Overview of available evidence

One systematic review was identified.

Narrative summary

The question of when bruises in children are diagnostic or suggestive of abuse was investigated in a narrative systematic review¹⁴ that included 23 studies. Owing to a lack of comparative studies (only two studies were comparative), the authors undertook a comparison by using nine studies that addressed bruising in non-abused children (two case–control studies, four cross-sectional studies and three case series) and 16 studies that addressed bruising in abused children (two case–control studies, one cross-sectional study and 13 case series).

Apart from the age and developmental stage of the child, the location and pattern of bruising was found to be important for distinguishing between accidental and non-accidental bruising.

The conclusions of this paper were that the following patterns of bruising are suggestive of physical child abuse:

• bruises in children who are not independently mobile
• bruising in babies
• bruises to the face (with the exception of the forehead), back, abdomen, arms, buttocks, ears and hands
• bruises that are seen away from bony prominences
• multiple bruises in clusters
• multiple bruises of uniform shape
• bruises that carry the clear imprint of the implement used or a ligature.

The authors emphasise that the interpretation of bruising always needs to take the context of medical and social history, the developmental stage, the explanation given and other available information into account. [EL = 2+]

An update to the above review included a paper that investigated whether petechiae are six times more likely to be seen in physical abuse than non-abuse in children. There was no difference in the distribution of petechiae in the two groups.¹⁵

Evidence statement

A systematic review has summarised findings from studies on bruising.

GDG considerations

The GDG supports the conclusions of the systematic review but notes that it is important to exclude bruises from everyday activity, accidental injury, meningococcal septicaemia and other blood disorders that may appear as signs of bruising before suspecting child maltreatment. Drawing on its clinical experience, the GDG suggests that inflicted bruising can occur on more than one plane of the body, for example both sides of the face, as well as in clusters. The GDG believes that the age of a bruise cannot be judged reliably from interpretation of the colour of a bruise and should not be used in the assessment of bruises. The developmental stage of the child, however, is a reasonable indicator for suspicion, in that if a child is unable to move independently, bruising is unlikely to be accidental unless there is good history of an accident. The GDG also believes that bruises of uniform appearance (bruises with very similar or identical appearances) imply that they may have been caused in the same way on more than one occasion by the same mechanism and as such are unlikely to be accidental.

There was no evidence identified regarding love bites. Bruising from ‘love bites’ may be identified as oval-shaped lesions with a bruised or petechial appearance. The GDG believes that love bites should be interpreted in a similar way to other bruises. An assessment of the age of the child or distribution (for example, over the breast area) may suggest child sexual abuse (CSA).

There was consensus within the GDG about the recommendations in this section. The Delphi panel's views were sought in relation to love bites (see Statement 2a in Section 4.1.2 on bites).

Overview of available evidence

One systematic review was identified.

Narrative summary

A systematic review of abusive bite marks in children (end search date June 2007) identified five case studies where bites had been inflicted.^16,17 Four of the children were younger than 30 months and one was in her teens. The perpetrator was a child in one case. [EL = 2+] No suitable published literature was found that links animal bites to maltreatment.

Evidence statement

The literature on abusive bite marks in children is sparse, with only five reported incidents of abusive bite marks.

Delphi consensus (see also Appendix C)

The GDG sought the opinions of the Delphi panel on bite marks. The statements below were drafted.

GDG considerations

The evidence base in this area is weak and thus the GDG made consensus-based statements and sought the opinions of the Delphi panel on this topic (see above and Section C.2.1).

It can be difficult for healthcare professionals to ascertain the provenance of a bite mark, whether from an adult, an older child or a young child. The GDG acknowledges that bites can be caused by young children in their play activities and that older children can inflict abusive bite marks. Once it seems unlikely that a bite mark was caused by a young child, the GDG concludes that inflicted injury has occurred and maltreatment should be strongly suspected.

Animal bites can occur when a child has not been adequately supervised and, if there is evidence of a lack of supervision, the GDG believes that healthcare professionals should consider neglect.

The GDG accepted Statements 1a and 4b from the Delphi survey but amended Statement 1a because of the difficulty among frontline healthcare professionals of ascertaining who has inflicted a bite based on appearance alone.

Overview of available evidence

No suitable published literature was identified that documented associations between cuts and abrasions and child maltreatment.

