Recurrent Vulvovaginal Candidosis and Cluster Analysis of Clinical Signs and Symptoms: A Laboratory-Based Investigation

Recurrent vulvovaginal candidosis (RVVC) represents a major health problem that significantly affects a patient’s quality of life (QoL). This infection presents with a plethora of clinical manifestation, and this is the first study that carries out a cluster analysis of these signs and symptoms (SS). The goals are to evaluate the distribution of species causing RVVC, their in-vitro susceptibility to antifungals, and the patient’s QoL. Additionally, the clinical characteristics are analyzed using cluster analysis. Prospective analysis of data was performed for women diagnosed with RVVC in the period from January 2016 to December 2019 based on the analysis of data from a single-center’s records. The standard mycological methods and antifungal susceptibility testing were done. Clinical characteristics and QoL were examined by appropriate questions. The cluster analysis was used to identify clusters of SS. A total of 320 women were diagnosed. The dominant species was Candida (C.) albicans. Non-albicans Candida (NAC) yeast was found in 24.4%, and the most common was C. glabrata. Interestingly, Saccharomyces (S.) cerevisiae was detected in 2%. All of the isolated species, except C. parapsilosis and C. kefyr, demonstrated reduced susceptibility to antifungals. We confirmed the emergence of the NAC species and S. cerevisiae with reduced susceptibility to antifungals. Cluster analysis represented by a dendrogram revealed three SS clusters: irritation, uncommon, and discharge, but further studies are needed to examine the relationship between clusters, Candida strains, and outcomes.


Introduction
Vulvovaginal candidosis (VVC) is one of the most common infections in women, with prevalence between 15% and 30% [1], though even higher prevalence rates were noted in some countries [2]. Furthermore, it is assumed that 70-75% of women suffer at least one episode of VVC in their lifetime [1]. All this is understandable since molecular studies have proven that vulvovaginal Candida-colonization is present in up to 64% of women [3,4].
Another problem is the fact that about 9% of the female population suffer from a recurrent form of VVC [5], which is defined as more than three symptomatic Candida genital infections over a one-year period [1]. The recurrent vulvovaginal candidosis (RVVC) requires more attention from the gynecologists, with an individual approach to each woman. Moreover, it is desirable to perform mycological analyses with antifungal susceptibility testing for every case [6,7]. In addition, contributing factors, namely, intestinal Candida-overgrowth and sexual transmission, should also be excluded.
Clinically, VVC can be manifested as a nonspecific genital infection with a plethora of signs and symptoms (SS) such as discharge, itching, burning, and erythema. However, it can also be a complicated condition, such as RVVC, characterized by chronicity or pronounced SS [8].
It is generally known that predominant species causing this infection is C. albicans [6]. However, in recent years, non-albicans Candida (NAC) species, such as C. glabrata or C. krusei, have been increasingly recognized as the cause of both sporadic and recurrent VVC [9,10]. Moreover, in the literature, there is evidence that yeast S.cerevesiae could be the causative agent of genital fungal infection as well [11,12].
Having in mind the impact of RVVC on global health, the increasing prevalence of this infection in women, as well as the emergence of Candida-resistance to commonly used antifungal drugs, we analyzed women with RVVC in order to determine the most frequent causative agents of RVVC and in-vitro susceptibility to antifungals. Additionally, using cluster analysis, SS of women with proven RVVC were evaluated for the first time.

Materials and Methods
The research was designed as a single-center prospective study. In the period from 2016 to 2019, women with four or more clinically and laboratory-confirmed episodes of fungal genital infection in the previous year were included in the study in their relapse period. Vaginal swabs were analyzed both bacteriologically and mycologically. In addition, a parasitological examination was performed.
The isolation and identification of yeast were done using standard mycological procedures. Fungal growth material was inoculated in Sabouraud dextrose agar (SDA; Liofilchem Diagnostici, Roseto degli Abruzzi, Italy) and chromogenic Candida media (Liofichem/Bacteriology products, Roseto degli Abruzzi, Italy) (both incubated at 37 • C for up to 7 days). Species of genus Candida were identified using the germ tube test, the chromogenic Candida media (Liofichem/Bacteriology products, Roseto degli Abruzzi, Italy), and Auxacolor TM (BioRad, Marnes-la-Coquette, France).

