Introduction
Candida spp. vulvovaginitis, or vulvovaginal candidiasis (VVC), considered the second cause of genital infection among women in menacme [
1], occurs by excessive multiplication, favored by predisposing factors of yeast, found in the vaginal microbiota of women in reproductive phase [
2,
3].
Despite the high prevalence and a large number of risk factors (RFs) associated with infection, the pathogenic VVC mechanisms and recurrent vulvovaginal candidiasis (VVCR) have not yet been fully elucidated [
4].
Candida yeasts migrate from the lower gastrointestinal tract to the adjacent vestibule and vagina, a route similar to that taken by vaginal
Lactobacillus species [
5‐
7]. The vaginal colonization, as well as yeast growth and germination, is enhanced by an estrogen-rich environment, which occurs in menacme, combined oral contraceptive use, and hormone replacement therapy [
6].
Besides the estrogen-rich environment, other risk factors contribute to VVC development, including genetic factors determining host susceptibility to infection, inflammatory response development, vaginal microbiota dysbiosis [
3,
8], sexual activity, hygiene and clothing habits, diseases such as diabetes mellitus [
4], and atopy [
9‐
11].
Among VVCR risk factors listed, vaginal microbiota dysbiosis, defined as an imbalance in this microbiota, has been most recently described as directly associated with infection pathogenesis. Vaginal microbiota (VMB) plays an important role in preventing colonization by pathogens [
12] and is affected by internal and external factors including hormonal changes (estrogen), menstruation, intestinal microbiota (near the rectum), intimate hygiene habits, sexual interaction (sharing with a partner), and contraceptive use (COC) [
13].
In this context, studies have shown that
Lactobacillus spp., in particular
L.
crispatus,
L.
jensenii, and
L.
gasseri, are predominant in healthy VMB [
14], acting through different mechanisms, and they form an important defensive barrier against
Candida spp. infection [
15‐
20].
Concerning diagnosis, although the clinical presentation—characterized by discharge, intense itching, and burning, accompanied or not by vulvitis with edema and fissure presence—is frequent in VVC, these symptoms are not specific and may be associated to other clinical conditions, such as desquamative vaginitis, cytolytic vaginitis, aerobic and anaerobic vaginitis, trichomoniasis, gonorrhea, and allergic symptoms [
21].
Fresh examination is based on direct microscopy, although an important diagnosis means fungal infections [
22] have low sensitivity, varying from 40 to 70%, because the result is dependent on microscopist experience [
6]. In recurrent cases, culture is recommended; however, despite allowing species identification, it has low sensitivity [
6,
23].
More recently, with the molecular age advent, methods employing DNA for fungal identification have been used which are able to provide accurate and faster results, demonstrating > 90% sensitivity in identifying
Candida species [
24‐
27].
Another alternative is mass spectrometry, an analytical technique that consists of the atoms or molecule ionization, their separation according to their mass/charge ratio (
m/
z), and then their identification and quantification [
28]. At present, one of the most widely used methods for biomolecule analysis is matrix-assisted laser desorption ionization (MALDI), followed by detection in a time-of-flight (TOF) analyzer, MALDI TOF [
29,
30]. MALDI TOF has been recognized as a fast and reliable tool for accurately identifying yeasts (more than 90% species identification specificity) since the spectra generated at identification are unique signatures of each microorganism [
31‐
33].
Using these more accurate diagnostic methods, it has been possible to evaluate species prevalence involved in VVC etiology. Although about 80 to 90% have
Candida albicans isolated as the etiological agent [
2,
34‐
38], non-
albicans (NA) species, such as
Candida glabrata,
Candida krusei,
Candida parapsilosis, and
Candida tropicalis, are no longer considered commensal species to be understood as etiological VVC agents [
3,
4,
39] which may be responsible for up to 30% of VVCR episodes [
21,
40].
The more accurate determination of
Candida species is particularly important in VVCR cases, whose prevalence demonstrated by some authors ranged from 5 to 10% [
41‐
43]. A recent systematic review about overall VVCR prevalence has estimated that approximately 138 million women are affected annually, which corresponds to 3871 cases per 100,000 women/year and about 372 million women will be affected by life-long VVCR [
44]. In this context, it is noteworthy that the recurrence prevalence obtained by prospective studies with laboratory diagnosis is unclear since most studies calculate the rates based on epidemiological questionnaires applied to patients, whose symptoms are self-described and can be confused with similar symptoms from another vaginitis.
Thus, facing a problem affecting women from different ethnicities worldwide, every effort should be made to study VVC and VVCR. This infection is very prevalent, and causes great discomfort, especially in the recurrent form, limiting the quality of life, affecting mental health and sexual life [
45]. In this sense, this study aimed to identify the etiologic agent of infections as accurately as possible, characterizing new RFs associated with infection and evaluating recurrence with laboratory diagnosis.
