- Email: [email protected]
Rib fractures identified at post-mortem examination in sudden unexpected deaths in infancy (SUDI)
Forensic Science International 189 (2009) 75–81
Contents lists available at ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Rib fractures identified at post-mortem examination in sudden unexpected deaths in infancy (SUDI) Martin A. Weber a, R. Anthony Risdon a, Amaka C. Offiah b, Marian Malone a, Neil J. Sebire a,* a b
Department of Paediatric Pathology, Great Ormond Street Hospital for Children and UCL Institute of Child Health, Great Ormond Street, London WC1N 3JH, UK Department of Radiology, Great Ormond Street Hospital for Children, Great Ormond Street, London, UK
A R T I C L E I N F O
A B S T R A C T
Article history: Received 27 August 2008 Received in revised form 22 January 2009 Accepted 14 April 2009 Available online 23 May 2009
Rib fractures may be associated with non-accidental injury (NAI) in infancy, but the possible significance of fresh rib fractures in relation to resuscitation remains undetermined. Consequently, it is important to detect and confirm the presence of rib fractures when performing a post-mortem examination, particularly in the context of sudden unexpected death in infancy (SUDI). At our centre, it has been local policy to perform routine radiological skeletal surveys and detailed post-mortem examination of the ribs in all SUDI autopsies. The aim of this study is to establish the characteristics of all rib fractures identified at autopsy in the setting of SUDI from a large series of cases examined at a single specialist centre. As part of a larger review of paediatric post-mortem examinations performed at a single specialist institution over a 10-year period (1996–2005), all cases presenting as SUDI (aged 7–365 days) were identified and their anonymised records searched to identify all cases in which rib fractures were recorded. Over the 10-year period, 546 post-mortem examinations were performed for the indication of SUDI, including 94 forensic autopsies. Rib fractures were identified in 24 cases (4%). 15 infants (3% of SUDI) demonstrated healing rib fractures, of which 10 infants (67%) showed additional features suggestive of NAI. The other 9 infants (2% of SUDI) demonstrated fresh rib fractures only with no surrounding tissue reaction histologically; in 7 (78%) of these there were no other injuries and the fresh fractures were interpreted to have been caused by resuscitation-related trauma. All of the resuscitation-related fractures were situated in the anterolateral chest, in contrast to NAI-associated fractures, which were located in the anterolateral and/or posterior chest. Anterior costochondral junction fractures were also seen in a minority of NAI-associated cases, but such fractures were not seen in apparent resuscitationrelated cases. Compared to healing rib fractures, which were detected on skeletal survey in 93%, fresh rib fractures were only detected in 22% of skeletal surveys. Rib fractures are uncommon in infancy and may indicate NAI, particularly when healed or healing, posterior or involving the costochondral junction. Fresh rib fractures may be missed on skeletal survey, but can be reliably detected at post-mortem examination following stripping of the pleura and detailed examination of each rib. Fresh anterolateral fractures, which may be multiple, contiguous and even bilateral, are highly likely to be related to resuscitation if there are no other associated injuries. ß 2009 Elsevier Ireland Ltd. All rights reserved.
Keywords: Rib fracture Infancy Sudden unexpected death Resuscitation Non-accidental injury
1. Introduction The presence of rib fractures is well-recognised as a component of non-accidental injury (NAI) in infancy. Classically, in this setting, such fractures are often multiple, of varying ages and usually associated with other features of abuse or trauma [1]. However, the clinical significance of fresh rib fractures in relation to terminal cardiopulmonary resuscitation (CPR) remains undetermined, even though incorrect interpretation of such fractures may have major
* Corresponding author. Tel.: +44 020 7829 8663; fax: +44 020 7829 7875. E-mail address: [email protected] (N.J. Sebire). 0379-0738/$ – see front matter ß 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2009.04.015
medicolegal implications. In the United Kingdom, it is recommended that all autopsies of sudden unexpected deaths in infancy (SUDI) are performed on behalf of Her Majesty’s Coroner (HMC) by specialist paediatric pathologists, unless there are significant concerns about possible NAI, in which case it is recommended that the autopsy is carried out in conjunction with a Home Officeaccredited, forensic pathologist [2]. All SUDI post-mortem examinations are carried out according to a standard, published protocol [2], which includes a radiological skeletal survey reported by a paediatric radiologist prior to the autopsy to detect bony injuries which may be associated with NAI. The detection and confirmation of rib fractures is also an integral part of the postmortem examination, both by macroscopic inspection of the ribs
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
76
and by means of histopathological examination of apparent fractures for definitive confirmation and dating of such fractures. However, the prevalence of such fractures in infancy, and in particular, the frequency of probable resuscitation-related fractures, remains uncertain. The aims of this study are, therefore, to establish the characteristics of all rib fractures identified at autopsy in the setting of SUDI from a large series of cases examined at a single specialist centre, and, specifically, to differentiate between fresh and healing rib fractures in relation to other findings at autopsy.
2. Materials and methods Great Ormond Street Hospital for Children, London, is a tertiary specialist hospital for the investigation and treatment of a range of paediatric diseases; the pathology department performs autopsies on the behalf of HMC from a wide geographical area, currently carrying out >300 paediatric post-mortem examinations per annum. As part of a large systematic review of paediatric post-mortem examinations performed during a 10-year period, from 1996 to 2005 inclusive, all autopsies carried out for the indication of SUDI (aged 7–365 days) were identified and their anonymised records searched to identify cases in which rib fractures were present. In most cases in this series, a radiological skeletal survey was carried out prior to the post-mortem examination, whilst in a small proportion of infants, usually due to a period of post-resuscitation survival, antemortem radiographs were available for radiological review. Radiological skeletal surveys comprised a panel of plain radiographs of the entire skeleton, including multiple images of skull, limbs and thoracic cage. Radiographs were reported by specialist paediatric radiologists with expertise in NAI. All post-mortem examinations were carried out by specialist paediatric pathologists, in addition to Home Office-accredited forensic pathologists for forensic autopsies. The ribs were routinely examined during the post-mortem examination as part of the standard autopsy protocol, which included stripping of the parietal pleura from the thoracic cage and careful macroscopic inspection of individual ribs, as well as microscopic examination of any suspected rib fractures. Further specific details were extracted from the database for infants in whom rib fractures were identified, including whether the fractures were identified radiologically prior to the autopsy, the number and distribution of the fractures, comments on radiological and histopathological dating of the fractures, as well as
the presence of other associated abnormalities suggestive of NAI detected at postmortem examination. For the purposes of this study, NAI was defined as the presence of demonstrable injuries not explained by the clinical history, the suspicion having been raised either at clinical presentation or by the pathologist at the time of the post-mortem examination, and was not based on the outcomes of medicolegal or criminal proceedings. Deaths in which the injuries were deemed insufficiently severe to account for the sudden death of the infant were classified as ‘undetermined.’ Rib fractures were either classified as ‘fresh’ or ‘healing,’ the former defined as the absence of a vital reaction (with or without some surrounding haemorrhage), indicating that such fractures occurred in the perimortem or immediate post-mortem period, whilst ‘healing’ fractures showed radiological, macroscopic and/or histological evidence of inflammation, granulation tissue formation, fibrosis and/or callus (or osteoid) formation, indicating that the fracture occurred some time prior to death. The study was approved by the Local Research Ethics Committee.