GDG considerations

The GDG found no suitable published literature on the question of when cuts, abrasions, scars and scratches are reasons to suspect child maltreatment. The GDG consensus is that, similar to other soft-tissue injuries, a healthcare professional should consider the site, pattern, distribution, characteristics, presentation and explanation of the injuries in order to decide whether to suspect maltreatment. The GDG recognises that these presentations can be consistent with deliberate self-harm (see Section 7.2.1 on self-harm).

There was consensus within the GDG about the recommendations in this section and thus the views of the Delphi panel were not sought.

GDG considerations

In the absence of a body of evidence, the GDG recognises that strangulation and suffocation are serious injuries. Any clinical signs of suffocation or strangulation should be a cause for serious concern regarding child maltreatment (see Section 4.1.1 on bruises, Section 4.1.3 on lacerations (cuts), abrasions and scars, Section 5.3 on apparent life-threatening events and Section 7.2.1 on self-harm), but the GDG was unable to make a specific recommendation in this section.

Overview of available evidence

The question ‘What patterns of burns in children are seen in physical abuse?’ was investigated separately for scald burns and non-scald burns in two systematic reviews by the same research group.^18,19

Narrative summary

The identification of intentional scald burns in children in contrast to accidentally sustained scalds was investigated in a well-conducted systematic review that included 26 studies, comprising one case–control study, eight cross-sectional studies and 17 case series and case studies.¹⁸

In addition to the usual exclusion criteria such as review papers and personal experiences, the authors excluded scalds that were due to neglect and studies that combined scald and contact burn data.

There was no evidence of a difference in gender, age of the child or the total body surface area affected between intentional and accidental scalds. Other features were grouped according to whether a scald was likely to be intentional based on the evidence level of the studies reporting those features.

The following features indicate that intentional scalds are likely:

• immersion scalds or scalds from hot tap water indicated by:

  –

  the presence of clear upper limits or symmetric scalds on the extremities

  –

  an isolated scald on the buttock or perineum with or without scald injuries on the lower extremities

  –

  isolated scald injuries on the lower extremities

• the child presents with associated unrelated injuries
• the history given is incompatible with examination findings
• there are coexisting fractures or other injuries
• the child is passive, introverted or fearful
• a history of previous abuse or domestic violence
• numerous prior accidental injuries.

The presence of one or more of the following features indicates that intentional scalds should be considered as a possibility:

• the scald is of uniform depth, flexures are spared, the centre of the buttock is spared, or the scald appears like a glove or stocking on one or more limbs
• a previous burn injury
• neglect/faltering growth
• a history inconsistent with assessed development
• historical/social features such as:

  –

  a trigger such as soiling, enuresis, misbehaviour

  –

  differing historical accounts

  –

  a lack of parental/carer concern

  –

  an unrelated adult presenting the child

  –

  the child is known to children's social care.

The strength of evidence for this review was limited by the small number of good-quality studies containing comparative data, the relatively small number of children included, the retrospective design and the lack of consistency between studies that does not allow a formal meta-analysis. [EL = 2+]

A systematic review about non-scald burns consisted of 25 case series or studies.¹⁹ The conclusion of the review was that the history should be taken carefully, the clothing should be examined for suspected caustic burns and the burn should be matched to the potential burn agent. The review was limited through the scarce evidence base and it thus describes a small number of children (84 children in total, of which 59 were abused). There were no comparative studies of cigarette burns and a lack of comparative data for contact burns. [EL = 2+]

GDG considerations

Burn injuries can be inflicted or accidental, and some burn injuries can be due to neglect through lack of supervision. The GDG believes that it is difficult to untangle these issues and therefore the story that accompanies a burn injury should be scrutinised for consistency with the injury.

Despite the low evidence level of the literature reviewed in the published systematic review, the GDG agrees with the recommendations made therein, based on the GDG members' own clinical experience. The GDG also believes that parents/carers may delay seeking medical attention when a burn injury has been intentional.

There was consensus within the GDG about the recommendation in this section and thus the views of the Delphi panel were not sought.

See also Chapter 6 on neglect – failure of provision and failure of supervision.

Overview of available evidence

No suitable published literature was identified that documented associations between cold injury and child maltreatment.