Materials and methods
Study design and ethics
The sample was composed of 278 women attending gynecology outpatient clinics (Brasília, DF, Brazil). Of these, 105 were asymptomatic and 173 had at least one vulvovaginitis symptoms, such as discharge, itching, burning, initial dyspareunia, dysuria, or malodor.
This study was conducted according to ethical principles (Declaration of Helsinki). All subjects signed an informed consent agreement to participate in this study, which was previously approved by the Research Ethics Committee (Faculdade de Medicina, Universidade de Brasília (CEP-FM-UNB), protocol number 1,572,449).
Epidemiological, clinical, and laboratorial screening
Epidemiological and clinical
At the consultation, the researcher conducted an interview with the patient and completed the symptom and risk factor questionnaire (Table
1), and during the physical examination realized by an experienced gynecologist, a vaginal discharge sample was collected, in a swab with Stuart transport medium, and was sent for laboratory identification.
Table 1
Symptom and risk factor questionnaire
Demographics | Age |
Marital status |
Presence at of least one symptom | Yes |
No |
Number of similar episodes in the last 12 months | ≤ 1 |
≤ 3 |
≥ 4 |
Symptoms (answers Yes or No) | ≥ 4 |
Medication use in the last 30 days | Discharge, itching, burning, dyspareunia, foul odor, dysuria |
Antibiotic, antifungal, or other |
Antecedents and habits |
Antecedent obstetric | Number of pregnancies, number of births and number of abortions |
Contraceptive method | Combined hormone, other (IUD, condom, other) |
Time with current sexual partner | ≤ 6 months |
6 months to ≤ 1 year |
≥ 1 year |
NA |
Number of sexual partners in the last 12 months | 0 |
1–2 |
3–4 |
Medication use, diseases (answers Yes or No) | Diabetes, use of oral hypoglycemic drugs, use of corticosteroids or immunosuppressants |
Personal habits (answers Yes or No) | Smoking, regular condom use, vaginal shower, daily panty protector |
Milk ingestion and derivatives | < 1 portion per day |
≥ 1 portion per day |
Gut transit | Normal and altered (constipation or diarrhea) |
Phenotypic identification of Candida species
For the phenotypic identification, the germ tube test (GTT) and the presumptive identification in a chromogenic medium (CM) were performed.
GTT was performed to differentiate Candida albicans (GTT positive) from non-albicans Candida species (GTT negative). Four pure fresh colonies subcultured on SDAC for 24 h at 35 °C were inoculated into 0.5 mL of human serum in test tubes (5 mL) and incubated at 35 °C for 3 h. Results were observed by an optical microscope (×400) using slide and coverslip preparations. Positive results were considered as isolates that presented elongated projections, called “Germ Tubes.” A reference American Type Culture Collection (ATCC) strain of Candida albicans (ATCC 90028) was employed as a positive control.
CM was used for presumptive identification of Candida albicans, Candida tropicalis, and Candida krusei. These Candida species are differentiated from the production of differently colored colonies due to reaction with the chromogenic substrates present in the medium. From subcultures on SDAC for 24 h at 35 °C, pure and fresh colonies were obtained and inoculated at CM with handle and then incubated for 24 to 48h. Plate reading was performed by visual means, where colony presence of specific color for each Candida species was observed. As a positive control, reference strains of Candida albicans ATCC 90028, Candida tropicalis ATCC 28707, and Candida krusei ATCC 34135 were used.
Identification using matrix-assisted laser desorption ionization time-of-flight
Mass spectrometry technique (MALDI TOF) use, a chemotaxonomic method for phenotypic identification of fungi, allowed the confirmatory identification of Candida spp.
The methodological technique principle uses whole yeast cells, which have molecular biomarkers, such as peptides or ribosomal proteins that are detected in the spectra generated after equipment reading. These protein spectra are characteristic of each fungal species, and functioning as the fingerprint of each microbial species, which are then compared to spectra deposited in databases that allow identification by comparability [
46,
47].
Spectra were analyzed using VITEK MS® Knowledge Base (BioMérieux, Marcy-l’Étoile, France) version 3.0. For comparing peaks formed with the characteristic patterns of the microorganism species, genus, or family were performed, resulting in its identification. Results were considered valid when the percentages of identification probability were equal to or higher than 90.0% [
48].
VVCR evaluation
During the study period, forty-one symptomatic patients were evaluated for recurrence, with the researcher reapplying the symptom and risk factor questionnaire and collecting another sample for laboratory evaluation.
