the fractures more than 14 days old at the time of death: SS
detected 69% of the anterior metaphyseal fractures, 82% of shaft
fractures and 67% of posterior metaphyseal fractures, with CR

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)

Femur 31 (26%) 14 Both 22 (61%) 9 (90%)

Tibia 29 (24%) 10 SS only 5 (14%) 1 (10%)
Humerus 28 (24%) 8 CR only 5 (14%) 0 (0%)
Radius 14 (12%) 8 Total by SS 27 (75%) 10 (100%)
Ulna 11 (9%) 7 Total by CR 27 (75%) 9 (90%)
Fibula 6 (5%) 5 Neither 4 (11%) 0 (0%)
Total 119 (100%) 20 Total 36 (of 46 [78%]) 10 (of 10 [100%])

11
E. Raynor et al. Journal of Forensic and Legal Medicine 60 (2018) 9–14

we have presented all our data on SS and CR as a proportion of histo-

logically detected fractures. In very general terms skeletal survey, the
mainstay of fracture detection in infants suspected of abusive injury,
failed to detect 39% of all long limb bone fractures and 34% of all rib
fractures. In early fractures (defined here as fractures occurring 14 days
or less prior to death, as assessed histologically) SS failed to detect 31%
of long limb bone fractures and 80% of rib fractures.
In this study neither radiological technique could detect a re-
fracture. 44 of the 318 fractures (14%) were refractures and, as such,
each was evidence of 2 fracturing events at different times, a char-
acteristic of non-accidental injury. We have not included refractures in
the data analyses in this study, but had we done so their inclusion
would have further reduced the ability of SS to detect fractures im-
portant in identifying abusive injuries.
That histology identified all the fractures in specific parts of a bone,
whilst SS (even supplemented by CR) did not, evidences that histology
Graph. 1. Distribution of Rib Fractures by bone. should be considered the definitive test by which to identify fractures.
However, because it is impossible to sample the entire skeleton histo-
Table 4 logically, selection of bone for histological examination is entirely re-
The position of rib fractures by site within the ribs. liant on skeletal survey and post mortem examination findings to target
Anterior metaphysis Shaft Posterior metaphysis bone for histological assessment. Therefore within the context of
identifying fractures, skeletal survey, post mortem examination and
Both 20 (31%) 8 (32%) 26 (41%) histological examination are complimentary and, as such, all are es-
SS only 6 (9%) 2 (8%) 10 (16%)
sential for investigation of an infant suspected of having sustained non-
CR only 11 (17%) 10 (40%) 8 (13%)
Total by SS 26 (41%) 10 (40%) 36 (56%)
accidental injury.
Total by CR 31 (48%) 18 (72%) 34 (53%) From this study we were able to show that, in addition to radi-
Neither 27 (42%) 5 (20%) 20 (31%) ologically suspicious and autopsy findings, there are areas of the ske-
Total 64 (42%) 25 (16%) 64 (42%) leton that warrant sampling for histology (or more detailed imaging
studies in live infants) in all cases of suspected non-accidental injury in
infants, even when SS fails to identify a fracture. They are the anterior
increasing the detection rate of anterior metaphyseal fracture to 88%,
and posterior metaphyses of the 5th, 6th, 7th and 8th ribs bilaterally
shaft fractures to 82% and posterior metaphyseal fractures to 78%.
and the metaphyses around the knees. These are important sites as
Skull Fractures: The total number of skull fractures identified was
metaphyseal and posterior rib fractures are highly correlated with a
10 (3%). One fracture was dated > 14 days. All were detected by both
non-accidental cause.2–16 In our study, these are also areas of the ske-
SS and CR.
leton in which histology detected fractures unrecognised by skeletal
Other Fractures: The total number of ‘other’ fractures identified was
