Second trimester vaginal Candida colonization among pregnant women attending antenatal care in Bukavu, Democratic Republic of the Congo: prevalence, clinical correlates, risk factors and pregnancy outcomes

Introduction Vaginal Candida colonization (CC) can lead to vulvovaginal candidiasis, the second most prevalent vaginal condition worldwide, and has been associated with adverse birth outcomes. However, no data on CC in the Democratic Republic of the Congo are available. We investigated the prevalence, Candida species, clinical correlates, risk factors and pregnancy outcomes in women with CC in the second trimester of pregnancy. Material and methods In Bukavu, the Democratic Republic of the Congo, pregnant women were recruited during antenatal care between 16 and 20 weeks of gestation from January 2017 to October 2017 and followed until delivery. Sociodemographics, sexual behavioral, hygienic and clinical characteristics, microbiological data and pregnancy outcomes were collected. Candida detection and speciation was performed with microscopy (Gram-stained smears and wet-mount) and/or quantitative PCR. Multivariate regression models were used to estimate the different associations with CC. Results The prevalence of CC by wet mount, microscopy of Gram-stain smears and qPCR was 27.9%, 28.1% and 38.2%, respectively. C. albicans was the most prevalent Candida species (91.0%). Previous genital infections, an intermediate vaginal microbiota, bacterial vaginosis, and the use of pit toilets were risk factors for CC. Clinically, CC was associated with itching only. Women with CC had twice the odds for preterm birth, if Candida concentration was high, the odds were four times higher. Conclusions In Bukavu, the Democratic Republic of the Congo, the prevalence of CC was high and associated with microbiological and modifiable risk factors. Screening and treatment for CC during antenatal care should be investigated as a possible strategy to reduce preterm birth.


Introduction
A healthy vaginal microbiota (VMB) of women of reproductive age is typically dominated by only a single Lactobacillus species, mostly L. crispatus, L. gasseri or L. jensenii (1).The role of L. iners in reproductive health is under debate (2).These single Lactobacillus species or consortia of different Lactobacillus species (3) produce lactic acid by metabolizing the amylase breakdown products of glycogen present in the vagina.This acidifies the vagina and is a crucial first-line defense protecting women against urogenital tract infections.Furthermore, a healthy VMB is associated with healthy pregnancy outcomes.Vice versa, bacterial vaginosis (BV) is a dysbiosis of this healthy VMB predominated by anaerobes such as Gardnerella and Fannyhessea and linked with adverse pregnancy outcomes such as preterm birth, low birth weight and premature ruptures of membranes (1).
The yeast Candida can colonize the VMB as a commensal.This vaginal Candida colonization (CC) is more prevalent in case of elevated endogenous estrogen levels, such as during pregnancy (4).CC can transition from a commensal state to a pathogenic state-vulvovaginal candidiasis-under certain conditions that are incompletely understood (4,5).Clinical therapy of vulvovaginal candidiasis is difficult due to the capacity of Candida to form biofilms (6) and withstand the host immune system (7).
In most populations, vulvovaginal candidiasis and bacterial vaginosis are the leading causes of vaginal dysbiosis in women of reproductive age (8).In by far most cases, vulvovaginal candidiasis is caused by Candida albicans (9), the remainder by non-albicans Candida spp., the most common of which is Candida glabrata.The usual presenting symptoms are acute pruritus and vaginal discharge, but none is specific (9).Vaginal discharge is present inconsistently and is often neglible, not uncommon symptoms include vaginal soreness, irritation, vulvar burning, dyspareunia (9).Importantly, asymptomatic CC be present in up to 50% of women (10).Besides clinical suspicion, microscopy has been used to diagnose vulvovaginal candidiasis for decades, however, culture has remained the gold standard (11).
In the Democratic Republic of the Congo (DRC), in case of absence of laboratory capacity, the syndromic management is implemented as recommended by the WHO (12).This syndromic management is based on the identification of consistent groups of symptoms and easily recognized signs, such as the vaginal discharge syndrome, and the treatment of the majority of, or the most serious, pathogens responsible for the syndrome (12).In practice, women presenting with vaginal discharge are often treated for BV, vulvovaginal candidiasis and trichomoniasis (13).This strategy results in frequent undertreatment or overtreatment, which is frequently accompanied by recurring vaginal problems.
In observational studies, vulvovaginal candidiasis but also CC have been conflictingly associated with increased risk of spontaneous preterm birth (PTB), the birth before 37 weeks of gestation (14,15).PTB directly accounts for 35% of the annual neonatal mortality worldwide (16).A review approximated that 14.8 million infants were born preterm in 2014 with most cases occurring in Sub-Saharan Africa and Asia (17).In DRC, the neonatal mortality rate is estimated to be 2.8% (18) and 32% of these mortalities in DRC are caused by PTB complications (19).
Despite the importance of vulvovaginal candidiasis, and the potential detrimental role of CC in pregnancy, the latter is immensely understudied, especially in sub-Saharan Africa and DRC.There are no data on prevalence, Candida species distribution, clinical presentation, risk factors and the role of Candida in birth outcomes.Therefore, we aimed to address these major gaps in a population of pregnant women from Bukavu, DRC.