GDG considerations

In the absence of suitable evidence, the GDG suggests that injuries due to the cold such as swollen, red hands or feet where there is no medical cause can be reason to consider child maltreatment in the context of the persistent failure to provide adequate warmth, clothing or shelter over a period of time. Similarly, hypothermia without an adequate explanation in a child should be a reason to consider child maltreatment.

There was consensus within the GDG about the recommendations in this section and thus the views of the Delphi panel were not sought.

Overview of available evidence

No suitable published literature was identified that documented associations between hair loss and child maltreatment.

GDG considerations

Hair loss in children can be caused by hair-pulling or spontaneous hair loss. The GDG identified no literature that suggests spontaneous hair loss occurs secondary to maltreatment. In the GDG's opinion, hair loss caused by inflicted hair-pulling constitutes physical abuse. It is the GDG's experience that children can pull each other's hair while fighting so it is important to establish who has inflicted the hair-pulling. .

The GDG believes that hair loss due to self-inflicted hair-pulling in the absence of a medical cause or other definable stressor may be a sign of emotional distress that could be due to maltreatment (see Section 7.2.1 on self-harm).

The GDG believes that unexplained hair loss is an example of an unusual injury without explanation (see Section 4.1.14 on general injuries). The GDG was unable to make a specific recommendation about hair loss.

Overview of available evidence

One systematic review and five additional studies were included.

Narrative summary

A systematic review (1950 to April 2007) that included 32 comparative studies investigated ‘Which fractures are indicative of abuse?’.²⁰ The authors highlighted concerns about the quality of papers available. The main concerns were:

• considerable heterogeneity between studies
• wide age ranges studied
• variable radiological techniques employed
• wide variation in definitions of abuse used in studies.

The statistical methodology adopted for the meta-analysis acknowledged these concerns. A random effects model was used. This method models heterogeneity by assuming that each study has a probability of abuse associated with it and that these form a probability distribution between studies. This probability distribution was estimated by a Bayesian method, using WinBugs²¹ and a 95% credible interval (CrI) was derived to summarise the probability of abuse.

The review was able to report two general findings:

• fractures from child abuse are most common in children younger than 18 months
• multiple fractures are more suspicious of abuse.

The results for specific locations are outlined below.

GDG considerations

Evidence from one systematic review and five additional studies showed that fractures in children can be indicative of maltreatment. These studies confirmed that children younger than 18 months are at a heightened risk of sustaining a fracture from physical abuse. No one fracture is characteristic of physical abuse. The probability that fractures are due to maltreatment is increased where multiple fractures are present or the child is yet to gain independent mobility. However, the available evidence from observational studies is inherently open to bias and reported confidence intervals are likely to greatly underestimate the true variance. There are very few comparative data on metaphyseal fractures or fractures other than ribs, long bones or skull fractures.

There was consensus within the GDG about the recommendations in this section and thus the views of the Delphi panel were not sought.

Overview of available evidence

The GDG referred, with permission, to work in this area by the Welsh Child Protection Systematic Review Group that is, as yet, unpublished. Skull fractures and bruising to the head from physical abuse is addressed in Section 4.1.8 on fractures and Section 4.1.1 on bruises.

Narrative summary

Two systematic reviews (search end date 2007) were identified that compared features and neuroimaging of abusive head injury with non-abusive head injury in children.²⁹ Studies were included if the child presented to hospital alive and neuroimaging was completed. Fourteen studies were included in the clinical features review, representing 779 abused and 876 non-abused children. Eighteen studies were included in the neuroimaging review. [EL = 2]

Eight studies showed that the age of children with abusive head injury was statistically significantly younger than non-abused children and two studies found no difference. The mean age of abused children was less than 1 year in all studies, and for non-abused children ranged from 4.8 months to 35.5 months. Intracranial injuries considered in the studies were subdural haemorrhage, subarachnoid haemorrhage and traumatic brain injury. The inclusion criteria for the comparison groups varied across studies.

Eight studies recorded whether there was an explanation of trauma and they all noted a statistically significantly greater number of children in the abuse group with no explanation of trauma. Seven studies recorded minor trauma (a fall under 4 feet): of these, three were general head injury studies and showed no difference between groups. Three of the four studies of children with traumatic brain injury or subdural haemorrhage showed that more children in the abuse group gave a history of minor injury and seven studies found that a history of major trauma was reported statistically significantly more often in non-abused compared with abused children. In five studies there were recorded cases of ‘admitted assault’.