Clinical Characteristics and Quality of Life (QoL) Assessment
Data regarding clinical characteristics were collected by a gynecologist who filled in a semistructured questionnaire. All women with a laboratory-based diagnosis of RVVC and with SS of infection during the physician's examination were included in the study. The study of the clinical characteristics and QoL also involved 109 control subjects (mean age 30.73 ± 9.3) who were healthy women who came for a control examination without a history of RVVC, pregnancy, or use of oral contraceptives or corticosteroids.
The Serbian version of self-completion generic questionnaires EuroQol-5 Dimension (EQ-5D) and EuroQol-VAS (EQ-VAS), developed by the EuroQoL Group [13], was used for the assessment of a patient's QoL. The modified EQ-5D was used for the assessment of usual activities, pain/discomfort, and anxiety/depression, while instead of mobility and self-care, sexual problems and self-confidence were evaluated. Respondents described these five features as having no problems, some problems, or severe problems. Women rated their health status using EQ-VAS, a vertical, visual analog scale, where 0 points represent "the worst" and 100 points "the best imaginable" health states. This instrument provided us with a quantitative measure of the patients' wellbeing.

Statistical Analysis
Continuous data are presented as arithmetic mean ± standard deviation (SD), and categorical variables are expressed as number (n) with percentage (%). The t-test was used for comparison of the continuous variables between the groups. Intergroup comparisons of categorical variables were made using the chi-square test. The agglomerative hierarchical cluster analysis with Ward's method was used to identify SS clusters, with the Euclidean distance being used to measure the similarity between variables. Clustering results are displayed with the dendrogram. Women suffering from RVVC were allocated to one of the three groups (having all of the SS, having some of the SS, or having none of the SS of that cluster) for each of the established SS clusters. All statistical analyses were performed using R software (version 3.4.3; The R Foundation for Statistical Computing, Austria) [14]. The significance level was set to p < 0.05.
All of the investigated SS, except unpleasant smells, were significantly more common in women with RVVC (Table 3).
Most frequently, these women complained about increased and whitish discharge, which disturbed 78.8% and 71.2% of infected women, respectively. On the other hand, edema was the least common SS, noted in only 10% of patients.
Signs and symptoms were then subjected to cluster analysis. The three-cluster solution was elected as optimal, based on the dendrogram (Figure 1). Signs and symptoms commonly associated with vaginal infection, burning, erythema, and itching, were clustered together into a cluster we named the "irritation cluster". The second cluster also comprised three SS: unpleasant smell, edema, and soreness/tenderness. This cluster of atypically grouped SS was named the "uncommon cluster". The third cluster consisted of two SS: excessive and whitish discharge, so it was labeled as the "discharge cluster". Forty-eight women presented with all SS of the "irritation cluster", while only 4 women presented with all SS of the "uncommon cluster". Both SS of the "discharge cluster" occurred in 208 women; therefore, this was the most common cluster (Table 4). The assessment of the QoL data showed that usual activities, pain/discomfort, self-confidence, and sexual problems were equally present in the examined groups. The statistically higher number of subjects without anxiety/depression was the RVVC group compared to the control group (38.8% vs. 16.5%, p = 0.001). Values of the EQ-VAS score were not significantly different between RVVC and the control group (73.9 ± 15.2 vs. 74.2 ± 14.7, p = 0.862)     There were no statistical differences in the frequency of presence of SS in any of the three clusters (p = 0.382, p = 0.249, p = 0.102) when comparing C. albicans and the NAC species (included S. cerevisiae).

Discussion
The assessment of the QoL data showed that usual activities, pain/discomfort, self-confidence, and sexual problems were equally present in the examined groups. The statistically higher number of subjects without anxiety/depression was the RVVC group compared to the control group (38.8% vs. 16.5%, p = 0.001). Values of the EQ-VAS score were not significantly different between RVVC and the control group (73.9 ± 15.2 vs. 74.2 ± 14.7, p = 0.862)

