Objective: To evaluate the validity of clinical algorithms proposed in Benin for the diagnosis of gonococcal or chlamydial infections among men with urethral discharge or dysuria and women with vaginal discharge consulting health services in Benin. These algorithms were adapted from those proposed by the World Health Organisation.
Methods: Consecutive patients with these symptoms were enrolled at three primary healthcare centres. The reference test for gonorrhoea was a combination of results from culture and polymerase chain reaction and chlamydial infection was ascertained by enzyme linked immunosorbent assay and PCR. In women, two algorithms were evaluated, one based on symptoms and risk assessment (algorithm A), the other relying also on speculum examination (algorithm B). The first algorithm evaluated in men relied on clinical examination only (algorithm C) whereas the other used microscopic examination of urethral smears (algorithm D). Sensitivity, specificity, and positive and negative predictive values of these algorithms were assessed using standard methods.
Results: In 192 women, the prevalence of gonococcal and chlamydial infections was 5.7% and 2.1% respectively (combined prevalence of 7.8%). The sensitivity, specificity, positive and negative predictive values of algorithm A (algorithm B) were respectively 86.7% (93.3%), 41.8% (34.5%), 11.2% (10.8%), and 97.4% (98.4%). In 105 men, the corresponding prevalences were 39.0% and 7.6% respectively (for a combined prevalence of 44.8%). The sensitivity, specificity, positive and negative predictive values of algorithm C (algorithm D) were respectively 91.5% (87.2%), 60.3% (67.2%), 65.2% (68.3%), and 89.7% (86.7%).
Conclusion: A syndromic approach for the diagnosis of urethritis in men appears appropriate. In women, the diagnosis of gonococcal or chlamydial infection without specific laboratory tests, which are not easily affordable in developing countries, remains highly problematic.