survey, particularly recent fractures. Our recommendation of focused
11 (3.5%; 3 clavicle, 4 cervical spine, 2 lumbar spine, 1 condylar ankle,
sampling of these areas of the skeleton in all cases of suspected non-
1 first metacarpal). SS and CR detected fractures in the peripheral bones
accidental injury is based on the following findings from this study:
but neither detected the spinal fractures all of which were metaphyseal
and predominantly of the transverse processes.
• Skeletal survey detected limb long bone metaphyseal fractures in
75% of cases in which histology dated the fracture as 14 days old or
4. Discussion less. This compares with 100% at fractures older than 14 days.
When the anterior and posterior rib metaphyses are taken together,
Despite the large number of cases received in our laboratory over SS detected 23% of fractures ≤14 days old and 68% > 14 days old.
the study period, only 38 met all the inclusion criteria. The largest • Fractures of the distal femoral and proximal tibial metaphyses ac-
single cause of non-inclusion was an absence of a comprehensive ske- counted for 28% of all limb long bone fractures.
letal survey report by a specialist radiologist. We deliberately chose • Fractures of the anterior and posterior ends of the 5th, 6th, 7th and
only cases reported by specialist radiologists to minimise the known 8th ribs accounted for two thirds of all rib fractures not detected by
inter-observer variation in skeletal survey reporting,37 and to try to SS.
provide a method by which maximum comparability could be brought
between radiological and specialist histological analyses. We also re- We have used a cut off of 14 days to distinguish “recent” and “old”
cognise that as our samples are taken by Forensic/Home Office Pa- fractures. Aging of the fractures comes from a histological assessment
thologists from around the UK, and that the sampling did not conform based upon a synthesis of the literature and our own published ex-
to a set protocol, bone sampling and post-mortem fracture detection perience. It must be acknowledged that there are no definitive pub-
therefore include selection bias. In the setting, statistical analysis would lished studies of the appearances of fractures of known ages in infants,
not have had validity and was not performed. an issue that needs to be urgently addressed. However there is a re-
Nevertheless, because we have made a detailed analysis comparing cognised sequence of events in fracture healing that enables the histo-
the detection by SS, CR and histology of the same 318 fractures, we pathologist to say whereabouts in the healing process a fracture has
believe the data are of considerable value to forensic pathologists, reached. Even if one were to ignore giving a specific age to a fracture
forensic/paediatric/musculoskeletal radiologists examining images in there is no doubt that early fractures, particularly non-displaced frac-
cases of potential child abuse, and all those considering the legal im- tures in which there is no mineralised matrix, are more difficult to
plications of differences between fractures detected by skeletal survey detect by skeletal survey.
and histological examination of bone in infants. Our study may help radiologists to focus on specific areas of the
Because of the design of this study, histology was the referent for skeleton for detailed analysis in live infants with existing or future in-
detecting fractures, an approach not used previously. As both SS and CR vestigative techniques. It also flags up the number of “old” fractures in
failed to demonstrate any fractures that were not detected by histology, infants later dying from a death suspected of being non-accidental. In

