Evaluation of the Hologic Aptima® Vaginitis and Seegene Allplex™ assays for microbiological diagnosis of bacterial vaginosis and Candida vaginitis in women in France, relative to Nugent score and culture

Introduction

Vaginitis is a common gynaecological condition characterized by inflammation of the vagina [1]. Symptoms include abnormal discharge, odour, pruritus and burning [2]. Where a microbiological cause is identified, the most common causes of infectious vaginitis are bacterial vaginosis (BV; 40–50% of cases), Candida vaginitis (CV; 20–25%) and Trichomonas vaginitis (TV; 15–20%), all of which result in a shift in the composition of the vaginal microflora and disruption to vaginal health [3].

BV is defined by vaginal dysbiosis, whereby the normal Lactobacillus-dominated microbiota shifts towards an anaerobe-enriched environment, dominated by genera including Gardnerella, Atopobium, Mobiluncus, Prevotella, Bacteroides and Peptostreptococcus species [4–6]. In CV, >90% of infections are due to C. albicans, with the remainder consisting of C. glabrata and other species [7]. The anaerobic protozoan T. vaginalis is the cause of TV [2].

Although treatment of vaginitis can achieve high cure rates, recurrent and persistent infections are common [8–10]. Carriers can also be asymptomatic, often remaining undiagnosed and untreated until their condition is picked up during routine testing for sexually transmitted infections (STIs) [11]. Complications of undertreated vaginitis can be severe, such as pelvic inflammatory disease, which can be associated with chronic pelvic pain, ectopic pregnancy and infertility [12]. In pregnancy, BV and TV have been linked to complications such as premature birth, low birth weight and infections of the amniotic fluid [1, 13]. Chronic vaginitis also increases the risk of contracting and transmitting STIs [1, 13, 14]. As such, early and reliable diagnosis is critical to informing appropriate treatment and achieving optimal outcomes for patients.

In France, BV is routinely diagnosed using either Nugent or Hay–Ison scoring of Gram-stained vaginal smears [15, 16]. The Nugent system assigns a score from 0 to 10, reflecting the relative abundance of Lactobacillus versus BV-associated morphotypes [15]. Hay–Ison scoring classifies Gram-stained vaginal smears into four categories (Grades 0–III), reflecting the spectrum from normal Lactobacillus-predominant flora to BV-associated flora [16]. For CV, microscopy is used in first-line diagnostics, with culture reserved for species identification where required [17]. Wet mount microscopy or culture can still be used for the diagnosis of TV, though molecular diagnostic assays, such as nucleic acid amplification tests, are now recommended in European guidelines [17].

Each of the aforementioned non-molecular methods faces challenges in practice. Microscopy depends on technical expertise and has variable sensitivity, while Nugent scoring is inherently subjective and often requires consensus between reviewers [18–20]. CV culture is also resource-intensive and may fail to reliably identify non-albicans species [21, 22]. Lastly, TV detection by wet mount is highly time-sensitive, requiring immediate examination of fresh samples [22, 23]. As such, the latter method is often associated with reduced sensitivity due to organism degradation during transportation [22, 23].

In recent years, commercially available molecular diagnostic assays have been developed to address these limitations [19, 22, 24–31]. Compared with conventional diagnostics, molecular assays aim to offer enhanced specificity and sensitivity, with the potential to detect multiple pathogens simultaneously. The Hologic Aptima^® BV and Aptima^® CV/TV multiplex assays (Aptima^® Vaginitis Panel) use transcription-mediated amplification (TMA) of ribosomal RNA rRNA targets from vaginal swabs collected via Hologic’s proprietary Aptima^® Multitest Swab Specimen Collection Kit and processed on the Hologic Panther^® platform [32, 33]. In Canada, the USA, the UK and France, Aptima Vaginitis assays have demonstrated sensitivities of 91.1–98.8% and specificities of 77.7–96.3%, relative to standard-of-care (SOC) diagnostics [22, 34–37].

By contrast, the Seegene Allplex^™ Vaginitis assay uses multiplex real-time (RT) polymerase chain reaction PCR to simultaneously detect the major pathogens that cause vaginitis [27, 38]. It uses Seegene’s automated All-In-One Platform, which includes automatic interpretation of infectious vaginitis. Vaginal swabs for testing with this assay can be collected using standard collection methods (i.e. non-proprietary kit).

Despite the availability and clinical use of both CE-approved assays in France, published evidence comparing the performance of both assays is currently limited to one relatively small study, which focused solely on BV, as opposed to the full spectrum of infectious vaginitis [39]. While the aforementioned study also included an analysis of non-proprietary swab use, further evidence is needed to better characterize Aptima’s performance with routinely used swabs and its potential impact on laboratory workflow.