GDG considerations

There is a strong evidence base that states that abusive head injury occurs primarily in babies and infants. These children have varied clinical presentation, ranging from non-specific symptoms such as vomiting and irritability to infants who are unconscious. Intracranial injury includes subdural haemorrhages, with or without subarachnoid haemorrhages, which are often small, multiple and widely distributed. Hypoxic ischaemic injury is more commonly associated with abusive head injury than accidental head injury. There is a strong association between intracranial injury and retinal haemorrhages, apnoeic episodes and skeletal fractures. Children with abusive head injury may present with impaired neurology and no external sign of injury.

There was consensus within the GDG about the recommendation in this section and thus the views of the Delphi panel were not sought.

See also Section 4.1.1 on bruises, Section 4.1.3 on lacerations (cuts), abrasions and scars, Section 4.1.5 on thermal injuries and Section 4.1.8 on fractures.

Overview of available evidence

Many papers retrieved on injuries to the eye discuss retinal haemorrhage in the context of head trauma (see Section 4.1.9 on intracranial injuries).

Narrative summary

In one report³⁰ and an update to it,³¹ the Ophthalmology Child Abuse Working Party of the Royal College of Ophthalmologists has considered questions relating to the effects on the eye of shaking or indirect trauma to the head in infants and young children. [EL = 4]

The Working Party concluded that:

• retinal haemorrhages are more likely to be due to non-accidental injury than accidental injury
• unilateral retinal haemorrhages can occur in child abuse
• in children younger than 2 years, retinal haemorrhage is highly unlikely to be caused by rough play or an attempt to arouse an apparently unconscious child
• birth-related retinal haemorrhages are common.

One prospective cohort study (n = 150) of consecutive referrals for craniocerebral traumatic lesions reported data on retinal haemorrhage in 129 children (median age 3.6 months) excluding neonates.³² Fifty-six children were found to have been abused and, of these, 75% had a retinal haemorrhage. Of the 73 children in the accidental trauma group, 7% had a retinal haemorrhage. There was a high level of confirmation of abuse. [EL = 2+]

No papers that met the inclusion criteria were retrieved on subconjunctival haemorrhage.

GDG considerations

The evidence about eye injury in maltreatment is largely confined to retinal haemorrhages which are closely associated with inflicted head trauma The GDG supports this association and is of the opinion that retinal haemorrhages in a young child should alert healthcare professionals to the possibility of inflicted head trauma and should be interpreted in that context (see Section 4.1.9 on intracranial injuries). In the absence of evidence relating other eye injuries to maltreatment, the GDG came to a consensus decision that other injuries to the eye should be assessed in the light of the explanation given. If the explanation is absent or not typical of accidental injury, maltreatment should be suspected.

There was consensus within the GDG about the recommendations in this section and thus the views of the Delphi panel were not sought.

See also Section 4.1.1 on bruises and Section 4.1.3 on lacerations (cuts), abrasions and scars.

Overview of available evidence

A systematic review was identified that set out to characterise the signs and symptoms of abusive spinal injury.

Narrative summary

A systematic review (search dates 1975 to 2006) included 15 studies representing information on 33 children.²⁹ Mortality was high, with 26 of the 33 children fatally injured; two of the seven survivors had quadriplegia. The median age of presentation was 6 months (range 1.2–48 months). Diagnosis was delayed in seven cases as the condition was not suspected. Statements of witnesses and confessions of the perpetrators were recorded. [EL = 2+]

Evidence statement

One systematic review suggested that spinal injury is uncommonly reported in child abuse and that it may easily be missed. More than 50% of cases with cervical trauma were younger than 6 months and had associated significant head injury and retinal bleeds. Given the subtle presentation of cervical injuries, these may be masked by associated symptoms or may remain asymptomatic and go undiagnosed. The thoracolumbar lesions occurred in older infants or toddlers and did not show the same association with abusive brain injury. Here, there were clinical signs (neurological or orthopaedic) yet diagnosis was frequently delayed.

GDG considerations

Vertebrospinal injuries of all causes are rare in children and most are associated with a history of significant trauma such as an MVC or sports injury. Abusive spinal cord injury causes significant morbidity and mortality. The substantiated cases of maltreatment in the literature were where there were confessions of perpetrators or statements of witnesses. Therefore the GDG concludes that the absence of an appropriate explanation should be a cause for concern and thus a reason to suspect maltreatment.

