Routine universal testing versus selective or incidental testing for oropharyngeal Neisseria gonorrhoeae in women in the Netherlands: a retrospective cohort study.

BACKGROUND: Women are not routinely tested for oropharyngeal Neisseria gonorrhoeae. At present, selective testing based on sexual behaviour or risk groups is advocated by international guidelines. Many oropharyngeal infections are asymptomatic and thus remain undetected, establishing a reservoir for ongoing transmission. Data on effectiveness of routine testing are scarce, thus we aimed to assess the optimal testing strategy for oropharyngeal N gonorrhoeae in women.

METHODS: In this retrospective cohort study, we used surveillance data obtained from all sexually transmitted infection (STI) clinics in the Netherlands between Jan 1, 2008, and Dec 31, 2017. We collected consultation-level data, and individual-level patient data from 2016 onwards, on sociodemographic characteristics, sexual behaviour in the past 6 months, self-reported symptoms, and STI diagnoses. We compared the prevalence of oropharyngeal N gonorrhoeae between women who attended clinics that offered routine universal testing (defined as >85% of women tested per clinic-year) and women who attended clinics that offered selective testing (defined as 5-85% of women tested per clinic-year) or incidental testing (<5% of women tested per clinic-year). We calculated the number of infections missed by selective testing by extrapolating prevalence for the routine universal testing group to that of weighted and unweighted samples of all selectively tested women. We used multivariable generalised estimating equations to identify independent risk factors for oropharyngeal N gonorrhoeae to identify the optimal selective testing strategy.

FINDINGS: 554 266 consultations with at least one N gonorrhoeae test were recorded, of which 545 750 consultations (including repeat visits) were included in the analyses. Of 545 750 consultations, routine universal testing was used in 57 359 (10·5%), selective testing in 444 283 (81·4%), and incidental testing in 44 108 (8·1%). The prevalence of oropharyngeal N gonorrhoeae was 1·4% (95% CI 1·3-1·5; 703 of 50 962 consultations) in the routine testing group compared with 1·4% (1·3-1·5; 1858 of 132 276) in the selective testing group (p=0·68) and 2·8% (1·9-3·9; 30 of 1088) in the incidental testing group (p<0·0001). The prevalence of oropharyngeal-only infections was 47·7% (335 of 703 women) in the routine testing group, 53·3% (991 of 1858) in the selective testing group, and 60·0% (18 of 30) in the incidental testing group. Selective testing would have missed an estimated 4363 (70%; 95% CI 69-71) of all 6221 oropharyngeal N gonorrhoeae infections. Independent risk factors for oropharyngeal N gonorrhoeae were being notified for any STI (adjusted odds ratio 2·1, 95% CI 1·5-3·0), reporting sex work (4·0, 2·3-6·7), and having concurrent genital (51·5, 34·1-77·7) or anorectal (2·6, 1·4-4·8) N gonorrhoeae. Selective testing of women notified for any STI, or who reported sex work, would have led to 5418 (27·8%) of 19 455 women being tested and would have identified 119 (55·6%) of 214 oropharyngeal N gonorrhoeae infections.

INTERPRETATION: Selective testing potentially misses more than two-thirds of oropharyngeal N gonorrhoeae infections in women, of whom half have oropharyngeal infections without concurrent genital or anorectal infections. Using independent risk factors for oropharyngeal infection to guide testing is a minimal testing strategy. Routine universal testing is the optimum scenario to detect the majority of infections. However, future studies are needed to assess the cost-effectiveness of routine testing and its effect on antimicrobial resistance.

FUNDING: Public Health Service South Limburg.