3. Results During the 10-year study period, a total of 1516 paediatric autopsies were carried out in the Department of Paediatric Pathology, including 546 post-mortem examinations performed for the indication of SUDI. These included 452 infant deaths in which there was no preceding clinical suspicion of NAI and 94 forensic (‘special’) autopsies jointly carried out by forensic and paediatric pathologists. The majority of infants presenting with sudden collapse underwent varying degrees of CPR, including chest compressions, by carers, ambulance crew and/or medical staff, but specific details of the resuscitative attempts with regards to personnel, duration and vigour were not recorded. Overall, rib fractures were present in 25 (5%) of all SUDI (Table 1). One case (Case No. 25, Table 1) was a co-sleeping associated sudden unexpected death of an 18-day-old neonate which occurred abroad. The initial autopsy was performed in that country, and a second post-mortem examination was subsequently carried out at our centre on the embalmed body of the
Table 1 Table of all rib fractures. Rib fractures described as ‘anterior’ (excluding CCJ fractures) or ‘lateral,’ and fractures located in the ‘midclavicular line’ or in the ‘mid-axillary line,’ are categorised as ‘anterolateral’ fractures. Specific details, including dating estimates, of the fractures are based on radiological or histological appearances, as reported at the time of the post-mortem examination. In the small number of cases in which there were discrepancies between the radiological and macroscopic findings relating to the site and/or number of rib fracture(s) or the estimated fracture age(s), the findings in the table are based on the macroscopic examination unless indicated otherwise. (*Based on antemortem radiographs, usually due to a period of post-resuscitation survival in hospital following the initial collapse; **detected radiologically only and not commented on in the macroscopic description or on histological examination; CCJ = costochondral junction). Case number
Cause of death
Age in months
Rib fractures Detected on skeletal survey
Site(s)
Specific details (estimated age of fractures, if given)
Healing rib fractures 1 Head injury (NAI)
2.7
Yes
Anterolateral
Left anterolateral 5, 6, 7 (2–5 weeks)
2
Undetermined
2.2
Yes
Posterior
Left posterior 6, 7, 8 (2–3 weeks)
None
3
Undetermined
5.5
Yes
Anterolateral, posterior
Left anterolateral 7, 8
Multiple fractures of thoracic vertebrae Fracture upper left tibia with early healing reaction Retinal haemorrhage
Left posterior 5, 6, 7 Right anterolateral 7, 8 (‘varying ages—some quite old’)
Features suggestive of NAI other than rib fractures
Head injury
4
Head injury (accidental)
7.3
Yes*
Posterior
Left posterior 10 (2–5 weeks)
None
5
Head injury (NAI)
7.8
Yes
CCJ, anterolateral, posterior
Left anterior (CCJ) 1, 3, 4, 5, 6, 7, 8 (1 week for most fractures) Left posterior 7, 10 (3–5 weeks) Right anterior (CCJ) 1, 4, 5, 6, 7 (1 week for most fractures) Right anterolateral 7 (10–12 weeks)
Head injury, facial bruising
6
Undetermined
7.4
Yes
Anterolateral
Right anterolateral 6, 7, 8 (3–6 weeks)
None
7
Head injury (NAI)
4.9
Yes
CCJ, anterolateral, posterior
Left anterolateral 4, 5, 6, 7, 8 Left posterior 7, 9, 10 Right anterior (CCJ) 5, 6, 7, 8 Right posterior 6, 7, 8 (all around 2–4 weeks)
Head injury
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
77
Table 1 (Continued ) Case number
8
Cause of death
Head injury (NAI)
Age in months
Rib fractures Detected on skeletal survey
Site(s)
Specific details (estimated age of fractures, if given)
2.9
Yes
Anterolateral, posterior
Left anterolateral 7 with recent re-fracture Left posterior 11, 12 Right posterior 10 with recent re-fracture (all about 4-6 weeks, re-fractures more recent)
Features suggestive of NAI other than rib fractures
Head injury Healing metaphyseal fracture left radius Healing metaphyseal fracture right tibia Recent metaphyseal fracture left tibia Torn frenulum Cigarette burns
9
Undetermined
3.6
Yes
Anterolateral
Left anterolateral 4, 5, 6
Healing tibial fracture Bruise of forehead
10
Undetermined
5.6
Yes
Posterior
Right posterior 9 (4–6 weeks)
None
11
Head injury (NAI)
2.8
Yes
CCJ**, posterior
Left anterior (CCJ) 12** Left posterior 7, 9, 10, 11 (2) (4–6 weeks for most fractures) Right anterior (CCJ) 11, 12** Right posterior 9, 10 (4–6 weeks)
Head injury
12
Inflicted asphyxia (NAI)
1.4
Yes
CCJ, anterolateral, posterior
Left posterior 6, 7, 8 (1–2 weeks)
None
Right anterior (CCJ) 10 (1–2 weeks) Right anterolateral 3,4 (fresh) Right anterolateral 5, 6, 7 (fresh re-fractures of previous fractures 1–2 weeks old) Right posterior 10, 11, 12 (1–2 weeks) 13
Traumatic liver rupture (NAI)
9.9
Yes
Posterior
Left posterior 10, 11 (8 weeks)
Traumatic liver rupture
Right posterior 7, 8 (‘recent’) Right posterior 9 (‘several weeks’ with fresh re-fracture)
Skull fracture Widespread bruising
14
Head injury (NAI)
1.4
No*
Posterior
Left posterior 2, 3, 4, 5, 6, 7 (2–3 weeks)
Head injury
15
Undetermined
4.8
Yes
Posterior
Left posterior 7, 9
Healing microfractures of both radii and ulnae Gum scar
Right posterior 4, 5, 6, 10, 11 (varying ages, 4–6 weeks to >2 months) Fresh rib fractures 16 Inflicted asphyxia (NAI)
<1
Yes
Anterolateral, posterior
Left anterolateral 1, 2, 3, 4, 5, 6, 7, 8 (fresh)
(Second autopsy): extensive subdural and subarachnoid haemorrhage around spinal cord
Right anterolateral 2, 3, 4, 5, 6, 7, 8 (fresh) Right posterior 1, 2 , 3, 4, 5, 6, 7 (fresh) 17
Head injury (NAI)
4.5
No
Anterolateral
Right anterolateral 5, 7 (fresh)
Head injury
18
Undetermined
<1
No
Anterolateral
Left anterolateral 3, 4, 5, 6 (fresh)
None
19
Undetermined
2.5
No
Anterolateral
Left anterolateral 2, 3, 4, 5 (fresh)
None
20
Undetermined
1.4
No
Anterolateral
Right anterolateral 4, 5, 6 (fresh)
None
21
Undetermined
2.6
No
Anterolateral
Left anterolateral 4, 5, 6 (fresh)
None
22
Undetermined
2.0
No
Anterolateral
Right anterolateral 4, 5 (fresh)
None
23
Undetermined
2.0
Yes
Anterolateral
Left anterolateral 3, 4, 5, 6 (fresh) Right anterolateral 3, 4, 5, 6 (fresh)
None
24
Undetermined
2.3
No
Anterolateral
Left anterolateral 3, 4, 5, 6 (fresh)
None
25
Undetermined
<1
No
Anterolateral
Left anterolateral 1, 2, 3, 4, 5, 6, 7 (fresh)
(Second autopsy; embalmed, rib fractures interpreted to be post-mortem): none
infant. The second autopsy revealed seven fresh rib fractures, but these were interpreted to have occurred post-mortem, and the case was thus excluded from further analysis. Of the remaining 24 infants, there were healing fractures in 15 (3% of all SUDI; 63% of those with rib fractures; Table 2 and Fig. 1), and of these, 10 (67% of all healing fractures) were associated with additional features at post-mortem examination suggestive
of NAI. In the other nine infants (2% of all SUDI; 37% of those with rib fractures), only fresh rib fractures were identified, with no histological evidence of a surrounding tissue reaction (Table 2 and Fig. 2). In seven of these (<2% of all SUDI; 29% of those with rib fractures; 78% of those with fresh rib fractures), there were no other injuries and no additional features, either clinically or at autopsy, suggestive of NAI or major trauma, suggesting that
78
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
Table 2 Summarised overview of the 24 cases with rib fractures included in the analysis, categorised into cases with healing and exclusively fresh rib fractures, the latter subdivided into fresh rib fractures associated with NAI and presumed resuscitation-related fractures (CCJ = costochondral junction).
