Clinical evaluation of Candida albicans enolase 1 as a novel biomarker for vulvovaginal candidiasis: a pilot study demonstrating rule-out capability

Abstract

Purpose

Vulvovaginal candidiasis (VVC) affects approximately 75% of women during their lifetime, yet current diagnostic methods have limitations. This pilot study evaluated Candida albicans enolase 1 (CaEno1) as a diagnostic biomarker for C. albicans-associated VVC, with a view to informing future point-of-care test (POCT) development.

Methods

A total of 131 vaginal secretion samples were collected from patients presenting with vulvovaginal symptoms at a gynecology department. Samples underwent routine microscopic examination, fungal culture, and CaEno1 detection using a previously developed double-antibody sandwich ELISA (DAbS-ELISA) with sensitivity of 0.33 ng/mL. ROC curve analysis determined optimal diagnostic performance parameters.

Results

Among 131 samples, 33 (25.2%) showed positive fungal growth, with C. albicans predominating (72.7% of positive cultures). CaEno1 levels were significantly higher in C. albicans-positive samples compared to negative samples (Mann-Whitney U test, P < 0.0001). ROC analysis yielded an area under the curve of 0.727 (95% CI: 0.613–0.842). At an optimal cutoff value of 0.052 optical density units at 450 nm (OD450), the assay demonstrated 70.8% sensitivity, 78.5% specificity, 42.5% positive predictive value (PPV), and 92.3% negative predictive value (NPV).

Conclusion

This study provides the first clinical evidence that CaEno1 can be detected in vaginal secretions and suggests potential rule-out utility for Candida-associated VVC in settings where C. albicans predominates. The favorable NPV observed in this cohort supports its potential use as a screening tool in comparable clinical settings, though cross-reactivity with non-albicans species warrants cautious interpretation alongside clinical findings.

Background

Vulvovaginal candidiasis (VVC) is a fungal infection caused by Candida species, particularly Candida albicans, that affects women of reproductive age. The condition is prevalent among women, with approximately 75% experiencing at least one episode during their lifetime [1–3]. Diagnosing VVC in a laboratory setting is straightforward, as multiple detection techniques can be effectively employed. Commonly used methods include microscopic examination and pH measurement of vaginal secretions. While DNA-based assays have emerged as valuable adjuncts to culture techniques, their widespread adoption remains limited [4, 5]. This is due to stringent requirements for laboratory instrumentation and challenges in standardizing fungal DNA sample preparation [6].

While laboratory-based microscopy and fungal culture remain effective and widely used diagnostic tools in well-equipped clinical settings, a significant proportion of women with VVC symptoms are managed empirically without microbiological confirmation, particularly in primary care, pharmacy, and self-care settings where laboratory infrastructure is unavailable [7]. This diagnostic gap contributes to inappropriate antifungal use and potential resistance development [8, 9]. The development of a POCT reagent for VVC would address this specific unmet need by enabling objective, accessible testing outside the traditional laboratory environment. Consequently, identifying an appropriate biomarker for VVC-specific POCT represents an important translational goal.

C. albicans enolase 1 (CaEno1) represents a pivotal and highly conserved enzyme within the glycolytic pathway, playing a significant role in C. albicans as a multifunctional protein [10, 11]. It mediates the adhesion of human cells and invasive infection [12, 13], as well as immune evasion through complement inactivation [14]. Notably, CaEno1 exhibits robust immunogenicity, capable of eliciting high titers of antibodies in the host and even activating lymphocytes through direct binding [15, 16]. In the context of laboratory diagnostics, Walsh et al. [17]. demonstrated in the 1990s that the detection of CaEno1 antigen serves as an efficient means for diagnosing invasive candidiasis. That work, however, was confined to immunocompromised patients with hematogenous dissemination and did not extend to mucosal surfaces. Whether CaEno1 is even detectable in vaginal secretions, let alone whether it holds diagnostic value for VVC in otherwise healthy women, has remained an open question. Consequently, both CaEno1 antigen and antibody detection are recognized as valuable strategies for diagnosing Candida infections.

Since CaEno1 is secreted by C. albicans and has been detected in other clinical specimens, it is reasonable to hypothesize that VVC secretions may contain measurable levels of CaEno1. Building on our previous development of a highly sensitive DAbS-ELISA (detection limit: 0.33 ng/mL) using in-house developed monoclonal antibodies [18], we conducted this pilot study to provide the first clinical validation of CaEno1 as a diagnostic biomarker for VVC. This represents a critical translational step from laboratory assay development to clinical application.