There was consensus within the GDG about the recommendation in this section and thus the views of the Delphi panel were not sought.

Overview of available evidence

There was a paucity of comparative studies and large case series in this area. Two retrospective studies investigating differences between inflicted and non-inflicted injuries^33,34 and one concentrating on abdominal injuries³⁵ were identified. All three studies provided epidemiological information.

Narrative summary

A retrospective review of patients attending a trauma centre (n = 121 younger than 6 years) found 13 children in whom injuries had been inflicted, 77 who had suffered a high-velocity accident and 31 who had suffered a low-velocity accident.³⁵ Children were excluded from the study if they had an associated neurological injury, an abdominal injury secondary to severe thoracic injury, injuries that could not be classified as accidental or inflicted or, in some child abuse cases, where there was a level of denial that trauma had occurred. Despite the small sample, injuries to the hollow viscus were found to be more common in child abuse cases than accidental injury cases. There was no statistically significant difference between the groups in incidence of injury to solid organs. Eight-two percent of accidental injuries were brought to medical attention within 12 hours compared with 46% of inflicted injuries. The median abbreviated injury scale (AIS) score was statistically significantly higher in the inflicted group compared with the high-velocity trauma and low-velocity trauma groups. [EL = 2−]

A review of data from the US-based National Pediatric Trauma Registry selected children younger than 5 years who had been hospitalised over a 10 year period.³³ A diagnosis of child abuse was ascertained at the treating hospital. There were 1997 abuse cases and 16 831 children who had suffered unintentional injury. Thoracic injury was more likely in children who had been maltreated than in those who had not (OR 1.70; 95% CI 1.39 to 2.08). Similarly, abdominal injury was more likely in the maltreated group (OR 2.71; 95% CI 2.23 to 3.29). [EL = 2−]

A follow-up to this study retrieved records from 1997 to 2001.³⁶ There were 927 children younger than 5 years who had suffered blunt abdominal trauma. Of these, 63% were due to MVCs, 16% were due to abuse, 14% were due to a fall and 8% were due to other causes. After excluding MVCs, abuse accounted for 79% of injuries in children younger than 12 months, 61% in children aged 13–24 months, 39% in children aged 25–36 months and 25% in children aged 37–48 months. [EL = 2−]

The fourth study reviewed medical records from a children's hospital over a 9 year period.³⁴ There were 5733 cases of accidental trauma and 453 cases of non-accidental trauma. The incidence of thoracic injury was lower in the accidental trauma group than in the non-accidental trauma group (6.0% of children versus 17.0% of children; P < 0.001). There was no statistically significant difference between the groups in the incidence of abdominal injury (7.6% accidental versus 8.6% non-accidental). [EL = 2−]

Evidence statement

The evidence base suggested that visceral injuries do occur in cases of maltreatment and that, after MVCs are excluded, maltreatment is the most common cause in young children.

GDG considerations

Visceral injuries are found in cases of child maltreatment. Injuries to hollow viscus and delayed presentation were more common in cases of child maltreatment. Visceral injuries can present as acute pancreatitis. The GDG's opinion is that visceral injuries due to child maltreatment can sometimes be missed because of the way they present; there may be no bruises even if the injury was inflicted. The GDG found no reason to specify age categories for the suspicion of maltreatment. As with other abusive injuries, the explanation given for the injury may not be compatible with the child's developmental stage. An abusive visceral injury may be in association with other injuries or in isolation.

There was consensus within the GDG about the recommendation in this section and thus the views of the Delphi panel were not sought.

Overview of available evidence

One systematic review was identified.

Narrative summary

One well-conducted systematic review of the literature identified 19 studies (603 children) that reported oral injuries associated with child maltreatment.³⁷ Twenty-seven abused children had torn labial frena, of whom 22 were younger than 5 years. Two children had non-abusive torn labial frena. The review listed a number of oral injuries that were identified in 580 cases of child abuse: laceration or bruising to the lips, mucosal lacerations, dental trauma, tongue injuries and gingival lesions. The authors presented no comparative data and concluded that oral cavities should be examined in suspected child abuse. [EL = 2+]

Evidence statement

The systematic review indicated that oral injuries can occur in child abuse but that there are no oral injuries that are specific to maltreatment.