these fractures were most likely caused by resuscitationassociated trauma. These apparent resuscitation-related fresh rib fractures were multiple, all involving the (left or right) fourth and fifth ribs, and almost exclusively limited to the third, fourth,
fifth and/or sixth ribs, and were bilateral in one case (Fig. 3). Importantly, all such resuscitation-related fractures were located anterolaterally, compared to the predominantly posterior rib fractures in infants in which other features of NAI were present
Fig. 1. Healing rib fractures. (A) Chest radiograph (faxitron) showing healing rib fractures with callus formation. (B) Macroscopic appearances of healing rib fractures in an infant with suspected NAI. (C and D) Histological features of healing rib fractures, with subperiosteal new bone formation. In (D), a recent re-fracture has occurred (white arrow).
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
79
Fig. 2. Fresh rib fractures. (A) Macroscopic appearances. (B) (H&E), (C and D) (Masson Trichrome). Photomicrographs demonstrating the fracture line (arrows) associated with minor fresh haemorrhage only (higher power in D). There is no surrounding inflammatory reaction or periosteal reaction.
(Fig. 4). Furthermore, none of the apparent resuscitation-related fractures involved the costochondral junction; such fractures, which were described in four infants, were only found in association with other healing fractures and, in three cases, with other injuries suggestive of NAI. The presence of healing rib fractures was identified prior to post-mortem examination on routine post-mortem radiological skeletal survey in 14 (93%) cases, the one case in which the presence of healing rib fractures was not detected prior to autopsy being an infant who, owing to a period of post-resuscitation survival, had an antemortem chest radiograph, which, even on review, failed to disclose the presence of posterior healing rib fractures. In contrast, of the nine cases with fresh rib fractures, the presence of fractures was detected on post-mortem radiological skeletal survey prior to the autopsy in only two (22%) cases. These infants all demonstrated linear, non-displaced fractures with no or minimal surrounding haemorrhage, the fractures being identifiable only on careful inspection of the ribs following stripping of the pleura.
4. Discussion The findings of this study demonstrate that rib fractures are present in up to 5% of cases referred for post-mortem examination for the indication of SUDI. This may be an overestimate of the true prevalence in an unselected population of SUDI, since Great Ormond Street Hospital is a tertiary referral centre and may preselect cases in which there is a preceding history or clinical suspicion of NAI or trauma. Nevertheless, the findings of this study have further demonstrated that in about two thirds of infants in whom rib fractures are present, such fractures show features of healing, whereas in about one third, only fresh rib fractures are present. In the cases in which only fresh rib fractures are identified, the majority are not associated with other injuries or additional features to suggest NAI and are therefore likely to represent resuscitation-related rib fractures; such fractures are seen in 1–2% of otherwise uncomplicated SUDI. These fresh, linear fractures are present anterolaterally in the chest wall and affect contiguous ribs, whereas in infants with other features suggestive of NAI, rib
80
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
Fig. 3. Schematic representation of the distribution of the fresh, presumed resuscitation-related rib fractures. These fractures are located in the anterolateral chest only. Each circle represents one fresh fracture, and each colour represents one case. The image is not drawn to scale, and the circles do not indicate the exact location of specific rib fractures.
fractures are more likely to be healing, posterior or, when anterior, affect the costochondral junction. Due to the limited mobility of infants, accidental trauma, and therefore rib fractures, are uncommon in infancy but may result from traumatic forceful squeezing or compression of the chest. Their presence in life is generally considered to be a strong indicator of possible NAI, particularly when such fractures are healed or healing, multiple and/or of different ages. Several causes of rib fractures in children have previously been described in studies reporting on series of children attending emergency departments. These include NAI, major trauma (such as motor vehicle accidents or MVAs) and rare cases associated with birth injury, neonatal physiotherapy for bronchiolitis or pneumonia, and underlying conditions such as rickets/osteoporosis or osteogenesis imperfecta [3–6]. Pooled data from five series reporting on cohorts of infants with rib fractures indicate that 62% (87 of 140) are associated with NAI [3–5,7,8], and it has been suggested that, due to their rarity in the overall population in this age group, the
positive predictive value of rib fractures for NAI is >95% in children below the age of 3 years [9]. Rib fractures associated with accidental injury, on the other hand, are almost always seen in conjunction with major trauma of comparable magnitude to that associated with MVAs [5]. In one study of 2080 children presenting to a trauma unit, there were 33 (1.6%) children with rib fractures. This group had a mortality rate of 42% since the children with rib fractures were significantly more severely injured, the presence of rib fractures on admission therefore being a marker of severe traumatic injury of whatever cause [8]. Several previous studies have also attempted to specifically determine the frequency of rib fractures following CPR, these being reported but rare, with a prevalence of <1% of witnessed CPR episodes [3,10,11]. A recent systematic review reporting on a possible association of CPR and rib fractures in children identified a total of six studies (with 923 children who had undergone CPR), with a rib fracture prevalence of 3/923 (0.3% or about 1 in 300 episodes); all three infants had anterior fractures [12]. It was also previously noted that resuscitation-related injuries, although rare, are more frequently caused by physicians rather than non-medical persons [10]. The prevalence of apparent resuscitation-related fractures was greater in the present study than reported above (1– 2% vs. 0.3%), almost certainly because our series only comprised infants that had undergone an autopsy, which included detailed examination of the ribs with the specific aim of documenting the presence of such fractures. Fresh rib fractures may not reliably be identified on the radiographic skeletal survey, but such fractures can be detected at post-mortem examination following stripping of the pleura and detailed examination of each rib. It should therefore be routine practice in all SUDI autopsies to specifically examine the thoracic cage for the presence of rib fractures and any possible rib fracture(s) identified should be further investigated histologically in order to confirm their presence and to provide further information regarding possible timing of the fracture(s) relative to the time of death. Histologically, fresh haemorrhage is the earliest finding following rib fracture, soon followed by the presence of a localised inflammatory reaction, with osteoblastic and osteoclastic activity starting as early as two to four days following the initial injury; mineralisation of the newly formed osteoid occurs approximately 1 week later [13]. Fresh rib fractures will not demonstrate any histological evidence of an inflammatory response but may be associated with a small amount of local haemorrhage, although resuscitation-related rib fractures, in cases
Fig. 4. Schematic representation of the distribution of all rib fractures associated with other injuries suggestive of NAI. Each circle represents one fresh fracture, and each triangle a healing fracture. Each colour represents one case. The image is not drawn to scale, the circles and triangles do not indicate the exact location of specific rib fractures, and the triangles do not distinguish between fractures of different healing ages. Such NAI-associated fractures include fresh and healing fractures, posterior fractures and anterior fractures, as well as anterior fractures affecting the costochondral junction.