Materials and methods

Study design and subjects

This retrospective pilot study analyzed residual vaginal secretion specimens from 131 patients who had visited the gynecology department of Bethune International Peace Hospital between March 2020 and May 2020 due to genital discomfort. Specimens were selected based on the following pre-specified criteria. Inclusion criteria were: (1) female patients aged 18–65 years; (2) presenting with vulvovaginal symptoms; and (3) availability of adequate residual vaginal secretion sample for all planned analyses. Exclusion criteria were: (1) pregnancy; (2) menstruation at the time of sampling; (3) severe immunosuppression; and (4) insufficient sample volume. Relevant clinical data, including age and results of routine vaginal microscopy, were extracted from the hospital’s electronic medical records.

As this was an exploratory pilot study with no prior data available on CaEno1 levels in vaginal secretions, a formal a priori sample size calculation was not performed. The sample size of 131 was determined by the availability of residual specimens during the study period and is consistent with the scope of a pilot diagnostic study intended to generate preliminary estimates of assay performance for future power calculations.

Vaginal secretion examination

Macroscopic evaluation included visual inspection of discharge color, consistency, and odor. Vaginal pH was measured using test strips or a calibrated meter. Microscopic analysis involved saline/KOH wet mounts and Gram staining to detect pathogens. White blood cells (WBCs) were quantified per high-power field (HPF: occasional/<10, +/10–20, ++/20–30, +++/>30). Cleanliness scores (I–IV) were assigned based on epithelial cells, leukocytes, and flora balance, where grades I–II indicated normal lactobacilli dominance and III–IV reflected dysbiosis.

Fungal culture

The vaginal secretion sample was diluted with 0.5 mL sterile normal saline, thoroughly vortexed, and then 20 µL of the suspension was spread onto a Chromogenic Candida Agar (KOMA) plate. The plate was incubated at 35 °C with 5% CO₂ for 48 h. The colonies grown on the KOMA plate were counted, and the preliminary identification of Candida species was based on their colony morphology and color characteristics. Further confirmation of Candida species relied on the use of the API 20 C AUX Yeast Identification System (BioMerieux, Craponne, France).

Detection of CaEno1 with the DAbS-ELISA

After centrifugation of the vaginal secretion suspension prepared with normal saline, the supernatant was taken for DAbS-ELISA analysis as previously described [18]. The reproducibility of the DAbS-ELISA has been previously characterized: intra-assay coefficient of variation (CV) was < 8% and inter-assay CV was < 12% across the assay’s dynamic range, demonstrating acceptable precision for a research-grade immunoassay.

Briefly, 100 µL of in-house-developed monoclonal antibody 9H8 [10] (3 µg/ml in carbonate-bicarbonate buffer, pH 9.6) was coated in each well of a 96-well plate and incubated at 4℃ overnight. After washing the plate three times with PBST, 100 µL of PBST containing 3% BSA was added to each well and incubated at 37℃ for 30 min. Then, 50 µL of prepared vaginal secretion sample (or diluted standard) and 50 µL of 1:2000 diluted HRP-10H8 were added to each well and incubated at 37℃ for 45 min. After washing the plate three times, 3,3,5,5-tetramethylbenzidine solution (Beyotime, Shanghai, China) was added, and the plate was incubated at room temperature for 15 min. The reaction was stopped with 50 µL of 2 N H₂SO₄, and the OD₄₅₀ values were finally measured with the VersaMax plate reader (Molecular Devices, CA, United States). The results were analyzed using ELISACalc V0.2 software. We conducted a series of 20 consecutive tests to detect background levels. Samples with test values exceeding the background level + 3 standard deviations (SD) were considered positive.

Statistical analysis

Statistical analysis was performed with GraphPad Prism 7.00 and SPSS Statistics 22.0. Kappa coefficient was used to determine the consistency of fungal culture and CaEno1 detection. Mann-Whitney U test was used to analyze the difference between C. albicans CFU and CaEno1. ROC curve analysis was performed to evaluate diagnostic performance. The optimal cutoff value was determined using Youden’s index. P < 0.05 was considered significant.

Results

Baseline characteristics of the subjects enrolled in the study

A total of 131 vaginal secretion samples were collected and analyzed in this study. The mean age of the patients was 35.5 ± 12.0 years, with 81.7% of them being of reproductive age (20–49 years). Of the samples analyzed, 33 cases (25.2%) had a white blood cell count of “+++” or higher. The majority of the samples (111 cases, 84.7%) had a cleanliness score of degree III, while 13 cases (9.9%) had a score of degree IV. Additionally, 21 (16.0%) vaginal secretions contained fungal yeast cells or hyphae upon microscopic examination (Table 1).