12
E. Raynor et al.
our study 27 of the 38 infants (71%) had at least one fracture > 14 days
old. This emphasises the proportion of cases of infants suspected as
having died an abusive death who have evidence of previous, probably
inflicted, skeletal injuries.
We have included contact/specimen radiography here because of a
previously reported study comparing SS, CR and histology in 3 in-
fants.33 In that study CR was the referent and it noted that all fractures
detected by CR were also detected by histology. If histopathology was
superior to CR in detecting fractures the study design would not have
allowed this to become apparent. Equally the study design would have
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
that SS and CR detect similar numbers of fractures but that the 2
techniques are complimentary, together recognising a greater number
of fractures that either technique alone. However, even in combination,
they are less efficient at identifying fractures than histology. An argu-
ment could be made that the CR data appear less accurate than his-
tology and different from SS because the contact radiographs are re-
ported by histopathologists lacking skill in interpreting radiographs. We
have tried to control for that by having a specialist musculoskeletal
radiologist review our findings. The relative influence of the detection
of fractures by the use of CR within the context of this paper which is
designed to compare SS and histopathology is probably small as in our
experience CR is not used to directly detect previously unrecognised
fractures so much as to guide the histopathologist to areas of bone to
sample histologically.
Although most of the cases coming to our laboratory have SS as the
only imaging study of non-cranial bones, a small, but increasing min-
ority of cases, have undergone post mortem CT (PMCT). The potential
of this technique is discussed in Arthurs et al.38 In one study,39 com-
paring thoracic radiography, thoracic PMCT, and post mortem ex-
amination (including histology) fractures were identified by a specialist
radiologist radiographically in 7/12 patients who had fractures found at
autopsy, and on CT in 11/12 patients. In total, 46% of fractures were
reported on radiography, and 85% of fractures were reported on PMCT
(a senior staff radiologist recorded 29% of fractures on the same
radiographs and 51% on CT). Whilst supporting our contention that
histology is essential in detecting fractures, this also emphasises the
potential of PMCT.
Increasingly, bones received in our laboratory have been in-
vestigated by PM micro-CT. This is a largely research tool, but one that
has the potential to demonstrate the microstructure of fracture callus at
a resolution similar to that of low power microscopy.40 It requires a
similar approach to sampling of bone tissue as histology. A careful
comparative study of PM micro-CT and histology is awaited.
5. Conclusion
This study has shown that post-mortem bone histology is superior to
post mortem skeletal survey and contact radiography for detecting
fractures in infants, particularly re-fractures, which have forensic im-
plications, and acute fractures of ribs and metaphyses. However, our
experience is that skeletal survey (together with macroscopic autopsy
findings) is essential in directing histopathologists to sites of potential
fractures. Without these findings to guide the histopathologist, histo-
logical assessment would be almost impossible. We have shown, how-
ever, that there are specific sites in the body (namely ribs 5, 6, 7 and 8
and metaphyses around the knee) that should be routinely sampled and
submitted for histological assessment even if no fracture is suspected on
SS or at autopsy as these often yield vital data that would otherwise be
missed.
A significant majority of the 38 infants studied here had evidence of
healing fractures not associated with the terminal event, many of which
were however typical of non-accidental injury.
13
Journal of Forensic and Legal Medicine 60 (2018) 9–14
Declarations of interest
Miss Emma Raynor: None
Praveen Konala: None
Professor Anthony Freemont: None
Professor Freemont is supported by the MRC and EPSRC (MRC
Grant MR/N00583X/1) for work on molecular biomarkers and is part of
the Musculoskeletal Division of the Manchester NIHR BRC.
Acknowledgements
This research did not receive any specific grant from funding
agencies in the public, commercial, or not-for-profit sectors.
References
1. Caffey J. The classic: multiple fractures in the long bones of infants suffering from
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
abuse (non-accidental injury). Pediatr Radiol. 2009;39(5):461–470.
3. Kleinman PK, Marks SC, Blackbourne B. The metaphyseal lesion in abused infants: a
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
postmortem radiologic-histopathologic study in 31 infants. AJR Am J Roentgenol.
1995;165(3):647–650.
5. Kleinman PK. Diagnostic imaging in infant abuse. AJR Am J Roentgenol.
1990;155(4):703–712.
6. Kleinman PK, Marks SC, Adams VI, Blackbourne BD. Factors affecting visualization of