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
in which CPR proved unsuccessful, would be expected to be associated with only minimal haemorrhage due to the lack of a circulation at the time of fracture. In one previous series of 70 infant autopsies, subtle fractures of the anterolateral aspects of the ribs were apparently identified in around 10% of cases, all of which showed little, if any, associated blood extravasation and might have been missed had the parietal pleura not been stripped [14]. There are a variety of reasons that rib fractures, especially fresh fractures, may be difficult to identify on routine skeletal radiology; these include superimposition of the transverse process over a posterior rib fracture site, a fracture line that crosses obliquely to the beam and non-displacement of fragments due to preservation of the periosteum, and therefore multiple projections may be required [15]. In a previous study of 31 infants who died due to NAI, only 30 of 84 (36%) rib fractures were identified on skeletal survey examination [16]. These findings are also consistent with recent data from an animal model of severe inflicted trauma, in which detection of rib fractures was more frequent at autopsy compared to radiological examination [17]. In the present study, of the 24 rib fractures present in total, 16 (67%) were detected, this apparent improvement in detection rate probably representing a combination of improvements in imaging technologies and increasing operator expertise. Nevertheless, the presence of fresh rib fractures, with absence of a surrounding tissue reaction, may not be identified on routine radiographic skeletal survey, even when performed by specialist paediatric radiologists with an interest in NAI. These cases would only be recognised at autopsy on detailed examination of the ribs but, as the majority of these otherwise ‘undetectable’ fractures are likely to be due to resuscitation-related trauma rather than NAI when located in the anterior chest, their ‘non-detection’ would arguably be of limited clinical significance. In accordance with the findings of the present study, previous data indicate that in infants the site of the rib fracture is important for correct interpretation of its significance. One study reported that fractures predominantly involving the posterior ends of the ribs occurred in 15 of 43 deaths due to severe trauma, and concluded that, in infants without metabolic bone disease and in whom there is no other history of severe trauma, fractures in the dorsal part of the chest wall should be interpreted as a strong indication of physical child abuse [10]. The mechanism for developing such posterior fractures is now well-recognised, with cadaver and animal studies indicating that posterior rib fractures are a consequence of severe anteroposterior chest compression and excessive posterior levering of the ribs, and would require forces of similar magnitude to those generated by a MVA [18,19], although it is recognised that such effects may be modified by additional mechanical factors, such as the time period and surface area over which the force is applied, and responses of the gripping, restraining or impact material. Pathologists must specifically search for rib fractures in all infant autopsies and must ensure that they evaluate the findings within the overall context of the case when interpreting the significance of fresh rib fractures at post-mortem examination. Although fresh anterior fractures may be seen in NAI, the presence
81
of fresh anterior fractures with no other associated injuries is highly likely to be resuscitation-associated and unrelated to the cause of death, but histological confirmation is required in all cases. Acknowledgements This study was supported by a project grant from the Foundation for the Study of Infant Deaths (FSID). The authors would like to thank Detective Constable Lorraine Gronbach and the Graphics Department of the Metropolitan Police for supplying the illustrations of the thorax used in Figs. 3 and 4. References [1] C.J. Hobbs, ABC of child abuse. Fractures, BMJ 298 (1989) 1015–1018. [2] H. Kennedy (Ed.), Sudden unexpected death in infancy. A multi-agency protocol for care and investigation. The report of a working group convened by The Royal College of Pathologists and The Royal College of Paediatrics and Child Health, The Royal College of Pathologists and The Royal College of Paediatrics and Child Health, London, 2004. [3] K.W. Feldman, D.K. Brewer, Child abuse, cardiopulmonary resuscitation, and rib fractures, Pediatrics 73 (1984) 339–342. [4] P. Schweich, G. Fleisher, Rib fractures in children, Pediatr. Emerg. Care 1 (1985) 187–189. [5] B. Bulloch, C.J. Schubert, P.D. Brophy, N. Johnson, M.H. Reed, R.A. Shapiro, Cause and clinical characteristics of rib fractures in infants, Pediatrics 105 (2000) E48. [6] M. Chalumeau, L. Foix-L’Helias, P. Scheinmann, P. Zuani, D. Gendrel, H. Ducou-lePointe, Rib fractures after chest physiotherapy for bronchiolitis or pneumonia in infants, Pediatr. Radiol. 32 (2002) 644–647. [7] S.P. Cadzow, K.L. Armstrong, Rib fractures in infants: red alert! The clinical features, investigations and child protection outcomes, J. Paediatr. Child Health 36 (2000) 322–326. [8] V.F. Garcia, C.S. Gotschall, M.R. Eichelberger, L.M. Bowman, Rib fractures in children: a marker of severe trauma, J. Trauma 30 (1990) 695–700. [9] K.A. Barsness, E.S. Cha, D.D. Bensard, C.M. Calkins, D.A. Partrick, F.M. Karrer, J.D. Strain, The positive predictive value of rib fractures as an indicator of nonaccidental trauma in children, J. Trauma 54 (2003) 1107–1110. [10] P. Betz, E. Liebhardt, Rib fractures in children—resuscitation or child abuse? Int. J. Legal Med. 106 (1994) 215–218. [11] M.R. Spevak, P.K. Kleinman, P.L. Belanger, C. Primack, J.M. Richmond, Cardiopulmonary resuscitation and rib fractures in infants. A post-mortem radiologic– pathologic study, JAMA 272 (1994) 617–618. [12] S. Maguire, M. Mann, N. John, B. Ellaway, J.R. Sibert, A.M. Kemp, Welsh Child Protection Systematic Review Group, Does cardiopulmonary resuscitation cause rib fractures in children? A systematic review, Child Abuse Negl. 30 (2006) 739– 751. [13] H. Klotzbach, G. Delling, E. Richter, J.P. Sperhake, K. Pu¨schel, Post-mortem diagnosis and age estimation of infants’ fractures, Int. J. Legal Med. 117 (2003) 82–89. [14] D. Dolinak, Rib fractures in infants due to cardiopulmonary resuscitation efforts, Am. J. Forensic Med. Pathol. 28 (2007) 107–110. [15] P.K. Kleinman, S.C. Marks, V.I. Adams, B.D. Blackbourne, Factors affecting visualization of posterior rib fractures in abused infants, AJR Am. J. Roentgenol. 150 (1988) 635–638. [16] P.K. Kleinman, S.C. Marks, K. Nimkin, S.M. Rayder, S.C. Kessler, Rib fractures in 31 abused infants: postmortem radiologic-histopathologic study, Radiology 200 (1996) 807–810. [17] C. Cattaneo, E. Marinelli, A. Di Giancamillo, M. Di Giancamillo, O. Travetti, L. Vigano, P. Poppa, D. Porta, A. Gentilomo, M. Grandi, Sensitivity of autopsy and radiological examination in detecting bone fractures in an animal model: implications for the assessment of fatal child physical abuse, Forensic Sci. Int. 164 (2006) 131–137. [18] P.K. Kleinman, A.E. Schlesinger, Mechanical factors associated with posterior rib fractures: laboratory and case studies, Pediatr. Radiol. 27 (1997) 87–91. [19] P.K. Kleinman, S.C. Marks, M.R. Spevak, J.M. Richmond, Fractures of the rib head in abused infants, Radiology 185 (1992) 119–123.