Table 1.

Baseline data of the 131 patients included in the study

Characteristics	Valuesa
Age (years, mean ± SD)	35.5 ± 12.0
< 20	3 (2.3%)
20–29	49 (37.4%)
30–49	58 (44.3%)
> 49	21 (16.0%)
Routine vaginal secretion test
WBC count (n, %)b
+	35 (26.7%)
++	47 (35.9%)
+++	22 (16.8%)
++++	11 (8.4%)
Occasional	16 (12.2%)
Cleanliness (n, %)
II	7 (5.3%)
III	111 (84.7%)
IV	13 (9.9%)
Yeast (n, %)
Positive	21 (16.0%)
Negative	110 (84.0%)

^aPercentages are calculated with the total number of patients 131 as the denominator

^bWBC, white blood cell

Fungal culture results and CaEno1 antigen distribution

Among the 131 vaginal secretion samples analyzed, 33 (25.2%) exhibited positive fungal growth, with C. albicans being the predominant species, accounting for 72.7% of positive cultures (Fig. 1A). The remaining positive cultures included C. glabrata (n = 5, 15.2%), C. krusei (n = 2, 6.1%), and single isolates of C. tropicalis and C. parapsilosis (3.0% each).

Fig. 1.

Distribution of fungal species in positive cultures and CaEno1 antigen levels in vaginal secretion samples. A Species distribution among fungal culture-positive samples (n = 33). B Comparison of CaEno1 antigen levels between C. albicans-negative (n = 107) and C. albicans-positive (n = 24) groups. The C. albicans-positive group had significantly higher CaEno1 levels than the negative group (Mann-Whitney U test, ***P < 0.0001)

The distribution of CaEno1 antigen levels (OD₄₅₀) across all samples demonstrated a markedly right-skewed distribution, with an interquartile range of 0.000-0.092 (range: 0.000-1.695). Comparative analysis revealed substantially elevated CaEno1 antigen levels in samples with positive C. albicans cultures compared to C. albicans-negative samples (Fig. 1B). The C. albicans-positive group (n = 24) showed a median CaEno1 level of 0.109 (IQR: 0.000-0.191), while the C. albicans-negative group (n = 107) demonstrated a median level of 0.000 (IQR: 0.000–0.000). Mann-Whitney U test analysis confirmed this difference to be statistically significant (U = 715.5, P < 0.0001), indicating that CaEno1 antigen levels, as detected by the monoclonal antibody-based DAbS-ELISA, are significantly higher in culture-confirmed C. albicans-positive vaginal secretions than in culture-negative samples, though potential cross-reactivity with non-albicans Candida species cannot be excluded.

ROC curve analysis and diagnostic performance

ROC curve analysis was performed to evaluate the diagnostic performance of the CaEno1 ELISA (Fig. 2A). For specific detection of C. albicans, the area under the curve (AUC) was 0.727 (95% CI: 0.613–0.842), indicating fair discriminatory ability. Using Youden’s index to determine optimal cutoff points, the threshold of 0.052 OD450 units was identified for C. albicans detection, maximizing the sum of sensitivity and specificity.

Fig. 2.

Diagnostic performance of CaEno1 ELISA for vulvovaginal candidiasis. A ROC curve analysis showing AUC of 0.727 (95% CI: 0.613–0.842). Optimal cutoff: 0.052 OD450 units determined by Youden’s index. B Diagnostic performance at optimal cutoff: sensitivity 70.8%, specificity 78.5%, PPV 42.5%, NPV 92.3%, overall accuracy 77.1%

At the optimal cutoff units, the diagnostic performance for C. albicans showed: sensitivity of 70.8% (17/24; 95% CI: 48.9%-87.4%), specificity of 78.5% (84/107; 95% CI: 69.4%-85.9%), positive predictive value (PPV) of 42.5% (17/40; 95% CI: 27.0%-59.1%), and negative predictive value (NPV) of 92.3% (84/91; 95% CI: 84.8% -96.9%). The overall accuracy was 77.1% with a Cohen’s kappa coefficient of 0.392, indicating fair agreement (Fig. 2B). The 2 × 2 contingency table for C. albicans detection revealed 17 true positives, 84 true negatives, 23 false positives, and 7 false negatives (Table 2). The assay demonstrates better NPV than PPV, suggesting potential utility as a screening tool to rule out VVC when negative, but limited value for definitive diagnosis when positive.

Table 2.