GDG considerations

The evidence did not show any means of distinguishing accidental oral injury from intentional injury. The GDG believes that as oral injuries may be inflicted and can be missed, all healthcare professionals who are concerned about maltreatment should inspect the child's mouth. The GDG recommends that, as with all injuries seen in child abuse cases, descriptions that are inconsistent with the injury should raise awareness about the possibility of child maltreatment.

There was consensus within the GDG about the recommendation in this section and thus the views of the Delphi panel were not sought.

Delphi consensus (see also Appendix C)

The small amount of relevant literature on genital and anal symptoms led the GDG to develop a number of statements for consideration by the Delphi panel. The GDG sought their opinions about genital and anal symptoms in general and asked questions about specific symptoms in order to offer better guidance to healthcare professionals.

Genital signs in girls

Erythema: In prepubertal girls, genital erythema has been found in sexual abuse cases (7/20) and non-abused controls (2/195) (separate studies). Proportions of sexually abused pubertal girls with erythema ranged from 13% (n = 204) to 32% (n = 214) in two case series. In one comparison study combining data on prepubertal and pubertal girls, erythema was reported in 34% (n = 119) of the CSA group, in 68% (n = 59) of girls with genital complaints and in 13% (n = 127) of girls undergoing routine examination. Abuse was not rigorously excluded from the comparison groups. The timing of examination after the alleged incident and skin pigmentation influence the finding of erythema.

Oedema: No studies were identified that reported the prevalence of oedema in non-abused girls. Oedema was noted in 19% (n = 214) of pubertal sexually abused girls. The timing of examination after the alleged incident influences the finding of oedema.

Bruising: In one comparative study, bruising was noted in one of 192 girls with a history of vaginal penetration and in none of 200 girls who had not been abused. In the abuse cases, examination took place on average 42 days after the abusive event.

In a case series (n = 43) of prepubertal girls with a history of vaginal penetration, 13 haematomas were found but it was unclear how many girls this involved. No genital bruising was reported in one study of prepubertal girls selected for non-abuse.

In a case series (n = 204) of pubertal girls with a history of penile vaginal penetration, 4% had bruising.

A case series (n = 155) of sexually abused prepubertal and pubertal girls examined within 72 hours of the abusive event reported 3% with genital bruising.

Abrasions: Genital abrasions were reported in one study of healing in sexually abused girls with a history of penile and/or digital vaginal penetration. No genital abrasions were reported in a study of non-abused prepubertal girls (n = 195). Abrasions were reported in 17% (n = 214) of pubertal sexually abused girls. The majority of the cohort reported penile vaginal penetration and had been examined within 72 hours of the incident. In a comparative study of prepubertal and pubertal sexually abused girls, three of 119 girls had abrasions; no abrasions were reported in the genital complaints group (n = 59) or the routine health check group (n = 127). Abrasions have been reported in one study of prepubertal girls with straddle injury. Abuse was not rigorously excluded from this group.

Lacerations: There was inconsistency of definitions of genital lacerations and tears to the hymen across the studies identified by the authors. Hymenal lacerations were reported in 33% (n = 205) of prepubertal sexually abused girls in a case series. The authors reported difficulty in distinguishing small lacerations from notches. Partial hymenal tears were reported in two of 24 girls reporting penile vaginal penetration and four of 19 reporting digital vaginal penetration. In a study of non-abused prepubertal girls, no hymenal lacerations were reported. In two studies of pubertal girls, hymenal lacerations/tears were reported in 3% (n = 204) and 6% (n = 214) where more than 90% of study participants reported penetrative abuse.

Posterior fourchette/fossa tears were reported in 14 of 24 prepubertal sexually abused girls. No genital lacerations were reported in the study of prepubertal non-abused girls (n = 195). Posterior fourchette/fossa tears were reported in 40% of pubertal sexually abused girls examined less than 72 hours after the incident and in 2% of those examined more than 72 hours after the incident (n = 204). In a study of prepubertal and pubertal sexually abused girls, one of 155 girls had a vaginal laceration (poor definitions used in this study).

Healing/healed injuries: Hymenal transection was inconsistently defined in the studies. Hymenal transections were found in some prepubertal girls with a history of penetrative abuse; none were found in non-abused girls. The evidence on the importance of scars in prepubertal girls is inconclusive.

