The clinical assessment of vaginitis is traditionally based on a combination of clinical symptoms, medical history, microscopy techniques and microbiological culture1,2. Some authors supported that G. vaginalis is the most frequent bacteria responsible for BV and a good predictor of infection; however, other microorganisms may be involved6,14. Indeed, the combination of G. vaginalis with A. vaginae is highly associated with biofilm formation causing recurrence and therapeutic failures5–7,11,12. Due to the different etiologies of the infection and the diverse bacteria involved, the algorithm for the diagnosis of BV is still under discussion. As previously reported14,19, the results obtained with the qPCR test showed a high percentage of G. vaginalis and A. vaginae coinfection. In this line, an advantage of the qPCR test is that it reports the presence of A. vaginae, which might be relevant for treatment choice.
The Nugent score categorizes the sample into three groups whereas the Vaginal Panel Realtime PCR Kit classifies the sample into two groups and each one into two grades (see Methods) so a direct comparison could not be performed. This two-group classification avoids the uncharacterized “intermediate” scoring of the Nugent criteria. This is why intermediate scores were excluded from the analysis shown in Table 2. A low rate of discrepancies between both methods was obtained. The exhaustive comparative analysis of discrepancies enhances the difference between reference methods and qPCR test to identify microorganisms and quantify bacteria loads.
Although Lactobacillus species have been defined as a negative predictor of BV14,20, BV is caused by a decrease in this microorganism6. Moreover, the vaginal microbiome fluctuates during women’s lives, increasing during pregnancy and decreasing during menopause21,22. Studies using in-house or commercial NAATs reported the advantages of including Lactobacillus as a positive marker of healthy microbiota16,22–25. In this way, the qPCR test includes the detection of the most common Lactobacillus species. We obtained 23.6% cases without anaerobic bacteria load with absence or very low levels of Lactobacillus spp. Its inclusion should be a good indicator of the vaginal microbiota state and when infection may be caused by another microorganism.
In our study, the sensitivity, specificity, PPV and NPV for the diagnosis of Candida species showed very good agreement between the qPCR test and the reference methods (Table 4). The prevalence of C. albicans was 87.7% higher than that of non-albicans Candida species (10.3%). Nevertheless, due to the increase in non-albicans Candida species in the recent years5,8, their diagnosis should be considered because they may be involved in recurrent infections and treatment failure2,8,26. VVC due to C. glabrata infection has been shown to have a higher risk of recurrence than VVC due to C. albicans27. The qPCR test detects the most frequent non-albicans Candida species, providing early detection compared to reference methods.
For T. vaginalis, the Vaginal Panel Realtime PCR Kit demonstrated high sensitivity and specificity compared with reference methods. When we compared the results obtained with the two NAAT tests, there were very similar results (twelve cases were detected by both methods and 3 cases were not confirmed by the third method). Although traditional techniques (wet mount or culture) for T. vaginalis remain widely used, we confirm that NAATs provide an accurate diagnosis18. We cannot conclude if the three positive cases reported by the qPCR test and not confirmed by a third method were false positives or due to a difference in sensitivity between both tests.
Despite the results obtained, this study had three main limitations. First, a direct comparison of BV could not be established due to the different classifications of vaginal microbiota between methods as previously described. Second, only one vaginal swab was collected, which could influence the lower detection rate by the molecular method. Finally, the anonymity of the participants in the study does not provide us with information about symptoms, ethnicity, contraceptive use, sexually transmitted diseases or menopausal state. All these factors are closely related and have a high influence on vaginal microbiome variations 2,8–10. A greater insight of clinical data would have provided a further analysis to better understand the relationship between the previous issues and vaginitis infections.
Accurate vaginitis diagnosis is needed to provide better treatment and reduce recurrent infections. NAAT use has increased in the recent years as it allows for the simultaneous detection of several microorganisms but Gram staining remains widely used. While traditional laboratory diagnosis takes at least 2–3 days to provide preliminary results, the qPCR test should be available within 3–4 hours. As it is thus far unknown if the high sensitivity of NAATs may lead to overdiagnosis, it is of utmost importance to combine data coming from these tests with the symptoms and clinical history of the patient. Indeed, reference methods are observer-dependent and require skilled laboratory technicians to identify microorganisms14,20.
In conclusion, the Vaginal panel Realtime PCR kit demonstrated appropriate performance in the diagnosis of vaginitis infections producing unbiased results compared to the reference methodology, and providing an increased accuracy in the diagnosis of T. vaginalis.