• CHLAMYDIA INFECTION
• CHILDREN
• TRACHOMA
• MOLECULAR GENETICS
• SEXUAL ABUSE

The notion that sexually transmitted infections (STIs) can be acquired through non-sexual means is a well-established meme in the public consciousness. The public toilet seat is the most commonly blamed vector. This is despite the lack of evidence for acquisition of STIs from toilet seats, and the ready availability of reputable advice to this effect.1

While the ‘toilet seat hypothesis’ is not currently regarded as supported by data, other modes of non-sexual acquisition of STIs, though rare, have been reported. Examples include pharyngeal gonorrhoea by sharing of confectionery,2 transmission of urogenital (UGT) gonorrhoea by sharing of a sex doll,3 and transmission of Trichomonas vaginalis via the fingers of a traditional healer in west Africa.4 This provides credibility to the notion that STI detection in UGT specimens from young children is not always a consequence of ‘conventional’ sexual contact or, for very young children, vertical transmission. Estreich and Forster5 reviewed relevant evidence and concluded that inferring the cause of an STI in a young child is a ‘complex subject’, with vertical transmission, child sexual abuse and accidental/fomite transmission all conceivable mechanisms. This begs the question that is critical to service providers: How strongly does the presence of an STI in a young child indicate sexual abuse? This dilemma remains a challenge.6 In individual cases, potential modes of non-sexual transmission may be invoked, and the plausibility of these can impact child protection and/or justice system responses. Incorrect interpretation has the potential for a child remaining in an unsafe environment, or conversely leading to a false conclusion that a crime has taken place, resulting in initiation of an investigation. The strong imperatives to avoid either of these errors can create instability in institutional responses, and inappropriate management of a child. There is therefore a strong case for gaining numerical understanding of the probabilities of non-sexual acquisition of STIs in young children.

In Australia, the ‘Little Children are Sacred’ report in 2007 detailed evidence for child sexual abuse in Aboriginal communities in remote regions of the Northern Territory of Australia.7 This report was the catalyst for the Australian Government interventions, the National Emergency Response and the subsequent Stronger Futures programme. The true nature and extent of child sexual abuse in this region remains unclear. Nevertheless, there is little doubt that extensive media reporting has considerable potential to negatively stereotype Aboriginal people and create a highly charged and complex socio-political climate8 that continues to this day. There are differences of opinion even among service providers regarding whether the detection of an STI agent in a paediatric UGT is a reliable indicator that sexual contact has occurred. This is particularly so for Chlamydia trachomatis. It is well understood that this species can be vertically transmitted from a maternal genital infection to her new-born child and occasionally cause a chronic infection in the child for years.9 However, another potential confounding factor is autoinoculation from ocular infections. This is particularly relevant to remote Aboriginal populations where trachoma may be endemic. The notion that C. trachomatis positive paediatric UGT specimens can result from self-inoculation from ocular infections is a widely held belief, which continues to be conveyed through informal channels. Self-inoculation in the other direction, from UGT to the ocular site is well established,9 adding credibility to this notion.

Giffard and co-workers have embarked on a programme to enhance the evidence base for responding to instances of C. trachomatis detection in paediatric specimens. They recently reported a study designed to illuminate the question of self-inoculation, by determining if trachoma genotypes were circulating in sexual networks in the remote Northern Territory of Australia. No such evidence was found.10 It was concluded that genotyping C. trachomatis from paediatric UGT specimens has the potential to indicate whether or not the organism in the specimen was epidemiologically connected to adult sexual networks. More specifically, the detection of a trachoma genotype in a paediatric specimen should perhaps not necessarily be classified as detection of an STI agent. While the absence of trachoma genotypes in UGT specimens may appear to be an unsurprising result, it is markedly different to what was found in a study performed in the late 1980s to early 1990s in the same geographical region, in which strong evidence for sexual transmission of the trachoma genotype B was found.11–13 The presence of trachoma genotypes, and in particular genotype B, in UGT specimens is not unprecedented;14 ,15 hence, in the absence of an ongoing surveillance, strict correspondence between genotype and site of infection cannot be assumed.

It is clear that a numerical description of the informative power of C. trachomatis detection and genotyping with respect to prior sexual contact would be of value. We are developing this conceptual framework and have reasoned that the exercise is simplified by defining the formal diagnostic target as ‘having acquired C. trachomatis from adult sexual networks’. This means that the difficult-to-estimate sensitivity of C. trachomatis presence for indicating sexual contact is eliminated. This is justified because patients who have experienced sexual contact without acquiring C. trachomatis are by definition invisible to this diagnostic approach and may be regarded as ‘true negatives’. The relationship between post-test probability of acquisition of C. trachomatis from adult sexual networks, and the proportions of ocular and UGT genotypes in paediatric and adult UGT specimens can be explored. For example, if (hypothetically) 40% of C. trachomatis detected in paediatric UGT specimens was of ocular origin and essentially absent in adult specimens, 10% of the paediatric UGT specimens contained C. trachomatis as a result of contact with adult sexual networks, and 94% of the C. trachomatis circulating in adult sexual networks was of UGT genotype (and 6% of ocular genotype), then genotyping has a very large effect on post-test probabilities. Specifically, simply detecting C. trachomatis in a paediatric UGT specimen (ie, no genotyping performed) changes the probability of having acquired C. trachomatis from adult sexual networks from 0.1 (pre-test probability) to just 0.22. However, the post-test probability when a UGT genotype is detected approaches 1.0, with the derived diagnostic odds ratio and positive likelihood ratio both approaching infinite. In contrast, detection of a trachoma genotype is not indicative of acquisition of C. trachomatis from adult sexual networks. Indeed, if it is assumed that where there is both UGT and trachoma genotypes in a specimen, then the result is classified as ‘UGT genotype’, then a ‘trachoma genotype’ result reduces the post-test probability of contact with adult sexual networks from 0.1 to 0.017. It is important to note that ‘contact with adult sexual networks’ is not necessarily equivalent to direct sexual contact, because other transmission routes, for example, derived from adult cases of conjunctivitis and transmitted first to the ocular site, can never be entirely ruled out. Resolving this issue is difficult; however, a prediction of self-inoculation from the eye to the UGT is genetically identical strains at both sites.

It is feasible to populate this analysis with experimental data. The key parameters are trachoma and UGT genotype frequencies in adult and paediatric specimens at any specific time period. The frequency of trachoma genotypes in adult specimens shows the frequency of trachoma genotypes of C. trachomatis circulating in adult sexual networks, while any excess of trachoma genotypes in paediatric specimens indicates the frequency of acquisition of C. trachomatis independent of contact with adult sexual networks. Ongoing surveillance to continuously compare the proportions of C. trachomatis genotypes in adult specimens with specimens from paediatric patients who are the subject of child protection investigations would be an efficient and effective strategy. Determination of the pre-test probability of acquisition of C. trachomatis from adult sexual networks is potentially problematic, but the frequency of paediatric specimens positive for UGT genotype C. trachomatis is a reasonable estimate. It is also important to note that assessing the probability of chronic infection related to maternal transmission, which is usually regarded as very low after 2–3 years, would be significant in cases such as this.

In summary, a consequence of this conceptual framework is that UGT genotype detection in a paediatric sample is a specific indicator of contact with adult sexual networks, with the sensitivity of a function of the UGT genotype frequency in adult UGT specimens. In contrast, the specificity provided by the detection of a trachoma genotype is inversely correlated with the excess of ocular genotypes of paediatric specimens in comparison to what is found in adult UGT specimens. Therefore, surveillance of C. trachomatis genotypes in adult and paediatric specimens can provide critical information regarding transmission mode.