Background
Child Sexual Abuse (CSA) diagnosis has been likened to a “jigsaw puzzle” [
1]. Whilst the child’s allegation is vital, physical evidence obtained by an appropriately qualified examiner [
2] can support criminal prosecution and child protection. Physical evidence has been the subject of consensus statements [
3] and systematic review [
4]. Anal findings are described following CSA [
5‐
13], in children selected for non-abuse [
14‐
16] and those with medical conditions affecting the anus [
17‐
21]. There have been two previous studies where anal signs in different groups of children were compared [
10,
11].
If present, anal signs may be used in children with a disclosure of CSA to provide corroboration for court proceedings, but it is not currently clear how much reliance can be placed on which signs. There is even less certainty about the extent to which anal signs seen in children with no disclosure or suspicion should raise concern about possible CSA and the need for further investigation.
This study aimed to compare the prevalence of anal findings as assessed by specialist forensically trained paediatricians in a group of children where the history included a statement by the child of anal abuse with a group of children with a history solely of non-sexual physical abuse or neglect and with no concerns re sexual abuse.
Results
A total of 19,785 children were seen and reported for child protection concerns in Leeds from January 1990 to December 2007, of whom 3,119 were categorized by the examining doctor as likely CSA. From these, 184 cases (105 boys, 79 girls) were identified with disclosure by the child of anal abuse, mean age 98.5 months, range 26 to 179 months, with only 7 younger than 3 years; 142 were identified from main database, 42 via the photographic database. There were 179 controls (94 boys, 85 girls, average age 83.7 months (range 35–193) from 1998 to 2007; 76 identified from the main database, 103 from photographic database.
Thirteen permanent paediatric staff examined 136 cases (74%) and 100 controls (56%) of whom three examined 35% of cases and 31% of controls. The remainder were examined by trainees supervised by forensically trained paediatricians.
In 134 cases where an object was specified, alleged penetration was penile for 64% (86) and digital for 30% (41). A majority of cases (74%) had one or more core Royal College of Paediatrics and Child Health signs [
4] (Table
2) and 43% two or more, compared to only 16% and 1% of controls respectively.
Table 2
Frequency of classic signs associated with anal abuse in cases and controls
Reflex anal dilatation | 41 | 22% | 0 | | 40.1 | 51.3 | <.0001 | 62.35 | 8.4 - 462 |
Gaping | 5 | 2.7% | 0 | | 4.9 | 5.0 | 0.12 | | |
Laxity/reduced anal tone | 49 | 27% | 5 | 2.8% | 9.6 | 4.9 | <.0001 | 13.7 | 5.3 - 35.8 |
Reddening/Erythema | 56 | 30% | 15 | 8.3% | 3.6 | 4.9 | <.0001 | 5.3 | 2.8 - 10.0 |
Perianal venous congestion | 66 | 36% | 1 | 0.6% | 59.8 | 99.6 | <.0001 | 101 | 13.8 - 743 |
Fissure/laceration | 26 | 14% | 2 | 1.1% | 12.8 | 14.6 | <.0001 | 13.5 | 3.1 - 58 |
Tag | 8 | 4.3% | 10 | 5.6% | 0.8 | 0.8 | >0.5 | | |
Scar | 10 | 5.4% | 0 | | 9.0 | 10.3 | 0.002 | 8.2 | 1.0 - 66.4 |
Anal or perianal bruising | 0 | | 0 | | | | | | |
None of the above signs | 48 | 26% | 150 | 84% | 0.31 | 0.07 | <.0001 | 0.059 | 0.03 - 0.10 |
More than one sign | 79 | 43% | 2 | 1.1% | 38.6 | 66.6 | <.0001 | 74 | 17.7 - 311 |
Total number | 184 | | 179 | | | | | | |
Training grade examiners reported fewer examinations where these signs were present than fully trained forensic paediatricians (cases: 68% versus 75%; controls 11% v 19%) but these differences were not significant (P = 0.4 and 0.2 respectively).
RAD and perianal venous congestion were seen commonly in cases but rarely or not at all in controls, resulting in high likelihood ratios. The estimated maximum horizontal diameter of dilatation was stated in 27 cases and was over 1 cm in 14 (52%) cases. Fissures and laxity were also seen more commonly in cases than controls. Anal tags were uncommon overall.
Of the less recognised signs, most were reported significantly more often in cases than controls (Table
3). Fold changes, were described fairly often, but the others were generally less common.