Therefore, the present study compared the diagnostic performance of the Hologic Aptima BV and Aptima CV/TV assays and the Seegene Allplex Vaginitis assay against reference standards of Nugent scoring and culture, using the Σ-TRANSWAB^® collection method. These findings aim to inform evidence-based optimization of vaginitis diagnostics in France.

Methods

Results

Discussion

We set out to evaluate the clinical performance of the Aptima Vaginitis Panel and the Seegene Allplex assay for the diagnosis of infectious vaginitis in women attending primary care centres in France. Previous studies have demonstrated the robust performance of the Aptima Vaginitis Panel using its proprietary Aptima Multitest swab, with high sensitivities and specificities for the detection of BV, CV and TV [22, 35, 36]. In the current study, our results demonstrate the high clinical performance of both assays for the detection of BV and CV, even when using a non-proprietary sample collection method (Σ-TRANSWAB) with the Aptima assays.

In particular, regarding the detection of BV, both assays reported sensitivities and specificities of ~90% or above, relative to Nugent score. Sensitivity for the Aptima BV assay relative to Nugent score was numerically higher than that of Seegene relative to Nugent, at 97% versus 90%. However, it is important to note that this study was descriptive in nature; therefore, further appropriately powered statistical analyses would be required to confirm potential performative differences between both assays. When the evaluation of Aptima BV was performed relative to the Seegene Allplex assay, the Aptima BV assay demonstrated a consistently high performance, with PPA and NPA values ≥90%.

When testing for CV, we observed similarly robust and comparable performances for both assays relative to SOC culture, with sensitivities of 88.6% for the Aptima CV/TV assay and 90.2% for Seegene Allplex, and specificities of 90.8% for the Aptima CV/TV assay and 93.7% for Seegene Allplex. With respect to TV, we observed an extremely low pervasiveness of TV among the sampled population (n=1), limiting any further evaluation of assay performance for the detection of TV. Notably, the low prevalence of TV in our study is in agreement with previous findings of a French cohort screened for STIs (N=41,507; TV-positivity rate of 0.1%) [40].

Although it was not the key focus of our analysis, we also performed evaluations of the Aptima Vaginitis Panel using its own proprietary Aptima Multitest Sample Collection Kit to provide context for our Σ-TRANSWAB findings. We report that the performance observed using the Aptima Multitest swab was consistent with that obtained using Σ-TRANSWAB samples, demonstrating high diagnostic accuracy irrespective of sample collection method.

Sensitivity and specificity data from the current study also align with previous reports from the literature. For example, the Aptima Vaginitis Panel has reported sensitivities ranging from 91.1 to 98.8% and specificities ranging from 77.7 to 96.3% relative to SOC using its own proprietary sample collection kit [22, 34–37]. Additionally, the Seegene Allplex assay has previously been reported to exhibit a sensitivity of 91.7% and specificity of 86.6% for the detection of BV, and 91.1 and 95.6% for the detection of CV [27]. Regarding the direct comparison of the Aptima Vaginitis Panel versus Seegene Allplex when using non-proprietary swabs for sample collection, a recent small-to-moderately sized study in the Netherlands reported a PPA of 77% and an NPA of 100% for the detection of BV [39]. The NPA observed in our study (97%) closely aligns with this finding, though the PPA was notably higher in our cohort (87% vs 77%). This difference may reflect variation in sample size, population characteristics or extraction protocols between studies, rather than a direct discrepancy in assay performance, and should be interpreted with caution given differences in study design. Collectively, these findings support the reproducibility of NPA and the overall comparability of both assays for the detection of infectious vaginitis across different study settings and sample collection approaches, despite some variability in PPA between cohorts.

However, even among distinct molecular assays with comparably high diagnostic accuracy, such as the Aptima Vaginitis Panel and Seegene Allplex assay, there are practical considerations that may influence the selection of one assay over another for a given laboratory. For example, use of a non-proprietary system, such as the Σ-TRANSWAB, allows a single sample to be deployed across multiple distinct assays, rather than requiring dedicated swabs for each test. This approach may help to reduce the need for repeat sampling and could improve overall efficiency in diagnostic settings. While the Seegene Allplex has previously been validated using a non-proprietary sample collection method [27], our study provides further evidence that the Aptima Vaginitis Panel performs consistently with both proprietary and non-proprietary sample collection methods for the detection of vaginitis [37]. These findings support the compatibility of Aptima with flexible sampling workflows already achievable with multiplex molecular platforms, such as Seegene Allplex, while also suggesting that laboratories may be able to select their preferred swab system when using the Aptima Vaginitis Panel without compromising diagnostic breadth across the spectrum of vaginal health testing. Notably, use of non-proprietary swabs with the Aptima Vaginitis Panel requires an additional step, whereby samples must be transferred into specimen transfer media prior to processing on the Panther detection system, which may introduce extra handling time in high-throughput settings.