Contents lists available at ScienceDirect
Forensic Science International journal homepage: www.elsevier.com/locate/forsciint
Rib fractures identified at post-mortem examination in sudden unexpected deaths in infancy (SUDI) Martin A. Weber a, R. Anthony Risdon a, Amaka C. Offiah b, Marian Malone a, Neil J. Sebire a,* a b
Department of Paediatric Pathology, Great Ormond Street Hospital for Children and UCL Institute of Child Health, Great Ormond Street, London WC1N 3JH, UK Department of Radiology, Great Ormond Street Hospital for Children, Great Ormond Street, London, UK
A R T I C L E I N F O
A B S T R A C T
Article history: Received 27 August 2008 Received in revised form 22 January 2009 Accepted 14 April 2009 Available online 23 May 2009
Rib fractures may be associated with non-accidental injury (NAI) in infancy, but the possible significance of fresh rib fractures in relation to resuscitation remains undetermined. Consequently, it is important to detect and confirm the presence of rib fractures when performing a post-mortem examination, particularly in the context of sudden unexpected death in infancy (SUDI). At our centre, it has been local policy to perform routine radiological skeletal surveys and detailed post-mortem examination of the ribs in all SUDI autopsies. The aim of this study is to establish the characteristics of all rib fractures identified at autopsy in the setting of SUDI from a large series of cases examined at a single specialist centre. As part of a larger review of paediatric post-mortem examinations performed at a single specialist institution over a 10-year period (1996–2005), all cases presenting as SUDI (aged 7–365 days) were identified and their anonymised records searched to identify all cases in which rib fractures were recorded. Over the 10-year period, 546 post-mortem examinations were performed for the indication of SUDI, including 94 forensic autopsies. Rib fractures were identified in 24 cases (4%). 15 infants (3% of SUDI) demonstrated healing rib fractures, of which 10 infants (67%) showed additional features suggestive of NAI. The other 9 infants (2% of SUDI) demonstrated fresh rib fractures only with no surrounding tissue reaction histologically; in 7 (78%) of these there were no other injuries and the fresh fractures were interpreted to have been caused by resuscitation-related trauma. All of the resuscitation-related fractures were situated in the anterolateral chest, in contrast to NAI-associated fractures, which were located in the anterolateral and/or posterior chest. Anterior costochondral junction fractures were also seen in a minority of NAI-associated cases, but such fractures were not seen in apparent resuscitationrelated cases. Compared to healing rib fractures, which were detected on skeletal survey in 93%, fresh rib fractures were only detected in 22% of skeletal surveys. Rib fractures are uncommon in infancy and may indicate NAI, particularly when healed or healing, posterior or involving the costochondral junction. Fresh rib fractures may be missed on skeletal survey, but can be reliably detected at post-mortem examination following stripping of the pleura and detailed examination of each rib. Fresh anterolateral fractures, which may be multiple, contiguous and even bilateral, are highly likely to be related to resuscitation if there are no other associated injuries. ß 2009 Elsevier Ireland Ltd. All rights reserved.
Keywords: Rib fracture Infancy Sudden unexpected death Resuscitation Non-accidental injury
1. Introduction The presence of rib fractures is well-recognised as a component of non-accidental injury (NAI) in infancy. Classically, in this setting, such fractures are often multiple, of varying ages and usually associated with other features of abuse or trauma [1]. However, the clinical significance of fresh rib fractures in relation to terminal cardiopulmonary resuscitation (CPR) remains undetermined, even though incorrect interpretation of such fractures may have major
* Corresponding author. Tel.: +44 020 7829 8663; fax: +44 020 7829 7875. E-mail address: [email protected] (N.J. Sebire). 0379-0738/$ – see front matter ß 2009 Elsevier Ireland Ltd. All rights reserved. doi:10.1016/j.forsciint.2009.04.015
medicolegal implications. In the United Kingdom, it is recommended that all autopsies of sudden unexpected deaths in infancy (SUDI) are performed on behalf of Her Majesty’s Coroner (HMC) by specialist paediatric pathologists, unless there are significant concerns about possible NAI, in which case it is recommended that the autopsy is carried out in conjunction with a Home Officeaccredited, forensic pathologist [2]. All SUDI post-mortem examinations are carried out according to a standard, published protocol [2], which includes a radiological skeletal survey reported by a paediatric radiologist prior to the autopsy to detect bony injuries which may be associated with NAI. The detection and confirmation of rib fractures is also an integral part of the postmortem examination, both by macroscopic inspection of the ribs
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
76
and by means of histopathological examination of apparent fractures for definitive confirmation and dating of such fractures. However, the prevalence of such fractures in infancy, and in particular, the frequency of probable resuscitation-related fractures, remains uncertain. The aims of this study are, therefore, to establish the characteristics of all rib fractures identified at autopsy in the setting of SUDI from a large series of cases examined at a single specialist centre, and, specifically, to differentiate between fresh and healing rib fractures in relation to other findings at autopsy.
2. Materials and methods Great Ormond Street Hospital for Children, London, is a tertiary specialist hospital for the investigation and treatment of a range of paediatric diseases; the pathology department performs autopsies on the behalf of HMC from a wide geographical area, currently carrying out >300 paediatric post-mortem examinations per annum. As part of a large systematic review of paediatric post-mortem examinations performed during a 10-year period, from 1996 to 2005 inclusive, all autopsies carried out for the indication of SUDI (aged 7–365 days) were identified and their anonymised records searched to identify cases in which rib fractures were present. In most cases in this series, a radiological skeletal survey was carried out prior to the post-mortem examination, whilst in a small proportion of infants, usually due to a period of post-resuscitation survival, antemortem radiographs were available for radiological review. Radiological skeletal surveys comprised a panel of plain radiographs of the entire skeleton, including multiple images of skull, limbs and thoracic cage. Radiographs were reported by specialist paediatric radiologists with expertise in NAI. All post-mortem examinations were carried out by specialist paediatric pathologists, in addition to Home Office-accredited forensic pathologists for forensic autopsies. The ribs were routinely examined during the post-mortem examination as part of the standard autopsy protocol, which included stripping of the parietal pleura from the thoracic cage and careful macroscopic inspection of individual ribs, as well as microscopic examination of any suspected rib fractures. Further specific details were extracted from the database for infants in whom rib fractures were identified, including whether the fractures were identified radiologically prior to the autopsy, the number and distribution of the fractures, comments on radiological and histopathological dating of the fractures, as well as
the presence of other associated abnormalities suggestive of NAI detected at postmortem examination. For the purposes of this study, NAI was defined as the presence of demonstrable injuries not explained by the clinical history, the suspicion having been raised either at clinical presentation or by the pathologist at the time of the post-mortem examination, and was not based on the outcomes of medicolegal or criminal proceedings. Deaths in which the injuries were deemed insufficiently severe to account for the sudden death of the infant were classified as ‘undetermined.’ Rib fractures were either classified as ‘fresh’ or ‘healing,’ the former defined as the absence of a vital reaction (with or without some surrounding haemorrhage), indicating that such fractures occurred in the perimortem or immediate post-mortem period, whilst ‘healing’ fractures showed radiological, macroscopic and/or histological evidence of inflammation, granulation tissue formation, fibrosis and/or callus (or osteoid) formation, indicating that the fracture occurred some time prior to death. The study was approved by the Local Research Ethics Committee.