Study design and population
The current study was part of the AVEONS study, described previously (20).AVEONS was a prospective observational study and included women between 16 and 20 weeks of gestational age [visit 1 (V1)], between 36 and 38 weeks (V2) and at delivery.The main aim of AVEONS was to investigate the role of vaginal infections in adverse pregnancy outcomes in Bukavu, DRC.Newborns were followed from delivery until day seven of life.This manuscript only considers CC at V1 and pregnancy outcomes.

Recruitment
Between January and October 2017, pregnant women were recruited when seeking antenatal care at the Provincial Referral Hospital of Bukavu (PRHB).Church announcements, radio and TV spots, community leader's meetings and posters were used to promote the study.Pregnant women who were interested in participating were informed individually and requested to sign an informed consent form.Participants were reimbursed for their transportation costs.

Inclusion and exclusion criteria
Pregnant women were considered for inclusion if (i) they were between 16 and 20 weeks of gestation, (ii) they accepted to be followed by a referral hospital team and were willing to deliver at PRHB and (iii) they accepted to be contacted by phone or other means.Women were not considered eligible in case (i) they planned to move out of Bukavu during their pregnancy, (ii) of twin pregnancies, (iii) of a fetus with a visible malformation at ultrasound examination, (iv) they had genital bleeding and/or (v) of use of antibiotics during the 2 weeks before recruitment.

Routine antenatal care and delivery procedures
A general physical and gynecological examination was performed.Urine (to diagnose urinary tract infections or bacteriuria) and five mL of total blood (to test for HIV, malaria and hemoglobin) was collected.Also, a gynecological examination, including a speculum examination with a sterile non-moistened speculum, was performed.During this examination, both the vaginal mucosa and cervix were inspected for the presence of sores and tumors.A vaginal infection was diagnosed according to the syndromic-based guidelines for the management of pregnancy issued by the Ministry of Public Health of DRC (21), which are based on WHO recommendations (22).This syndromic-based diagnosis was made in case the participant suffered from abnormal vaginal discharge (thick curdy), itching, redness, abnormal foul smell and/or burning sensation after intercourse.In case of vaginal infection, women were treated empirically with a combination of clotrimazole (200 mg) and clindamycin (100 mg).This combination is an efficient, safe and well-tolerated combination for treating mixed vaginitis caused by bacteria and fungi, also in pregnancy (23).Due to a lack of diagnostic tools in DRC, the syndromic method is employed to manage vulvovaginal diseases and a broad-spectrum combination medication capable of treating fungal, bacterial, and mixed infections is needed.
An ultrasound examination was performed to assess the viability of the fetus and to measure the cervical length.Women followed antenatal care as usual and were offered a single dose of 500 mg mebendazole (against intestinal worm infections) and a single dose of sulfadiazine-pyrimethamine (500 mg) against malaria.At delivery, pediatricians evaluated the newborns and assessed their anthropometric features.The WHO criteria were used to suspect an early newborn infection.Nurses and a senior assistant monitored labor and collected the relevant data, such as vital signs, vaginal delivery mode, obstetrical criteria of delivery, and symptoms of vulvovaginal infection.

Study specific procedures
Data on sociodemographics and reproductive health history, sexual behavior, vaginal practices and complaints were obtained by questionnaires.The vaginal pH was determined by indicator pH papers (Hilindicator ® pH paper).Three vaginal swabs and one cervicovaginal lavage (CVL) were taken.The first swab was rolled on three glass slides, one for direct wet mount microscopy and two for future Gram staining, the second swab was used for culturing and the third for molecular testing.CVLs were taken as previously described (24).Briefly, using a pipette, 10 ml of normal saline was flushed over the cervix and the lateral vaginal walls, aspirated as much as possible from the posterior fornix and collected in a vacutube.CVLs where then put in a cooled box (4 °C) and immediately transported to the laboratory, where they were frozen at −20 °C until shipment to the Laboratory Bacteriology Research (LBR) (Ghent University, Ghent, Belgium).

Microscopic detection of Candida and BV
The two vaginal smears were fixated, shipped to the LBR, Gram-stained using an automated Poly Stainer and scored by two independent readers according to the Nugent scoring system (26) to categorize the vaginal microbiota as healthy, intermediate microbiota or BV.In case of discrepancy in categorization, slides were reassessed by the two reviewers and discussed.In case no consensus was obtained, a third reader assessed the slide as a tie breaker.The presence of Candida (cells and/or hyphae) upon microscopic examination of Gram-stained smears was used as the case definition of CC.