Table 3
Frequency of other anal signs not discussed by RCPCH (2008) in cases and controls
Fold changes | 34 | 18.5% | 3 | 1.7% | 10.9 | 13.3 | <.0001 | 8.7 | 3.0 - 25 |
Twitching | 17 | 9.2% | 2 | 1.1% | 8.4 | 9.1 | <.0001 | 9.2 | 2 - 41 |
Swelling | 12 | 6.5% | 0 | 0 | 11.8 | 12.6 | <0.001 | 15.4 | 1.9 - 120 |
Funnelling | 8 | 4.3% | 1 | 0.6% | 7.2 | 8.1 | 0.037 | 6.4 | 0.75 - 53 |
Mucosal prolapse | 8 | 4.3% | 0 | 0 | 7.2 | 8.1 | 0.007 | 8.1 | 1.0 - 70 |
Abrasion | 7 | 3.8% | 0 | 0 | 6.9 | 7.1 | 0.015 | 10.6 | 1.2 - 90 |
Deficit | 5 | 2.7% | 0 | 0 | NA | | 0.061 | | |
Warts | 1 | 0.5% | 0 | 0 | NA | | 1 | | |
Soiling | 5 | 2.7% | 11 | 6.1% | NA | | 0.13 | | |
There were no significant effects of age or examiner grade on the prevalence of signs (data not shown) and simultaneous adjustment for age, gender and examination era made no meaningful difference to the results (Tables
2 and
3).
The prevalence of signs varied with interval to examination (Table
4). Erythema, swelling and fold changes occurred most commonly within 7 days of the alleged assault. RAD, laxity, venous congestion, fissure and twitching were seen up to 6 months.
Table 4
Anal findings in cases by time interval between last episode of abuse to examination
Reflex anal dilatation | 17 (29%) | 13 (22%) | 9 (20%) | 2 (9%) | 0.21 |
Laxity | 18 (31%) | 15 (25%) | 11 (25%) | 5 (23%) | 0.76 |
Reddening | 21 (36%) | 23 (39%) | 11 (25%) | 1 (4.3%) | 0.002 |
Venous congestion | 22 (38%) | 23 (39%) | 15 (34%) | 6 (26%) | 0.28 |
Fissure | 5 (9%) | 11 (19%) | 9 (16%) | 1 (4%) | 0.21 |
Scar | 6 (10.3%) | 2 (3.4%) | 2 (4.5%) | 0 | 0.57 |
Any core sign | 44 (76%) | 48 (81%) | 33 (75%) | 11 (48%) | 0.005 |
2 or more core signs | 30 (52%) | 30 (51%) | 14 (32%) | 5 (22%) | 0.008 |
Fold changes | 10 (17%) | 15 (25%) | 7 (16%) | 2 (8.7%) | 0.07 |
Twitching | 2 (3.4%) | 9 (15%) | 4 (9.1%) | 2 (8.7%) | 0.32 |
Swelling | 5 (8.6%) | 6 (10%) | 1 (2%) | 0 | 0.04 |
Funnelling | 2 (3.4%) | 2 (3.4%) | 3 (7%) | 1 (4.3%) | 0.69 |
Mucosal prolapse | 4 (6.9%) | 3 (5%) | 0 | 1 (4.3%) | 0.57 |
Abrasion | 1 (2%) | 6 (10%) | 0 | 0 | 0.018 |
Total | 54 | 58 | 55 | 17 | |
History of constipation was recorded in 15 cases (7 boys, 8 girls), of whom 5 had RAD and 2 had fissures. There were 3 constipated controls (all girls) and each had one of venous congestion, a fissure and tag.
Discussion
Martial wrote in 1
st century AD that “the favourite sexual use of children was not fellatio, but anal intercourse” [
26]. Summit wrote “Manual, oral and anal containment of the penis are the “normal” activities of incestuous intercourse, as they are also for the more typically out of family sexual assault of boys” [
27]. Anal signs were central in the Cleveland Inquiry [
28] which recommended further study which in turn lead to publications by the Royal College of Physicians which provided guidance for clinicians [
29,
30]. Allegations of anal abuse appear to be relatively rare, as these disclosed cases represented only 5% of all CSA cases seen. This possibly explains why the recent RCPCH review noted a serious lack of evidence on anal signs in children [
4]. The resulting uncertainty has limited doctor’s ability to provide clear opinions.
Identification of a group where CSA can be confidently diagnosed or excluded is always challenging. While we cannot be certain that all the children who alleged anal abuse were true cases, it is generally accepted that disclosure is strongly indicative of abuse. Ideally the non abused controls would be sampled from the general population, but in practice recruiting a truly representative group and excluding CSA can be problematic. Selection of children from the general population has proved quite difficult, but it also raises serious ethical considerations. In one of the few studies of this kind [
15] only 10% of parents approached participated and some later admitted that concerns that their child had been abused motivated them to participate. In that study perianal venous congestion was more commonly seen (16%) than in another study where 1% of younger children showed this sign [
14].
A different approach was used in a recent study [
11]. Children evaluated for possible sexual abuse were divided into 2 groups, one with a low probability (917 children) and one with a high probability (198 children) of having been anally penetrated. Comparison was made between these groups in terms of the physical signs observed. However identifying comparison children with a low risk of having been anally penetrated in a group of children referred for sexual abuse evaluation is problematic as suggested by the presence of anal bruising in 10, anal fissure in 25 and anal laceration in 3. Consequently, the solution of choosing as controls children examined with concerns about other forms of abuse where the routine practice was to include anal examination seemed overall the best solution to us.