posterior rib fractures in abused infants. AJR Am J Roentgenol. 1988;150(3):635–638.
7. Dwek JR. The radiographic approach to child abuse. Clin Orthop Relat Res.
2011;469(3):776–789.
8. Paddock M, Sprigg A, Offiah AC. Imaging and reporting considerations for suspected
physical abuse (non-accidental injury) in infants and young children. Part 1: initial
considerations and appendicular skeleton. Clin Radiol. 2017;72(3):179–188.
9. Barber I, Perez-Rossello JM, Wilson CR, Kleinman PK. The yield of high-detail
radiographic skeletal surveys in suspected infant abuse. Pediatr Radiol.
2014;45(1):69–80.
10. Barber I, Kleinman PK. Imaging of skeletal injuries associated with abusive head
trauma. Pediatr Radiol. 2014;44(Suppl4):613–620.
11. van Rijn RR, Bilo RA, Robben SG. Birth-related mid-posterior rib fractures in neo-
nates: a report of three cases (and a possible fourth case) and a review of the lit-
erature. Pediatr Radiol. 2009;39(1):30–34.
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.
13. Borg K, Hodes D. Guidelines for skeletal survey in young children with fractures.
Arch Dis Child Educ Pract Ed. 2015;100(5):253–256.
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
Radiol. 2014;44(2):124–140.
15. Maguire S. Which injuries may indicate child abuse? Arch Dis Child Educ Pract Ed.
2010;95(6):170–177.
16. Leventhal JM, Martin KD, Asnes AG. Incidence of fractures attributable to abuse in
young hospitalized children: results from analysis of a United States database.
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
as an indicator of nonaccidental trauma in children. J Trauma.
2003;54(6):1107–1110.
19. Lloyd-Roberts G. The diagnosis of injury of bones and joints in young babies. Proc
Roy Soc Med. 1968;61(12):1299–1300.
20. Bajaj M, Offiah AC. Imaging in suspected child abuse: necessity or radiation hazard?
Arch Dis Child. 2015;100(12):1163–1168.
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.
Lancet. 2012;380(9840):499–505.
22. Johnson JN, Hornik CP, Li JS, et al. Cumulative radiation exposure and cancer risk
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
MRI in suspected infant abuse. AJR Am J Roentgenol. 2010;195(3):744–750.
24. Mandelstam SA, Cook D, Fitzgerald M, Ditchfield MR. Complementary use of radi-
ological skeletal survey and bone scintigraphy in detection of bony injuries in sus-
pected child abuse. Arch Dis Child. 2003;88(5):387–390.
25. Kemp AM, Butler A, Morris S, et al. Which radiological investigations should be
performed to identify fractures in suspected child abuse? Clin Radiol.
2006;61(9):723–736.
26. Child abuse. BMJ Best Practice. 2017; 2017http://bestpractice.bmj.com/topics/en-
us/846.
27. Sieswerda-Hoogendoorn T, van Rijn RR, Robben SGF. Classic metaphyseal lesion
following vaginal breech birth, a rare birth trauma. J Forensic Radiology Imaging.
2014;2(1):2–4.
E. Raynor et al.
28. O'Connell A, Donoghue VB. Can classic metaphyseal lesions follow uncomplicated
caesarean section? Pediatr Radiol. 2007;37(5):488–491.
29. Lysack JT, Soboleski D. Classic metaphyseal lesion following external cephalic ver-
sion and cesarean section. Pediatr Radiol. 2003;33(6):422–424.
30. Kleinman PK, Nimkin K, Spevak MR, et al. Follow-up skeletal surveys in suspected
child abuse. AJR Am J Roentgenol. 1996;167(4):893–896.
31. Geller E. Imaging in Child Abuse. 2017; 2017http://emedicine.medscape.com/article/
407144- overview.
32. Worlock P, Stower M, Barbor P. Patterns of fractures in accidental and non-accidental
injury in children: a comparative study. Br Med J. 1986;293(6539):100–102.
33. Klotzbach H, Delling G, Richter E, Sperhake JP, Püschel K. Post-mortem diagnosis
and age estimation of infants' fractures. Int J Leg Med. 2003;117(2):82–89.
34. Kwong FN, Hoyland JA, Evans CH, Freemont AJ. Regional and cellular localisation of
BMPs and their inhibitors' expression in human fractures. Int Orthop.
2009;33(1):281–288.
14
Journal of Forensic and Legal Medicine 60 (2018) 9–14
35. Andrew JG, Hoyland JA, Freemont AJ, Marsh DR. Platelet-derived growth factor
expression in normally healing human fractures. Bone. 1995;16(4):455–460.
36. Andrew JG, Andrew SM, Freemont AJ, Marsh DR. Inflammatory cells in normal
human fracture healing. Acta Orthop Scand. 1994;65(4):462–466.
37. Thompson NB, Kelly DM, Warner Jr WC, et al. Intraobserver and interobserver re-
liability and the role of fracture morphology in classifying femoral shaft fractures in
young children. J Pediatr Orthop. 2014;34(3):352–358.
38. Arthurs OJ, Hutchinson JC, Sebire NJ. Current issues in postmortem imaging of
perinatal and forensic childhood deaths. Forensic Sci Med Pathol. 2017;13(1):58–66.
39. Hong TS, Reyes JA, Moineddin R, Chiasson DA, Berdon WE, Babyn PS. Value of
postmortem thoracic CT over radiography in imaging of pediatric rib fractures.
Pediatr Radiol. 2011;41(6):736–748.
40. Rutty GN, Brough A, Biggs MJ, Robinson C, Lawes SD, Hainsworth SV. The role of
micro-computed tomography in forensic investigations. Forensic Sci Int.
2013;225(1–3):60–66.