Hymenal bumps/mounds: There was inconsistency of definitions in the identified studies but, overall, hymenal bumps/mounds were found to be a normal variant.

Hymenal width and diameter: No conclusions could be drawn about the importance of hymenal width or diameter as signs of sexual abuse.

Friability of the genital tissues is not specific for sexual abuse in prepubertal girls and there is insufficient literature in pubertal girls.

Labial fusion has been found in both abused and non-abused prepubertal girls. There is insufficient evidence to determine the importance of labial fusion in sexual abuse of pubertal girls.

Vaginal discharge in prepubertal girls was observed more often in girls reporting penile vaginal penetration than those reporting digital penetration or no abuse in a case–control study where presence of a sexually transmitted infection (STI) was used to define abuse. Vaginal discharge was found in 1% to 2% of non-abused prepubertal girls.

Vaginal foreign bodies: No suitable comparative studies were identified that investigated vaginal foreign bodies. No studies of foreign bodies in pubertal or non-abused girls were identified. In prepubertal girls, three studies representing data on 47 girls (age range 2–10 years) with vaginal foreign bodies. Nine girls were defined as victims of CSA according to differing criteria.

Anal signs in girls and boys

No comparative studies of suitable quality were identified that reported on anal/perianal erythema, perianal venous congestion, anal/perianal bruising, anal fissures, lacerations, scars and tags, or reflex anal dilatation.

Anal or perianal erythema was observed in 1% (n = 310) to 10% (n = 189) of CSA cases. The timing of examination in relation to the incident was not stated. In non-abused children, redness was reported in 7% (n = 89) of infants and 11% (n = 276) of 5- to 6-year-olds.

Perianal venous congestion was observed in 8% (n = 50) and 36% (n = 50) of anally abused children; the timing of the examination after the incident ranged from 4 weeks to 6 years. In non-abused children, perianal venous congestion was reported in 1% of infants (n = 89) and 20% of 5- to 6-year-olds (n = 276).

Bruising: In a case series of anally abused children, bruising was observed in 10% (n = 50); the timing of examination after the incident was not reported. In another study, 1% of sexually abused children (n = 190) examined within 72 hours had anal/perianal bruising. There were no reports of bruising in non-abused children (n = 305).

Anal lacerations/tears defined as acute tears in the anus and tissues immediately surrounding it were not found in a study (n = 305) where abuse was excluded. Lacerations/tears were found in between 1% and 18% of sexually abused children (based on six case series).

Anal fissures were found in one child in a study of non-abused children (n = 89). In a study of abused children, 25 of 50 anally abused children had anal fissures, fissures were present in 7% of sexually abused children who denied anal abuse (n = 83) and 3% of children with no allegation of sexual abuse (n = 81).

Anal scars were not found in children selected for non-abuse (n = 305). In anally abused children, scars were found in 38% and 84% of children (n = 50) in two studies. In sexually abused children, anal scars were found in between 1% and 4%.

Anal tags were reported in between 3% and 7% of children selected for non-abuse (two studies) and between 4% and 32% of anally abused children (two studies) where the majority of tags were found away from the midline. In sexually abused children, tags were found in between 3% and 7%.

Reflex anal dilatation: In children selected for non-abuse, reflex anal dilatation has been reported in less than 1% of children examined in the left lateral position and in 5% of those examined in the knee-chest position. It was observed in 10% and 34% (two studies, each n = 50) of anally abused children and in 5% of sexually abused children.

Genital signs in boys

Genital injuries in boys following sexual abuse have not been well reported. Four case series of sexual abuse in boys have reported injury to the external male genitalia as a result of sexual abuse in between 0% and 7% of abuse cases. Genital injuries due to sexual abuse occur mostly to the penis. Testicular or scrotal injuries are more commonly due to accidents than abuse (based on one study where confirmation of abuse was unclear).

Evidence statement

The thorough review of the literature on physical signs of sexual abuse⁶ highlights important issues for the use of physical signs in suspecting abuse. The evidence base is lacking in both quality and quantity, in part due to difficulties in conducting research in this area. Observable signs are relatively uncommon and this could be because of the timing of the examination relative to the abuse.

