Subjects, methods, and results

Three hundred and nineteen children (77 boys, 242 girls) between the ages of 9 months and 16 years, attending an inner London paediatric department for evaluation of sexual abuse entered the study from July 1991 to July 1994. The diagnosis of child sexual abuse was based on the following: disclosure by the child; outcome of police and social services investigation; care and/or criminal proceedings; and the presence of physical findings supportive of abuse, such as hymenal tears. Swabs were taken of vaginal discharge (if present) or from the vaginal introitus, for Gram stain and culture for Neisseria gonorrhoeae, Chlamydia trachomatis,Trichomonas vaginalis, Candida spp, mycoplasmas, and other bacterial infections using standard methods.14 Bacterial vaginosis was diagnosed on two criteria, clinically abnormal vaginal discharge and clue cells on Gram stain. As prepubertal children have alkaline secretions, pH is invalid in this population as a diagnostic criterion for bacterial vaginosis. Anal swabs were taken from boys and girls reporting anal abuse, or with anal findings suggestive of abuse, for culture for N gonorrhoeaeand C trachomatis. Urethral swabs were sent for isolation of N gonorrhoeae and C trachomatis. Blood tests for human immunodeficiency virus (HIV) and treponemal antibodies were performed at the request of parents or child in cases involving multiple abusers or when the perpetrator was known to be HIV positive. Genital warts were diagnosed clinically.

Swabs could not be taken from all sites in all children. Swabs were not taken from 83 girls and 47 boys: 49 children refused; 10 samples were toxic to culture; in the remainder there were no physical findings and either no evidence of sexual abuse that involved touching or penetration, or a delay of more than a year to presentation from time of abuse. Vaginal swabs were taken from 159 girls, and anal swabs from 63 children (28 boys and 35 girls); two boys had both urethral and anal swabs and two adolescent boys had only urethral swabs taken. Two throat swabs were taken from girls giving a history of oral sex.

Fifteen boys were aged under 3 years, 42 were between 3 and 10 years, and 20 were over 10 years of age. No sexually transmitted organisms were detected in the 28 boys who had anal swabs taken or in the four boys who had urethral swabs taken. Six of these 28 boys were aged under 3 years, three of whom were infected with C albicans; 15 were between 3 and 10 years and seven were aged 10 years and older and they had no infections. Two boys were identified as abusers and had also been abused.

The infections identified according to the different age categories of girls are presented in table 1. Three girls (aged 11 months, 2½ years, and 13 years), all with vaginal discharge, were found to have gonorrhoea. Sexual contact was disclosed by the two oldest girls. Neither parent of the 11 month old girl had evidence of gonorrhoea and a neonatal infection screen, including eye swabs, was negative making vertical transmission unlikely. Two of 131 children screened for C trachomatis were positive, one by direct immunofluorescence only. Both had evidence of penetrative abuse and one had vaginal discharge. T vaginalis was found in four girls, all of whom were Afro-Caribbean, with oestrogenised hymena and vaginal discharge. One had anogenital warts. All had evidence of penetrative abuse (three hymenal tears, one enlarged hymenal opening, one previous termination of pregnancy).

Nineteen girls had candidal infection diagnosed on culture. Six of the younger children had a typical candidal nappy rash without laboratory confirmation. Mycoplasmas were identified in 21 (22%) of the 96 girls for whom swabs were available (14 Ureaplasma urealyticum, four Mycoplasma hominis, three dual infection). Three girls were prepubertal and 15 had evidence of penetrative abuse. One girl disclosed digital abuse but her father admitted penile penetration.

A heavy growth of anaerobes was found in seven children. Five of the girls disclosed abuse involving touching of the genital area, one disclosed fellatio, and one child did not describe the type of abuse. Bacterial vaginosis was diagnosed in three girls, one of whom disclosed licking, one penetrative sex, and one unknown.Mobiluncus spp were identified on Gram stain of vaginal discharge in a 6 year old girl who alleged touching. Of these 11 girls, six had discharge, four had erythema alone, and one was asymptomatic. A girl aged 11 years and a boy aged 9 years had anogenital warts. None of six children tested serologically for syphilis and HIV were positive.

Other organisms identified included staphylococci, streptococci,Haemophilus influenzae, and Escherichia coliin eight girls aged 3–10 years, all of whom had vulvovaginitis.

Overall, seven of the 159 girls (3.7%) had gonorrhoea, chlamydial infection, or trichomoniasis. One who was gang raped had all three infections plus mycoplasmas.

Comments

Our prevalence of 3.7% is similar to another British study.1 All the girls with gonorrhoea or trichomoniasis and two of three with chlamydial infection had vaginal discharge. The largest prospective study found 90% of girls with gonorrhoea and 33% of girls with C trachomatis had vaginal discharge.2

These results support mandatory screening for N gonorrhoeae, C trachomatis, and T vaginalis when there is a history of abuse and presence of vaginal discharge. Ideally, screening for these organisms should be included in any child sexual abuse investigation.

Mycoplasmas were present in 22% of girls swabbed in our study. All had either a history of or physical signs suggestive of penetrative sex. The identification of mycoplasmas cannot be deemed to be diagnostic of sexual abuse because they can be transmitted vertically and have been found in non-sexually active girls. Two reported studies comparing controls with sexually abused girls give conflicting conclusions on the significance of mycoplasmas in sexually abused children.15 16 Further controlled studies are needed to assess the significance of mycoplasmas. Similarly, the significance of a heavy growth of anaerobes and organisms associated with bacterial vaginosis9 needs evaluation in relation to abuse. Finding suitable control groups for such studies poses ethical and practical difficulties, particularly in relation to consent, and matching for previous antibiotic use and age. Identification of organisms (N gonorrhoeae, C trachomatis) using DNA amplification technology has been reported in prepubescent children using vaginal washings.17 Samples obtained non-invasively, such as urine, might be advantageous for screening abused children and might increase our understanding of genital infections in children.