Statistical analysis
Statistical tests were performed using GraphPad Prism version 8.00 software for Windows (La Jolla, CA, USA). To analyze VVC risk factors, Fisher’s exact test was used to verify if there was a correlation between VVC and the presence of each risk factor surveyed. The data were considered statistically significant when p ≤ 0.0500. In calculating positive predictive value (PPV), equation PPV = (a / (a + b)) was used, where a refers to patients with positive laboratory results, here called true positives, and b, false positive results.
Discussion
The diagnosis of VVC is very often based on the presence of characteristic symptoms. These diagnoses are used not only in clinical practice but also in most epidemiological studies on the prevalence of the disease. However, some studies indicate that, due to the low specificity, diagnosis based only on the presence of symptoms is very imprecise [
6,
49] and, in the case of a self-reported diagnosis, it can be wrong by up to 90% [
33,
50].
Despite this data, in practice, the treatment of VVC in most cases is based on clinical diagnosis, and there are few studies that assess the predictive value of symptoms compared to laboratory diagnosis. In the present study, the laboratory diagnosis confirmed by MALDI TOF, a highly sensitive method, was compared to the symptoms reported by patients, and, only about half of the symptomatic patients had positive laboratory results. Even when the classic triad of symptoms was reported, the PPV of symptoms to diagnosis VVC was also low confirming that similar symptoms appear in vulvovaginitis to other etiologies, as well as in cytolytic vaginitis, aerobic vaginitis, trichomoniasis, vulvar dermatosis, and atrophic vaginitis [
49,
51‐
57], which reinforces the need for laboratory diagnosis.
With respect to VVC-related risk factors, those frequently described in previous studies, such as oral contraceptive use, condoms, vaginal douche, and daily panty liners [
4,
6,
58], were evaluated and, as well, those mentioned less frequently (respiratory and skin allergies) [
10]. In addition to these risk factors already described, the study presented two new factors still little explored, namely the frequency of milk and dairy products, and changes in bowel habits. These factors were studied based on the current understanding that vaginal dysbiosis, defined as vaginal microbiota imbalance, may be directly implicated in the infection pathogenesis.
The laboratory method for identifying
Candida spp. was MALDI TOF, with more than 90% specificity [
32,
33,
59]. This method requires expensive equipment, not always accessible for diagnosis. The realization of classic laboratory methodologies such as GTT and CM, although currently uncommon, allowed an assessment of the sensitivity and specificity of each of them concerning the gold standard (MALDI TOF). This result makes it possible to carry out laboratory diagnostics in laboratories that do not have the MALDI TOF methodology available to use, which would justify the use of GTT and CM methodologies in the study, and would possibility the use of these in similar studies, with knowledge of the expected sensitivity and specificity in symptomatic and asymptomatic patients.
Another factor to be highlighted is that the CM in addition to permit the division into two distinct groups, the albicans species of the non-albicans, enables the identification of more than one species of Candida spp., when present in the same biological sample, which suggests the existence of VVC by more than one species, and does not represent contamination of the analyzed biological sample (data are not shown).
For many authors,
C.
albicans is considered a VMB and IMB commensal [
51,
53,
60], and the onset of symptoms is attributed to a dysbiosis that results in excessive
Candida growth [
6,
11]. However, data on asymptomatic colonization are quite limited. A study in Australia using CM for culture found a prevalence of
Candida spp. 21% [
61], while in the present study, the prevalence was 12.4%.
This species is also the etiological agent of most cases of VVC [
2,
33,
35‐
38,
62,
63]. Other authors pinpoint other species, such as
Candida glabrata,
Candida krusei,
Candida parapsilosis, and
Candida tropicalis as emerging [
3,
4,
38]. In this study, the prevalence of
Candida albicans (80.9%) and non-
albicans (18.1%) was similar to previous studies. An unrelated yeast,
Rhodotorula mucilaginosa, was also isolated. This species, considered non-virulent and saprophytic, was found in the genital region of 13.5% of asymptomatic people [
64]. However, new studies show that it can be opportunistic [
65,
66]. Another study identified this yeast in 10% of vaginal samples from pregnant women, 3.8% of these symptomatic [
67]. Although the prevalence of this species was very low here, it confirms the existence of VVC caused by this species.
VVC recurrence data based on prospective studies, including laboratory analyses of samples from the same patient collected at different times, are uncommon. Most studies assess recurrence based on questionnaires, such as a study that includes data from several countries, which indicated a VVCR of 9% [
44]. Another survey found a VVCR of 34% among women who answered a questionnaire [
50]. In this study, the examination of the same patient who remained symptomatic, performed at different times, revealed a recurrence rate of 28.6%, and 33.3% for non-
albicans species.
