Susceptibility of Trichomonas vaginalis to metronidazole and other compounds.

Abstract

Trichomonas vaginalis is the most common sexually transmitted infection caused by a single known organism worldwide; and has been associated with an increased risk of HIV acquisition and transmission. Despite its high prevalence in South Africa, limited information is available on the extent of T. vaginalis metronidazole resistance and genotypic variation in this setting. We therefore tested the susceptibility of local T. vaginalis isolates against metronidazole and drugs prescribed in combination in the context of syndromic management of vaginal discharge syndrome. Susceptibility testing of 40 isolates demonstrated that metronidazole as well as some of the other drugs tested showed inhibiting effect on T. vaginalis. We recommend that these drugs be tested for synergistic effect with metronidazole. In a different set of 160 isolates the minimum inhibitory concentrations (MIC) of metronidazole ranged from 1.1 μg/ml to > 34.2 μg/ml (6.25 μM to > 200 μM) in the aerobic assay. Interpretation of these MICs differed based on the different resistance breakpoints applied. There was no correlation between MIC and treatment outcome in the subset of 56 patients that returned for follow-up. The expected association between MIC and clinical outcome was only observed in one of eight patients with unsatisfactory treatment outcome. This patient‘s isolate had the highest MIC. In the remaining seven patients with unsatisfactory treatment outcome, no relation with the susceptibility test result was found. A possible reason for the poor correlation may be inadequate concentration of metronidazole at the site of infection. In view of this, we assessed a self-administered and collected vaginal tampon specimen for the investigation of metronidazole concentration in the vagina of five healthy volunteers, using high performance liquid chromatography (HPLC). Maximum values of metronidazole concentrations detected in both serum and vaginal fluid were obtained at two hours following oral administration of 2 g of the drug. This method can be applied in future clinical studies to correlate treatment outcome and MICs with metronidazole concentration at the site of infection. This may lead to the development of susceptibility assays and interpretation criteria that are better able to predict treatment outcome than the current methods. Another reason for the poor correlation between treatment outcome and in vitro resistance may be early reinfection. We used PCR-RFLP, targeting a 650-bp repeat region in the T. vaginalis genome, to genotype T. vaginalis isolates. Four genotypes were found in 100 T. vaginalis isolates using this method. Both the vaginal secretion of metronidazole and the strain typing methodology needs to be further investigated before a comprehensive study as outlined above can be executed.