Another key distinction between the Aptima Vaginitis Panel and the Seegene Allplex assay that may influence assay selection is the type of data output. The Aptima Vaginitis Panel provides a binary output (‘positive’ or ‘negative’), which may reduce interpretive complexity. In contrast, the Seegene Allplex assay provides detailed quantitative outputs, including Ct values and PCR curves, which may capture subtle shifts in vaginal microbiota, though these outputs require additional training and staff input to guide interpretation. While Seegene’s more nuanced output may be informative in research settings, in routine high-throughput diagnostic practice, the added complexity may not translate to improved patient management, potentially making binary output a more practical option for routine workflows in some laboratory settings. However, this simplicity may come at the cost of the additional biological detail captured by quantitative outputs, such as Seegene’s, which may be valuable in specific clinical or research contexts.

Aptima’s binary approach also does not include the intermediate category, unlike conventional Nugent scoring and the Seegene Allplex assay, each of which defines its own intermediate classification. While intermediate classifications may help to identify transitional or borderline states of vaginal dysbiosis, the clinical interpretation of these categories remains uncertain, which may contribute to diagnostic uncertainty [39]. In our study, we observed very limited overlap between samples classed as intermediate by Seegene versus Nugent score (1.5%), in line with previously published findings [27]. Though this likely reflects methodological differences between morphologic and molecular approaches to BV classification, it also highlights the lack of consistency in how intermediate categories are defined across diagnostic systems. Therefore, our observations suggest that the intermediate category may be highly assay-dependent, potentially contributing to limited interpretive consistency in routine clinical practice and the need for additional diagnostic review within laboratories.

Species identification is another factor to consider when selecting the most appropriate assay. As discussed throughout, the Seegene Allplex assay is capable of detecting the presence of Mobiluncus species, which have previously been reported to be associated with BV [41–43], whereas this marker is not included in the Aptima BV assay. As such, the broader microbiome profiling offered by Seegene could potentially detect atypical cases of BV that Aptima may miss, as well as guide antibiotic choice in recurrent/resistant cases. However, we detected a relatively low frequency of Mobiluncus positivity in our study. Moreover, BV-positive samples by Seegene, Aptima BV or Nugent score were more commonly Mob^-ve than Mob^+ve. It is also worth noting that in our study, the capability to detect Mobiluncus did not confer any additional benefits relating to a gain in sensitivity for the detection of BV for Seegene versus Aptima. Collectively, our data suggest that this marker is not necessary for the detection of BV, though further large-scale studies are warranted to verify our findings.

As with all studies, ours was subject to some limitations. Most pertinent to our overarching comparison is that this study was not powered for the detection of statistical differences between paired data groups, meaning that any observed numerical differences in diagnostic accuracy should be considered exploratory in nature and interpreted with appropriate caution. Furthermore, no discrepancy analysis was performed on discordant results, though both molecular assays demonstrated strong overall agreement with SOC methods, supporting their clinical utility. A particular strength of our study is that, to the best of our knowledge, it provides one of the first direct comparisons of the overall clinical performance of the Aptima Vaginitis Panel with the Seegene Allplex assay for the detection of both BV and CV. We also demonstrate these findings in a relatively large cohort.

Conclusions

Both the Aptima Vaginitis Panel and the Seegene Allplex assay demonstrate high diagnostic performance for the detection of BV and CV. Relative to SOC, performance was generally comparable between assays. Notably, the rarity of TV cases encountered during the study period meant that neither assay’s performance for this pathogen could be meaningfully assessed here, and dedicated studies in populations with higher TV burden remain needed. The Aptima Vaginitis Panel maintained its robust performance when using a non-proprietary Σ-TRANSWAB for sample collection, as opposed to Aptima’s proprietary sample collection kit, highlighting its clinical versatility and applicability across diverse laboratory workflows. Beyond diagnostic accuracy, practical factors such as sample collection flexibility, data output format and handling of intermediate or borderline results are likely to influence assay selection depending on laboratory context and workflow priorities. Larger, multi-site studies would help to confirm the generalizability of our findings and further clarify how these practical considerations translate into clinical and operational outcomes.

Supplementary material

Abbreviations

BV

bacterial vaginosis

CE

European conformity

CI

confidence interval

CV

Candida vaginitis

DNA

deoxyribonucleic acid

FN

false negative

FP

false positive

Mob ^-ve

Mobiluncus-negative

Mob ^+ve

Mobiluncus-positive

NPA

negative percent agreement

NPVs

negative predictive values

PCR

polymerase chain reaction

PPA

positive percent agreement

PPVs

positive predictive values

RNA

ribonucleic acid

rRNA

ribosomal ribonucleic acid

RT

real-time

SOC

standard-of-care

STIs

sexually transmitted infections

TMA

transcription-mediated amplification

TN

true negative

TP

true positive

TV

Trichomonas vaginitis