3. Results During the 10-year study period, a total of 1516 paediatric autopsies were carried out in the Department of Paediatric Pathology, including 546 post-mortem examinations performed for the indication of SUDI. These included 452 infant deaths in which there was no preceding clinical suspicion of NAI and 94 forensic (‘special’) autopsies jointly carried out by forensic and paediatric pathologists. The majority of infants presenting with sudden collapse underwent varying degrees of CPR, including chest compressions, by carers, ambulance crew and/or medical staff, but specific details of the resuscitative attempts with regards to personnel, duration and vigour were not recorded. Overall, rib fractures were present in 25 (5%) of all SUDI (Table 1). One case (Case No. 25, Table 1) was a co-sleeping associated sudden unexpected death of an 18-day-old neonate which occurred abroad. The initial autopsy was performed in that country, and a second post-mortem examination was subsequently carried out at our centre on the embalmed body of the
Table 1 Table of all rib fractures. Rib fractures described as ‘anterior’ (excluding CCJ fractures) or ‘lateral,’ and fractures located in the ‘midclavicular line’ or in the ‘mid-axillary line,’ are categorised as ‘anterolateral’ fractures. Specific details, including dating estimates, of the fractures are based on radiological or histological appearances, as reported at the time of the post-mortem examination. In the small number of cases in which there were discrepancies between the radiological and macroscopic findings relating to the site and/or number of rib fracture(s) or the estimated fracture age(s), the findings in the table are based on the macroscopic examination unless indicated otherwise. (*Based on antemortem radiographs, usually due to a period of post-resuscitation survival in hospital following the initial collapse; **detected radiologically only and not commented on in the macroscopic description or on histological examination; CCJ = costochondral junction). Case number
Cause of death
Age in months
Rib fractures Detected on skeletal survey
Site(s)
Specific details (estimated age of fractures, if given)
Healing rib fractures 1 Head injury (NAI)
2.7
Yes
Anterolateral
Left anterolateral 5, 6, 7 (2–5 weeks)
2
Undetermined
2.2
Yes
Posterior
Left posterior 6, 7, 8 (2–3 weeks)
None
3
Undetermined
5.5
Yes
Anterolateral, posterior
Left anterolateral 7, 8
Multiple fractures of thoracic vertebrae Fracture upper left tibia with early healing reaction Retinal haemorrhage
Left posterior 5, 6, 7 Right anterolateral 7, 8 (‘varying ages—some quite old’)
Features suggestive of NAI other than rib fractures
Head injury
4
Head injury (accidental)
7.3
Yes*
Posterior
Left posterior 10 (2–5 weeks)
None
5
Head injury (NAI)
7.8
Yes
CCJ, anterolateral, posterior
Left anterior (CCJ) 1, 3, 4, 5, 6, 7, 8 (1 week for most fractures) Left posterior 7, 10 (3–5 weeks) Right anterior (CCJ) 1, 4, 5, 6, 7 (1 week for most fractures) Right anterolateral 7 (10–12 weeks)
Head injury, facial bruising
6
Undetermined
7.4
Yes
Anterolateral
Right anterolateral 6, 7, 8 (3–6 weeks)
None
7
Head injury (NAI)
4.9
Yes
CCJ, anterolateral, posterior
Left anterolateral 4, 5, 6, 7, 8 Left posterior 7, 9, 10 Right anterior (CCJ) 5, 6, 7, 8 Right posterior 6, 7, 8 (all around 2–4 weeks)
Head injury
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
77
Table 1 (Continued ) Case number
8
Cause of death
Head injury (NAI)
Age in months
Rib fractures Detected on skeletal survey
Site(s)
Specific details (estimated age of fractures, if given)
2.9
Yes
Anterolateral, posterior
Left anterolateral 7 with recent re-fracture Left posterior 11, 12 Right posterior 10 with recent re-fracture (all about 4-6 weeks, re-fractures more recent)
Features suggestive of NAI other than rib fractures
Head injury Healing metaphyseal fracture left radius Healing metaphyseal fracture right tibia Recent metaphyseal fracture left tibia Torn frenulum Cigarette burns
9
Undetermined
3.6
Yes
Anterolateral
Left anterolateral 4, 5, 6
Healing tibial fracture Bruise of forehead
10
Undetermined
5.6
Yes
Posterior
Right posterior 9 (4–6 weeks)
None
11
Head injury (NAI)
2.8
Yes
CCJ**, posterior
Left anterior (CCJ) 12** Left posterior 7, 9, 10, 11 (2) (4–6 weeks for most fractures) Right anterior (CCJ) 11, 12** Right posterior 9, 10 (4–6 weeks)
Head injury
12
Inflicted asphyxia (NAI)
1.4
Yes
CCJ, anterolateral, posterior
Left posterior 6, 7, 8 (1–2 weeks)
None
Right anterior (CCJ) 10 (1–2 weeks) Right anterolateral 3,4 (fresh) Right anterolateral 5, 6, 7 (fresh re-fractures of previous fractures 1–2 weeks old) Right posterior 10, 11, 12 (1–2 weeks) 13
Traumatic liver rupture (NAI)
9.9
Yes
Posterior
Left posterior 10, 11 (8 weeks)
Traumatic liver rupture
Right posterior 7, 8 (‘recent’) Right posterior 9 (‘several weeks’ with fresh re-fracture)
Skull fracture Widespread bruising
14
Head injury (NAI)
1.4
No*
Posterior
Left posterior 2, 3, 4, 5, 6, 7 (2–3 weeks)
Head injury
15
Undetermined
4.8
Yes
Posterior
Left posterior 7, 9
Healing microfractures of both radii and ulnae Gum scar
Right posterior 4, 5, 6, 10, 11 (varying ages, 4–6 weeks to >2 months) Fresh rib fractures 16 Inflicted asphyxia (NAI)
<1
Yes
Anterolateral, posterior
Left anterolateral 1, 2, 3, 4, 5, 6, 7, 8 (fresh)
(Second autopsy): extensive subdural and subarachnoid haemorrhage around spinal cord
Right anterolateral 2, 3, 4, 5, 6, 7, 8 (fresh) Right posterior 1, 2 , 3, 4, 5, 6, 7 (fresh) 17
Head injury (NAI)
4.5
No
Anterolateral
Right anterolateral 5, 7 (fresh)
Head injury
18
Undetermined
<1
No
Anterolateral
Left anterolateral 3, 4, 5, 6 (fresh)
None
19
Undetermined
2.5
No
Anterolateral
Left anterolateral 2, 3, 4, 5 (fresh)
None
20
Undetermined
1.4
No
Anterolateral
Right anterolateral 4, 5, 6 (fresh)
None
21
Undetermined
2.6
No
Anterolateral
Left anterolateral 4, 5, 6 (fresh)
None
22
Undetermined
2.0
No
Anterolateral
Right anterolateral 4, 5 (fresh)
None
23
Undetermined
2.0
Yes
Anterolateral
Left anterolateral 3, 4, 5, 6 (fresh) Right anterolateral 3, 4, 5, 6 (fresh)
None
24
Undetermined
2.3
No
Anterolateral
Left anterolateral 3, 4, 5, 6 (fresh)
None
25
Undetermined
<1
No
Anterolateral
Left anterolateral 1, 2, 3, 4, 5, 6, 7 (fresh)
(Second autopsy; embalmed, rib fractures interpreted to be post-mortem): none
infant. The second autopsy revealed seven fresh rib fractures, but these were interpreted to have occurred post-mortem, and the case was thus excluded from further analysis. Of the remaining 24 infants, there were healing fractures in 15 (3% of all SUDI; 63% of those with rib fractures; Table 2 and Fig. 1), and of these, 10 (67% of all healing fractures) were associated with additional features at post-mortem examination suggestive
of NAI. In the other nine infants (2% of all SUDI; 37% of those with rib fractures), only fresh rib fractures were identified, with no histological evidence of a surrounding tissue reaction (Table 2 and Fig. 2). In seven of these (<2% of all SUDI; 29% of those with rib fractures; 78% of those with fresh rib fractures), there were no other injuries and no additional features, either clinically or at autopsy, suggestive of NAI or major trauma, suggesting that
78
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
Table 2 Summarised overview of the 24 cases with rib fractures included in the analysis, categorised into cases with healing and exclusively fresh rib fractures, the latter subdivided into fresh rib fractures associated with NAI and presumed resuscitation-related fractures (CCJ = costochondral junction).