Discussion
The prevalence of RVVC of women has increased over the past years. Regardless of numerous treatments that seem beneficial and effective, a large portion of women develop a recurrence of the infection that can have an impact on the patient's QoL and self-confidence. One of the concepts emphasizes the role of the decreased sensitivity of yeast to applied antifungal therapy conjoined with reduced vaginal defense mechanisms [15].
Species C. albicans still remains the dominant species identified in over 70% of cases. However, confirmed NAC-RVVC shows remarkably higher prevalence in comparison to the last 20 years when they were found in less than 5% of women [16]. The most commonly isolated species were C. glabrata (13.8%) and C. krusei (6.3%). Rarely, etiological agents of RVVC were C. tropicalis (2.2%), C. kefyr (0.9%), and C. parapsilosis (1.3%). It is assumed that low-dose systemic antifungal therapy, single-dose local antifungal therapy, as well as possibility of self-treatment easily available to everyone have contributed not only to high prevalence of NAC-RVVC but also to their decreased susceptibility to antifungals [1,12].
In addition to the rising prevalence of non-Candida RVVC, a "new" cause of fungal genital infection is emerging-species S. cerevisiae. This yeast is a part of the physiological microbiota of respiratory, intestinal, and vaginal mucosa and was, until recently, considered nonpathogenic. Although genital infections caused by S. cerevisiae are still only occasionally reported, in our study, this species was responsible for about 2% of RVVC in women. Following the SS of S. cerevisiae infection, it was determined that all patients have increased and milky-white secretion as the dominant clinical finding, which is no different from a Candida infection. Research performed in Greece has even found that this fungi is the third most common cause of RVVC [11]. Similar to NAC-RVVC, S. cerevisiae infection is also associated with a higher risk of recurrence and shows decreased sensitivity to antifungals [17,18].
According to the fact that low sensitivity to applied therapy could be the cause of relapse, in our study, all isolates were tested by two commercial antifungal susceptibility tests. These kits for antifungal susceptibility testing can be used as a screening test for the detection of antifungal efficacy. Results of in vitro testing showed that in a very high percentage of women where C. albicans or C. glabrata were the causative agents, triazoles and MCZ did not have satisfactory effectiveness since the determined MICs for this drugs were higher. The mentioned finding was followed by lower sensitivity to antifungals of C. krusei and C. tropicalis. Species C. krusei is already a recognizable resistant-species to FCZ, and in this study, besides this, it showed lower sensitivity to ITZ in a high percentage as well. Similarly, a big portion of C. tropicalis isolates demonstrated low susceptibility to triazoles and MCZ. Only two species, C. parapsilosis and C. kefyr, were sensitive to all applied antifungal drugs, but these isolates were established in low percentage of women. As for the "new" causative agent of fungal genital infection, S. cerevisiae, it can be pointed out that it could be a problem in treatment since the two isolates showed low sensitivity to all included antifungals except 5-FC.
Contrary to our results, research that included over 3000 strains of Candida spp. isolated from women suffering from VVC did not note the high prevalence of samples with decreased susceptibility to antifungals [19]. This discrepancy can be explained by the fact that the mentioned research did not include women with RVVC. Such an assumption is supported by studies that compared the sensitivity of strains detected in RVVC and VVC women. These studies found that in-vitro tested antifungals had lower efficiency on the isolates from RVVC women [12,[20][21][22].
When it comes to SS of RVVC, it is shown that they are the same as the ones that occur in acute VVC [23], including itching, discharge, burning, and erythema [8,24]. These SS occur with different prevalence in different studies [23,25].
Patients with RVVC experienced almost all of the examined SS significantly more often compared to the control. Only the presence of odor did not differ significantly between groups. Given the fact that it is a highly subjective sign, this should not surprise. The dominant SS among our patients was an increased and whitish discharge, which was also observed by research done in Italy [25].
Since RVVC rarely presents with a single, but rather, with numerous SS simultaneously, SS cluster analysis was performed. This meant the creation of SS cluster from scratch, with no clue of what to expect, since this is the first such analysis on the subject. Three clusters were formed and labeled according to the SS they comprised. Thus, we had irritation, uncommon, and discharge clusters. Irritation and discharge clusters both included SS that usually appear together and are considered typical for infection. The uncommon cluster, as its name says, included SS that rarely appear together, suggesting that even though it was the rarest of the three, it can occur and mislead clinicians. For this reason, skipping steps in the diagnostic algorithm of RVVC and leaning only on the SS is not advisable. This shortcut, alongside self-diagnosed RVVC, carries the risk of the wrong treatment or overtreatment of the patient's condition. Additionally, the analysis of the presence of different species as causative agents of RVVC did not show a statistically significant difference in examined clusters.
The quality of life survey revealed that the EQ-VAS score is not significantly different among groups. Even though the control group had a slightly better EQ-VAS score than RVVC, no significant difference was presented, suggesting that RVVC does not crucially influence daily routine.
The limitation of the study is the usage of screening tests for the determination of antifungal susceptibility in the absence of an established standardized dilution test. Additionally, the use of more specific questionnaires might provide better insight into the patient's QoL. This study shows the emergence of RVVC caused by non-Candida species such as S. cerevisiae. Therefore, new definitions and new therapeutic strategies are required, mainly because these species show reduced sensitivity to antifungals. As RVVC is still a complex medical topic, the performance of the cluster analysis may be useful for application to laboratory characteristics, separately or in connection with clinical data; further studies are needed.