GDG considerations

Among the various signs presented in the systematic review, few are commonly observed and, of those, many will only be seen on examination following a disclosure or report. In the context of this guidance and its intended audience, the GDG believes that the history that the child or parent/carer provides will be of the utmost importance. Therefore the GDG believes that genital or anal symptoms and their context are more likely to become apparent as features of maltreatment in a routine clinical situation than genital or anal signs. The GDG has considered signs and symptoms outside of the remit of the RCPCH guidance, as the RCPCH did not consider an exhaustive list.

The GDG acknowledges that it is common for newborns to have vaginal discharge and sometimes bleeding, especially if they are breastfed.

There are no studies reporting the prevalence of anal fissures in constipation or the passing of hard stools but the GDG's clinical experience suggests that these, together with Crohn's disease, should be excluded before suspecting anal abuse.

After discussion, the GDG decided that reflex anal dilatation is a sign that would be sought during the full assessment of a child where child abuse was suspected. Such an assessment would be conducted by a professional with expertise in the field and the GDG therefore concluded that reflex anal dilatation is not a sign that a frontline healthcare professional would be expected to recognise. However, a healthcare professional may come across a child with a gaping or dilated anus. In the absence of an obvious medical condition to explain this finding, such as a neurological disorder or severe constipation, the GDG believes that they should consider child maltreatment and seek advice from a more experienced colleague.

The GDG sought the opinions of the Delphi panel on statements about genital and anal symptoms (see above and Section C.2.6). 5a–14a, 15b, 19a and 20a were adopted for use in the recommendations. There was consensus within the GDG about the recommendations on genital and anal signs and thus the views of the Delphi panel were not sought.

Overview of available evidence

A systematic review for physical signs of CSA builds the evidence base for STIs.⁶ The chapter on STIs is treated as one systematic review for the purposes of this document.

Narrative summary

In a systematic review of some of the most frequent STIs that have been noted in CSA cases, 84 studies were reviewed.⁶ Conclusions were drawn from prevalence figures of:

• sexual abuse in children with the STI, and
• prevalence figures of the STI in sexually abused children. [EL = 2+]

None of the literature was able to establish the age at which mother-to-child (vertical) transmission can be excluded.

Delphi consensus (see also Appendix C)

The GDG sought the opinions of the Delphi panel on the circumstances under which an STI in a young person aged 13 years or over is a reason to suspect sexual abuse. They did not seek validation on the list of STIs that should prompt a concern. The following statements were drafted:

GDG considerations

It is the GDG's opinion that an STI in children as a direct result of sexual abuse falls within the legal framework outlined in the Sexual Offences Act 2003 (see below). Therefore, an STI in a child younger than 13 years should raise the suspicion of sexual abuse. The GDG was unable to make specific recommendations about the age at which mother-to-child transmission of infections can be ruled out as the evidence in this area is scarce. If vertical transmission is suspected, it is good clinical practice to trace the family member concerned. The GDG believes that hepatitis B can be transmitted non-sexually within households so this should be ruled out as a cause before sexual abuse is considered.

There is a high prevalence of sexual abuse among children with anogenital warts. However, it is not known at what age vertical transmission can be excluded. The GDG were also concerned that it can be difficult for healthcare professionals to tell the difference between cutaneous warts or molluscum contagiosum and anogenital warts in the perineal region. The GDG concluded that healthcare professionals should consider sexual abuse in all children where they are concerned about anogenital warts and seek advice from a more experienced professional.

There is insufficient information about bacterial vaginosis, genital mycoplasma and oral warts in the context of sexual abuse to warrant inclusion in a list of possible STIs due to sexual abuse.

The GDG believes that the issues around consensual experimentation among 13- to 15-year-olds outlined in Crown Prosecution Service guidance⁴⁰ should be taken into account when a young person of this age presents with an STI: that guidance indicates that an STI in this age group is not an immediate reason to suspect sexual abuse.

The GDG believes that to consider an STI in young people aged 16 or 17 years to be a direct result of sexual abuse will depend on the context and nature of the sexual act. Therefore, the presence of an STI in this age group needs to be evaluated in the context of consensual sexual activity.

The GDG sought the opinions of the Delphi panel on recommendations about young people between the ages of 13 and 18 years (see above and Section C.2.5). The GDG accepted statements 21a and 22b from the Delphi survey. Although agreement was reached on Statement 22b, the GDG amended the definition of a ‘discrepancy in power, emotional maturity or mental capacity’ to provide examples that are meaningful for healthcare professionals.