these fractures were most likely caused by resuscitationassociated trauma. These apparent resuscitation-related fresh rib fractures were multiple, all involving the (left or right) fourth and fifth ribs, and almost exclusively limited to the third, fourth,
fifth and/or sixth ribs, and were bilateral in one case (Fig. 3). Importantly, all such resuscitation-related fractures were located anterolaterally, compared to the predominantly posterior rib fractures in infants in which other features of NAI were present
Fig. 1. Healing rib fractures. (A) Chest radiograph (faxitron) showing healing rib fractures with callus formation. (B) Macroscopic appearances of healing rib fractures in an infant with suspected NAI. (C and D) Histological features of healing rib fractures, with subperiosteal new bone formation. In (D), a recent re-fracture has occurred (white arrow).
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
79
Fig. 2. Fresh rib fractures. (A) Macroscopic appearances. (B) (H&E), (C and D) (Masson Trichrome). Photomicrographs demonstrating the fracture line (arrows) associated with minor fresh haemorrhage only (higher power in D). There is no surrounding inflammatory reaction or periosteal reaction.
(Fig. 4). Furthermore, none of the apparent resuscitation-related fractures involved the costochondral junction; such fractures, which were described in four infants, were only found in association with other healing fractures and, in three cases, with other injuries suggestive of NAI. The presence of healing rib fractures was identified prior to post-mortem examination on routine post-mortem radiological skeletal survey in 14 (93%) cases, the one case in which the presence of healing rib fractures was not detected prior to autopsy being an infant who, owing to a period of post-resuscitation survival, had an antemortem chest radiograph, which, even on review, failed to disclose the presence of posterior healing rib fractures. In contrast, of the nine cases with fresh rib fractures, the presence of fractures was detected on post-mortem radiological skeletal survey prior to the autopsy in only two (22%) cases. These infants all demonstrated linear, non-displaced fractures with no or minimal surrounding haemorrhage, the fractures being identifiable only on careful inspection of the ribs following stripping of the pleura.
4. Discussion The findings of this study demonstrate that rib fractures are present in up to 5% of cases referred for post-mortem examination for the indication of SUDI. This may be an overestimate of the true prevalence in an unselected population of SUDI, since Great Ormond Street Hospital is a tertiary referral centre and may preselect cases in which there is a preceding history or clinical suspicion of NAI or trauma. Nevertheless, the findings of this study have further demonstrated that in about two thirds of infants in whom rib fractures are present, such fractures show features of healing, whereas in about one third, only fresh rib fractures are present. In the cases in which only fresh rib fractures are identified, the majority are not associated with other injuries or additional features to suggest NAI and are therefore likely to represent resuscitation-related rib fractures; such fractures are seen in 1–2% of otherwise uncomplicated SUDI. These fresh, linear fractures are present anterolaterally in the chest wall and affect contiguous ribs, whereas in infants with other features suggestive of NAI, rib
80
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
Fig. 3. Schematic representation of the distribution of the fresh, presumed resuscitation-related rib fractures. These fractures are located in the anterolateral chest only. Each circle represents one fresh fracture, and each colour represents one case. The image is not drawn to scale, and the circles do not indicate the exact location of specific rib fractures.
fractures are more likely to be healing, posterior or, when anterior, affect the costochondral junction. Due to the limited mobility of infants, accidental trauma, and therefore rib fractures, are uncommon in infancy but may result from traumatic forceful squeezing or compression of the chest. Their presence in life is generally considered to be a strong indicator of possible NAI, particularly when such fractures are healed or healing, multiple and/or of different ages. Several causes of rib fractures in children have previously been described in studies reporting on series of children attending emergency departments. These include NAI, major trauma (such as motor vehicle accidents or MVAs) and rare cases associated with birth injury, neonatal physiotherapy for bronchiolitis or pneumonia, and underlying conditions such as rickets/osteoporosis or osteogenesis imperfecta [3–6]. Pooled data from five series reporting on cohorts of infants with rib fractures indicate that 62% (87 of 140) are associated with NAI [3–5,7,8], and it has been suggested that, due to their rarity in the overall population in this age group, the
positive predictive value of rib fractures for NAI is >95% in children below the age of 3 years [9]. Rib fractures associated with accidental injury, on the other hand, are almost always seen in conjunction with major trauma of comparable magnitude to that associated with MVAs [5]. In one study of 2080 children presenting to a trauma unit, there were 33 (1.6%) children with rib fractures. This group had a mortality rate of 42% since the children with rib fractures were significantly more severely injured, the presence of rib fractures on admission therefore being a marker of severe traumatic injury of whatever cause [8]. Several previous studies have also attempted to specifically determine the frequency of rib fractures following CPR, these being reported but rare, with a prevalence of <1% of witnessed CPR episodes [3,10,11]. A recent systematic review reporting on a possible association of CPR and rib fractures in children identified a total of six studies (with 923 children who had undergone CPR), with a rib fracture prevalence of 3/923 (0.3% or about 1 in 300 episodes); all three infants had anterior fractures [12]. It was also previously noted that resuscitation-related injuries, although rare, are more frequently caused by physicians rather than non-medical persons [10]. The prevalence of apparent resuscitation-related fractures was greater in the present study than reported above (1– 2% vs. 0.3%), almost certainly because our series only comprised infants that had undergone an autopsy, which included detailed examination of the ribs with the specific aim of documenting the presence of such fractures. Fresh rib fractures may not reliably be identified on the radiographic skeletal survey, but such fractures can be detected at post-mortem examination following stripping of the pleura and detailed examination of each rib. It should therefore be routine practice in all SUDI autopsies to specifically examine the thoracic cage for the presence of rib fractures and any possible rib fracture(s) identified should be further investigated histologically in order to confirm their presence and to provide further information regarding possible timing of the fracture(s) relative to the time of death. Histologically, fresh haemorrhage is the earliest finding following rib fracture, soon followed by the presence of a localised inflammatory reaction, with osteoblastic and osteoclastic activity starting as early as two to four days following the initial injury; mineralisation of the newly formed osteoid occurs approximately 1 week later [13]. Fresh rib fractures will not demonstrate any histological evidence of an inflammatory response but may be associated with a small amount of local haemorrhage, although resuscitation-related rib fractures, in cases
Fig. 4. Schematic representation of the distribution of all rib fractures associated with other injuries suggestive of NAI. Each circle represents one fresh fracture, and each triangle a healing fracture. Each colour represents one case. The image is not drawn to scale, the circles and triangles do not indicate the exact location of specific rib fractures, and the triangles do not distinguish between fractures of different healing ages. Such NAI-associated fractures include fresh and healing fractures, posterior fractures and anterior fractures, as well as anterior fractures affecting the costochondral junction.