While physically abused and neglected children have a known increased risk of CSA [
31], in this study the fact that wide ranging sensitive information was available minimised the likelihood of including unrecognised CSA. However, it is possible that an occasional sexually abused child could unintentionally have been included in the control group and if so this would mean that the prevalence of signs seen in the controls would be overestimated. Control children with anal photographs were more likely to be included in this study than those without, and this could also have had the effect of overestimating the proportion of controls with positive findings. If this were the case that would imply that the true difference between groups was in fact even greater.
There were small differences in examiner status between cases and controls, cases were drawn over a longer time period than controls and the age range of cases and controls was slightly different, but statistical adjustment for all these factors made no meaningful difference to the results.
An important remaining concern is the possibility of examiner bias. When examining a child who has alleged anal abuse, a physician might be more confident in reporting abnormal findings than in a child with no such history. However both groups were examined by the same staff who would be alert to the possibility of undisclosed anal abuse and with experience of eliciting the signs in question. This makes it possible that examiners in this centre were more likely to detect signs in general, but this would apply to both cases and controls.
Thus while the limitations of the samples must be recognised, this remains the first case/control study in which a large group of children all of whom disclosed anal abuse was examined using the same techniques and examiners as controls, using well defined terminology. The difference in frequency of some signs between cases and controls suggest that they are likely to relate to abuse. In particular RAD and perianal venous congestion were seen frequently in cases, but rarely or not at all in controls. RAD is dramatic, involves dilatation of both sphincters, requiring observation for up to 30 seconds as it does not always appear immediately. Previous studies found RAD in 10% to 34% disclosing anal abuse and 5% to 20% reporting any sexual abuse (Table
5) [
4]. In children selected for non-abuse, RAD was noted in 5% examined in the knee chest, but less than 1% in the left lateral position [
16]. Another study [
15] found none with the sign. An earlier study which has influenced practice especially in North America [
17] described anal dilatation in 49% children selected for non-abuse examined in the knee chest position, for up to 8 minutes. But this position is rarely used in the UK. Apart from that study our figures for cases (22%) and controls (0%) lie within the range of other studies for both abused and “non-abused” in the left lateral position.
Table 5
Comparison with published studies reviewed by RCPCH[
4]
Reflex anal dilatation | 22.3% | | 0% | <1 – 3.6% (left lateral) [ 11, 14] |
Laxity/reduced anal tone | 26.6% | | 2.8% | No reports |
Reddening/erythema | 30.4% | | 8.3% | |
Perianal venous congestion | 35.9% | | 0.6% | |
Fissure/laceration | 14.1% | | 1.1% | |
Anal or perianal bruising | 0% | | 0% | |
Any signs | 74% | | 16% | No reports |
Anal laxity (reduced anal tone) was seen more commonly in our cases than in earlier studies [
7,
12,
36], but had never been previously considered in children selected for non-abuse (Table
5). Anal fissure and laceration are injuries in the perianal skin. There is a lack of agreed definitions to fully differentiate them. Our figures which combine fissures with lacerations gave prevalence for both cases and controls which were within the range described in other studies (Table
5). Perianal venous congestion was at the upper end of the range for cases in previous studies and the lower end for controls (Table
5). As with most previous studies, anal bruising was uncommon following abuse and rarely reported in “non-abuse”. Erythema was seen more commonly than in previous studies probably reflecting a higher proportion examined soon after an assault than in previous studies.
Anal dilatation and venous congestion were so rarely seen in controls, that it raises the possibility that they should be recognised as signs which should prompt further investigation, as long as they are interpreted in the broad context of a detailed medical, social and family assessment and the child’s behaviour and demeanour.
The highest frequency of signs was seen in those abused less than 7 days previously and in those where the timing of the abusive episode was not known. But none of the signs were seen only within 7 days of the alleged assault, suggesting that examination is worthwhile even some weeks after the alleged assault.
The majority of cases had at least one sign, though in many these were non-specific. This observation is consistent with previous studies reviewed by the RCPCH [
4]. Of seven studies reporting any abnormal signs, two found these in 61-95% [
5,
35] and two in 46% and 57% [
6,
33], despite widely differing methodology and definitions. However a quarter had no signs, so the absence of physical signs could not be said to negate a child’s history or exclude the possibility of abuse.
Competing interests
CJH is a retired NHS consultant who undertakes locum work that may involve child protection assessments and also provides expert medico legal opinions on Child Protection cases for which he receives a fee. CW is an honorary NHS consultant who advises on academic aspects of Child Protection and does not usually undertake paid medico legal work.
Authors’ contributions
CJH was involved in: the conception, design, analysis and interpretation of data. Drafting the article and revising it critically for important intellectual content. Final approval of the version to be published. CMW was involved in: analysis and interpretation of data. Drafting the article and revising it critically for important intellectual content. Final approval of the version to be published. Both authors have given final approval for the publication of this manuscript.