Contingency table for CaEno1 ELISA versus fungal culture for C. albicans detection

Fungal culture	CaEno1 detection			Kappa
	Positive	Negative	Total
C. albicans positive	17	7	24
C. albicans negative	23	84	107
Total	40	91	131	0.392

Analysis of discordant results: insights into assay characteristics

To understand the clinical limitations of CaEno1 ELISA, we analyzed discordant results between the assay and fungal culture. The seven false negative samples had significantly lower colony-forming units compared to true positives (P = 0.0301), indicating detection threshold limitations for low-burden infections (Fig. 3A). Twenty-three samples showed false positive results. Cross-reactivity with other Candida species (C. glabrata, C. krusei, C. tropicalis, C. parapsilosis) accounted for 21.7% (5/23) of cases. Recent antifungal treatment within 7 days explained 43.5% (10/23) of false positives, likely due to residual antigen detection post-therapy. The remaining 34.8% (8/23) had no identifiable cause (Fig. 3B).

Fig. 3.

Analysis of discordant results in CaEno1 ELISA testing for C. albicans associated vulvovaginal candidiasis. A False negative samples (n = 7) showed significantly lower CFU counts compared to true positive samples (P = 0.0301). B Distribution of causes for false positive results (n = 23)

These findings highlight important clinical considerations: the assay may miss low-burden C. albicans infections, cross-react with non-albicans species, and be influenced by recent antifungal therapy. Clinicians should consider these limitations when interpreting CaEno1 ELISA results.

Discussion

This study represents the first investigation of CaEno1 as a diagnostic biomarker specifically for C. albicans-associated VVC and provides preliminary evidence supporting its potential as a foundation for POCT development. Our findings reveal that CaEno1 levels are significantly elevated in C. albicans-positive vaginal secretion samples (P < 0.0001). While CaEno1 has been extensively studied as a biomarker for invasive candidiasis since Walsh et al.‘s seminal work in 1991 [17], no previous study has evaluated its diagnostic utility for mucosal candidiasis such as VVC. Our findings extend the clinical application of this well-characterized biomarker to a new disease context affecting millions of women worldwide, representing a constructive translational step in VVC diagnostics research.

The diagnostic performance of our CaEno1 ELISA is broadly comparable to current first-line diagnostic methods for VVC. Microscopy, the most commonly used initial diagnostic tool, demonstrates only 43.9–70% sensitivity with 88.9% specificity [7]. While our assay shows moderate discriminatory ability, it addresses a gap in VVC diagnosis by providing objective, quantitative results that could reduce the subjectivity inherent in microscopic interpretation [19]. The optimal cutoff value of 0.052 OD450 achieved balanced sensitivity and specificity. From a clinical perspective, the performance characteristics of a diagnostic test determine its optimal application in patient care. Given the observed cross-reactivity with non-albicans species, the NPV of 92.3% suggests that a negative result may help exclude Candida-associated VVC, though clinical interpretation should consider local species prevalence and other clinical findings. It should be noted that, as with all predictive values, the NPV is prevalence-dependent; in settings with substantially higher VVC prevalence, the rule-out confidence afforded by a negative result will be correspondingly diminished.

The analysis of discordant results provides insights indicating that assay sensitivity may be insufficient for detecting low-burden C. albicans infections. This finding aligns with previous observations that enolase secretion correlates with fungal load and pathogenic activity [20]. Cross-reactivity with non-albicans Candida species accounted for 21.7% of false-positive results, involving C. glabrata, C. krusei, C. tropicalis, and C. parapsilosis. Therefore, a positive result does not rule out non-albicans infection, and caution is warranted in settings with high non-albicans species prevalence. Additionally, the 43.5% false positive rate associated with recent antifungal treatment suggests that residual antigen persistence may interfere with accurate diagnosis, requiring careful clinical correlation and consideration of recent treatment history when interpreting results. The significantly lower CFU counts in false negative samples indicate that the current assay sensitivity may be insufficient for detecting low-burden C. albicans infections.

The choice of fungal culture as the reference standard warrants discussion. Although PCR offers higher analytical sensitivity, it is not currently recommended for VVC diagnosis in major clinical guidelines [21, 22]. Published PCR validation studies for VVC themselves rely on culture as the reference standard, confirming its primacy in current clinical practice [23, 24]. These considerations support our use of culture as the reference standard, while we acknowledge that culture sensitivity varies with sample handling and laboratory conditions. The composite reference standard comprising culture, clinical symptom scoring, and microscopy represents the current recommended approach for VVC diagnosis per the CDC 2021 Sexually Transmitted Infections Treatment Guidelines [22]. Future multicenter studies could follow this standard to strengthen diagnostic evaluation across diverse settings.