M.A. Weber et al. / Forensic Science International 189 (2009) 75–81
in which CPR proved unsuccessful, would be expected to be associated with only minimal haemorrhage due to the lack of a circulation at the time of fracture. In one previous series of 70 infant autopsies, subtle fractures of the anterolateral aspects of the ribs were apparently identified in around 10% of cases, all of which showed little, if any, associated blood extravasation and might have been missed had the parietal pleura not been stripped [14]. There are a variety of reasons that rib fractures, especially fresh fractures, may be difficult to identify on routine skeletal radiology; these include superimposition of the transverse process over a posterior rib fracture site, a fracture line that crosses obliquely to the beam and non-displacement of fragments due to preservation of the periosteum, and therefore multiple projections may be required [15]. In a previous study of 31 infants who died due to NAI, only 30 of 84 (36%) rib fractures were identified on skeletal survey examination [16]. These findings are also consistent with recent data from an animal model of severe inflicted trauma, in which detection of rib fractures was more frequent at autopsy compared to radiological examination [17]. In the present study, of the 24 rib fractures present in total, 16 (67%) were detected, this apparent improvement in detection rate probably representing a combination of improvements in imaging technologies and increasing operator expertise. Nevertheless, the presence of fresh rib fractures, with absence of a surrounding tissue reaction, may not be identified on routine radiographic skeletal survey, even when performed by specialist paediatric radiologists with an interest in NAI. These cases would only be recognised at autopsy on detailed examination of the ribs but, as the majority of these otherwise ‘undetectable’ fractures are likely to be due to resuscitation-related trauma rather than NAI when located in the anterior chest, their ‘non-detection’ would arguably be of limited clinical significance. In accordance with the findings of the present study, previous data indicate that in infants the site of the rib fracture is important for correct interpretation of its significance. One study reported that fractures predominantly involving the posterior ends of the ribs occurred in 15 of 43 deaths due to severe trauma, and concluded that, in infants without metabolic bone disease and in whom there is no other history of severe trauma, fractures in the dorsal part of the chest wall should be interpreted as a strong indication of physical child abuse [10]. The mechanism for developing such posterior fractures is now well-recognised, with cadaver and animal studies indicating that posterior rib fractures are a consequence of severe anteroposterior chest compression and excessive posterior levering of the ribs, and would require forces of similar magnitude to those generated by a MVA [18,19], although it is recognised that such effects may be modified by additional mechanical factors, such as the time period and surface area over which the force is applied, and responses of the gripping, restraining or impact material. Pathologists must specifically search for rib fractures in all infant autopsies and must ensure that they evaluate the findings within the overall context of the case when interpreting the significance of fresh rib fractures at post-mortem examination. Although fresh anterior fractures may be seen in NAI, the presence
81
of fresh anterior fractures with no other associated injuries is highly likely to be resuscitation-associated and unrelated to the cause of death, but histological confirmation is required in all cases. Acknowledgements This study was supported by a project grant from the Foundation for the Study of Infant Deaths (FSID). The authors would like to thank Detective Constable Lorraine Gronbach and the Graphics Department of the Metropolitan Police for supplying the illustrations of the thorax used in Figs. 3 and 4. References [1] C.J. Hobbs, ABC of child abuse. Fractures, BMJ 298 (1989) 1015–1018. [2] H. Kennedy (Ed.), Sudden unexpected death in infancy. A multi-agency protocol for care and investigation. The report of a working group convened by The Royal College of Pathologists and The Royal College of Paediatrics and Child Health, The Royal College of Pathologists and The Royal College of Paediatrics and Child Health, London, 2004. [3] K.W. Feldman, D.K. Brewer, Child abuse, cardiopulmonary resuscitation, and rib fractures, Pediatrics 73 (1984) 339–342. [4] P. Schweich, G. Fleisher, Rib fractures in children, Pediatr. Emerg. Care 1 (1985) 187–189. [5] B. Bulloch, C.J. Schubert, P.D. Brophy, N. Johnson, M.H. Reed, R.A. Shapiro, Cause and clinical characteristics of rib fractures in infants, Pediatrics 105 (2000) E48. [6] M. Chalumeau, L. Foix-L’Helias, P. Scheinmann, P. Zuani, D. Gendrel, H. Ducou-lePointe, Rib fractures after chest physiotherapy for bronchiolitis or pneumonia in infants, Pediatr. Radiol. 32 (2002) 644–647. [7] S.P. Cadzow, K.L. Armstrong, Rib fractures in infants: red alert! The clinical features, investigations and child protection outcomes, J. Paediatr. Child Health 36 (2000) 322–326. [8] V.F. Garcia, C.S. Gotschall, M.R. Eichelberger, L.M. Bowman, Rib fractures in children: a marker of severe trauma, J. Trauma 30 (1990) 695–700. [9] K.A. Barsness, E.S. Cha, D.D. Bensard, C.M. Calkins, D.A. Partrick, F.M. Karrer, J.D. Strain, The positive predictive value of rib fractures as an indicator of nonaccidental trauma in children, J. Trauma 54 (2003) 1107–1110. [10] P. Betz, E. Liebhardt, Rib fractures in children—resuscitation or child abuse? Int. J. Legal Med. 106 (1994) 215–218. [11] M.R. Spevak, P.K. Kleinman, P.L. Belanger, C. Primack, J.M. Richmond, Cardiopulmonary resuscitation and rib fractures in infants. A post-mortem radiologic– pathologic study, JAMA 272 (1994) 617–618. [12] S. Maguire, M. Mann, N. John, B. Ellaway, J.R. Sibert, A.M. Kemp, Welsh Child Protection Systematic Review Group, Does cardiopulmonary resuscitation cause rib fractures in children? A systematic review, Child Abuse Negl. 30 (2006) 739– 751. [13] H. Klotzbach, G. Delling, E. Richter, J.P. Sperhake, K. Pu¨schel, Post-mortem diagnosis and age estimation of infants’ fractures, Int. J. Legal Med. 117 (2003) 82–89. [14] D. Dolinak, Rib fractures in infants due to cardiopulmonary resuscitation efforts, Am. J. Forensic Med. Pathol. 28 (2007) 107–110. [15] P.K. Kleinman, S.C. Marks, V.I. Adams, B.D. Blackbourne, Factors affecting visualization of posterior rib fractures in abused infants, AJR Am. J. Roentgenol. 150 (1988) 635–638. [16] P.K. Kleinman, S.C. Marks, K. Nimkin, S.M. Rayder, S.C. Kessler, Rib fractures in 31 abused infants: postmortem radiologic-histopathologic study, Radiology 200 (1996) 807–810. [17] C. Cattaneo, E. Marinelli, A. Di Giancamillo, M. Di Giancamillo, O. Travetti, L. Vigano, P. Poppa, D. Porta, A. Gentilomo, M. Grandi, Sensitivity of autopsy and radiological examination in detecting bone fractures in an animal model: implications for the assessment of fatal child physical abuse, Forensic Sci. Int. 164 (2006) 131–137. [18] P.K. Kleinman, A.E. Schlesinger, Mechanical factors associated with posterior rib fractures: laboratory and case studies, Pediatr. Radiol. 27 (1997) 87–91. [19] P.K. Kleinman, S.C. Marks, M.R. Spevak, J.M. Richmond, Fractures of the rib head in abused infants, Radiology 185 (1992) 119–123.