Some RFs that already described for VVC [
4,
6,
10,
68] were evaluated. Confirming previous studies, the use of COC was greater when compared to controls. The use of a vaginal douche and condom did not differ between groups, perhaps due to the low prevalence of these habits.
The daily use of panty liners has been considered a risk factor for VVC [
69,
70], but two reviews have not confirmed this association [
71,
72]. In contrast, in this study, the use of panty liners was higher in the VVC group (
p 0.0003). However, it cannot be affirmed that using panty liners is a risk factor, since women with VVC may be more likely to use them due to the presence of abundant secretion.
Some studies have linked the co-existence of VVC to respiratory allergies (RA), although they do not explain a possible mechanism [
9,
73,
74]. The VVC group reported RA more frequently in the present study (
p 0.0042). The immune response in individuals with RA is mediated by Th17, which increases IL-10, an anti-inflammatory cytokine that transfers the immune response (IR) to a Th2 pole [
75]. Some components of the cell wall of glucan-rich fungi negatively modulate IR, displacing it to Th2, favoring the persistence of fungi, and reducing microbicidal effects on it [
76]. Thus, an individual with RA has more IL-10, and this cytokine-mediated IR is ineffective in eliminating fungi like
Candida spp., suggesting that this atopic individual would be more susceptible to VVC.
The study also presented two new RFs, frequency of ingestion of milk and dairy products, and changes in bowel habits. These factors were studied based on the current understanding that vaginal dysbiosis may be implicated in infection. VMB plays an important role in preventing colonization by pathogens and acts in maintaining women’s reproductive and gynecological health [
12]. In addition, it is known that there is a correlation between IMB and VMB, including the terminal portion of the intestine as a source of colonization of the vagina by the migration of
Lactobacillus spp. and yeasts [
6,
51,
77].
There is a peak of lactase expression in the small intestine villi at birth, and this expression decreases with age, decreasing the absorption of lactose, which in contact with the IMB is fermented, producing H
2, CO
2 and methane (CH
4), and short-chain fatty acids, which alter intestinal function [
51]. Considering the most prevalent age group in the study, this deficiency in lactose absorption may result in an imbalance of the IMB, favoring the growth of yeasts, with interference in the VMB, given the correlation between the microbiome of the two sites.
Patients were asked about their bowel habits and classified by Rome IV consensus [
71,
72]. In the VVC group, there was a higher frequency of changes in intestinal transit (
p 0.0002). Studies show that, in individuals with intestinal constipation, bacteria of the genera
Bifidobacterium and
Lactobacillus decrease [
48,
77‐
79]. Thus, the decrease of these species in the intestine, especially
Lactobacillus spp., which have a recognized role in maintaining vaginal health, possibly reduces their presence in the vagina, since the VMB is fed-back by the IMB.
Our study makes new contributions to the diagnosis and clinic of VVC. Laboratory diagnosis of VVC is essential for treatment based on the identification of the Candida species, since non-albicans species are present in many of these infections and some of them have intrinsic or acquired resistance to azoles, and this therapy would be ineffective against these pathogens. In addition, there is a need for this diagnosis to rule out the presence of another non-Candida vaginitis, which share the same classic symptoms of VVC (discharge, itching, and burning), the most frequently cited in the clinic, and which in our study proved that they did present a low PPV in the diagnosis of VVC in these patients.
The existence of new risk factors associated with CVV, such as milk and dairy products and bowel habits, in addition to the use of panty liners and the presence of respiratory allergies, is relevant to the diagnosis of CVV, and may be related to the presence of dysbiosis of vaginal and/or intestinal. And the inclusion of these risk factors can add information to the clinical anamnesis, especially in patients who have the recurrent form of candidiasis (VVCR).
Strengths and limitations of the study
Limitations in our study include the lack of a screening test, such as the Nugent criterion that could be performed using Gram stain, which would allow us to diagnose the presence of non-Candida vaginitis, in the absence and/or presence of CVV, making it possible to prove the existence of a single or mixed infection. Regarding the analysis of risk factors, the presence of a few patients using a vaginal shower and condom limited the analysis of these factors associated with CVV in the study.
The major strength of our study was to show the importance of laboratory diagnosis for the correct identification of CVV, and to compare it with the prevalence of the literature, which is almost always based on self-diagnosis in questionnaires answered by patients who describe the main symptoms of CVV. The study identified that the PPV based on the presence of symptoms is low when confirmed by the laboratory diagnosis, emphasizing the importance of the laboratory diagnosis for the clinical diagnosis. In addition, the description of new risk factors may contribute to the physician’s clinical conduct in conducting CVV treatment, namely the frequency of milk and dairy products, and changes in bowel habits, which may contribute to the understanding of vaginal dysbiosis.