The development of CaEno1-based diagnostics addresses several unmet clinical needs in VVC management. Current diagnostic delays often necessitate presumptive treatment, contributing to inappropriate antifungal use and potential resistance development [7, 25, 26]. A rapid, objective diagnostic test could enable targeted therapy, improve treatment outcomes, and support antimicrobial stewardship efforts. While the current DAbS-ELISA format is not itself a POCT, the underlying antibody pair and detection principle are compatible with adaptation to simpler formats. The detection limit of 0.33 ng/mL and simple sample preparation requirements are compatible with future development of lateral flow immunoassay, which has proven successful for other infectious disease applications [27, 28]. Recent advances in smartphone-based detection systems and artificial intelligence-enhanced result interpretation represent future possibilities that may further simplify deployment in resource-limited clinical settings [29, 30], pending validation in the VVC context. If successfully adapted to a lateral flow format, such a platform could potentially enable a turnaround time substantially shorter than the 24–48 h required for culture-based methods, though this would require dedicated assay development and performance validation.

This study has several limitations. It was a single-center study with a modest sample size, drawn from one tertiary gynecology department in Hebei Province, China. This sample size precluded multivariate analysis of confounding factors such as antifungal exposure, fungal burden, and species composition. Species identification used the API 20 C AUX system instead of the more accurate MALDI-TOF mass spectrometry, potentially misclassifying a small number of non-albicans isolates. Our observed C. albicans prevalence of 72.7% among culture-positive samples may not generalize to other regions or clinical settings. Therefore, external validation in multicenter cohorts is needed before clinical implementation. Technically, the current ELISA format requires laboratory infrastructure and trained personnel; adapting it to a POCT format would require separate assay development and validation. Moreover, although the specificity of the 9H8/10H8 antibody pair was validated against C. albicans lysates by Western blot and LC-MS/MS [10], systematic cross-reactivity testing against non-albicans Candida species was not performed, and the clinical cross-reactivity observed here suggests that species-level specificity may be limited in complex biological matrices, particularly in settings where non-albicans species are more prevalent than in our cohort. Finally, the moderate AUC of 0.727 suggests that diagnostic performance requires further improvement.

Several strategies may be considered to address these limitations. Assay sensitivity for low-burden infections could potentially be improved by signal amplification or by combining CaEno1 detection with complementary biomarkers such as mannan antigen in a multiplex format. Dedicated cross-reactivity characterization against a panel of non-albicans Candida isolates is warranted; if cross-reactivity is confirmed, the assay may be more accurately positioned as a broad Candida-associated VVC detection tool, which could broaden rather than undermine its clinical utility given the rising prevalence of non-albicans VVC [31]. Finally, the expected decline in CaEno1 levels following antifungal treatment could potentially be exploited for therapeutic monitoring, with serial measurements serving as a surrogate marker of treatment response.

Conclusion

This study provides the first clinical evidence that CaEno1 is detectable in vaginal secretions and suggests potential rule-out utility for Candida-associated VVC in settings where C. albicans predominates. These findings suggest its potential as a screening tool in symptomatic women, potentially reducing unnecessary antifungal treatment and supporting antimicrobial stewardship efforts. While further optimization to enhance sensitivity and species specificity remains necessary, this work provides a biomarker validation foundation that could inform the future development of accessible diagnostic tools, including potential adaptation to POCTs in resource-limited settings where laboratory-based diagnosis is unavailable.

Abbreviations

CaEno1

Candida albicans enolase 1

CFU

Colony-forming units

DAbS-ELISA

Double-antibody sandwich enzyme-linked immunosorbent assay

POCT

Point-of-care testing

HPF

High-power field

NPV

Negative predictive value

PPV

Positive predictive value

SD

Standard deviation

VVC

Vulvovaginal candidiasis

WBC

White blood cell

Authors’ contributions

**HB Meng & ZX He** conceived and planned the experiments. **HB Meng, HM Xing & QQ Sun** carried out the experiments. **F Li** contributed to sample preparation. **HB Meng, HM Xing & ZX He** contributed to the interpretation of the results. **ZX He** took the lead in writing the manuscript. All authors provided critical feedback and helped shape the research, analysis and manuscript.

Funding

This work was supported by grants C2021505001 from the Natural Science Foundation of Hebei province, China.

Data availability

The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was conducted in accordance with the Declaration of Helsinki and approved by the Medical Ethics Committee of Bethune International Peace Hospital (approval number: 2019-KY-23). Written informed consent was waived by the ethics committee due to the retrospective nature of the study using residual clinical specimens that would otherwise be discarded, with no additional risk to patients.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.