Molecular detection and speciation of Candida
DNA was extracted from CVLs using the RNeasy PowerMicrobiome Kit (Qiagen, Antwerp, Belgium) according to manufacturer's instructions with minor modifications.To detect and speciate Candida in the CVL DNA extracts, fungal DNA was amplified in a qPCR assay using universal fungal primers targeting the internal transcribed spacer-2 region followed by high resolution melting of the amplicons and analysis of the melting curves for speciation as described previously (27).Briefly, a total reaction volume of 10 µl was prepared containing the primers ITS4 (5'-TCCTCCGCTTATTGATATGC-3') and ITS86 (5'-GTGAATCATCGAATCTTTGAAC-3'), both at a final concentration of 0.5 μM, 1× LightCycler 480 HRM master mix (Roche) and two μl of DNA extract or control DNA.Amplification was carried out on a LightCycler 480 (Roche) by preincubating the reaction mixture for 10 min at 95 °C, followed by 45 cycles of 20 s at 95 °C, 30 s at 55 °C and 30 s at 72 °C.Hereafter, a high resolution melting curve was generated by first melting all DNA at 95 °C for 5 s, followed by renaturating DNA for 1 min at 60 °C, and gradually increasing the temperature from 60 °C to 97 °C at a ramp rate of 0.02 °C /s.Results were analyzed with the LightCycler 480 Software, version 1.5 (Roche), as described previously ( 27

Data analysis
A case of CC was defined based on the microscopic observation of Candida cells and/or hyphae on Gram-stained vaginal smears.PTB was defined as delivery before 37 completed weeks of gestation and after 28 completed weeks of gestation, low birth weight as birth weight <2,500 g (28-30).
For a comparison of the diagnostic performance of different assays, in case of absence of a true gold standard assay, a composite reference standard (CRS) can be used as the gold standard (31).For Candida detection, there is no gold standard, hence, a CRS for Candida colonization was used to assess sensitivity and specificity of the different Candida diagnostic methods.A case of Candida colonization using a CRS was defined if from the three diagnostic methods (wet mount microscopy, Gram-stain microscopy and qPCR) at least two of the assays were positive.This comparison of diagnostic performance of three methods was based on a subset of 330 samples.To investigate to which extent the sensitivity of Gramstain and wet mount-based microscopy were function of the Candida concentration, we (i) stratified the qPCR positive samples into three more or less equal groups, i.e., low (<9.0E+ 03 Candida cells/ml), moderate (9.0E + 03 Candida cells/ml-2.0E+ 05 Candida cells/ml) and high concentration (>2.0E + 05 Candida cells/ml) and (ii) assessed the Gram-stain and wet mount positive rate in each of the concentration categories.Fleiss Cohen's kappa test was used to determine the agreement between Candida assays (32).
The Chi-squared test was used to compare CC with categorical independent variables.A binary logistic regression was performed to assess the associations between CC and different sociodemographic, sexual behavioral, hygienic, clinical, and anthropometric characteristics and laboratory findings (independent variables).These associations were expressed as crude odds ratios and 95% confidence interval (CI).Then, a multiple logistic regression model (using a combination of the stepwise forward or backward method) was applied to select CC risk factors and calculate adjusted odds ratios and 95% CI.This model was based on both probability-to-enter of 0.05 and probability-to-remove of 0.10.
Another multiple logistic regression model stratified by vaginal microbiota status was employed to estimate the association of CC with clinical symptoms.The goodness-of-fit of the final models was tested using the Hosmer-Lemeshow test.
A modified Poisson regression model with robust standard error was built to further investigate the extent to which CC was as an independent risk factor for adverse pregnancy outcomes [preterm birth (PTB), low birth weight, and premature ruptures of membranes (PROM)].We corrected for a priori defined (possible) confounding factors influencing pregnancy outcome suggested by other studies (33-36) (vaginal microbiota, adjusted hemoglobin rate, cervix length, parity, body mass index, maternal age, education level and poverty).
All forms (laboratory forms, questionnaires, case reports forms) were scanned and stored digitally as a singular file.All data were entered by double data entry using the CSPRO software.The comparison data tool of CSPRO was used to compare the two data files (from the double data entry) for discrepancies.Discrepancies were resolved by consulting the original data in the scanned documents.Analyses were performed in STATA 14 (Stata Corp, College Station, US), p-values below 0.05 were considered as statistically significant.

Ethics statement
This study was ethically approved by the Internal Review Board of the Catholic University of Bukavu (UCB/CIE/NC/016/2016), by the Ministry of Public health (062/CD/DPS/SK/2017) and by the Ethical Committee of Ghent University Hospital (PA2014/003).An informed consent form was signed by each pregnant woman who accepted to participate in the study.

Sociodemographic description of the study population
The number of pregnant women and neonates withheld at each visit is shown in Figure 1.
A total of 750 pregnant women were screened of which 533 were found eligible and were enrolled in the cohort (V1).A total of 104 women (19.5%) who started in the study withdrew from the study before V2.A total of 26 (4.9%) women had a second trimester pregnancy loss.Of all newborns in the study, 288 (85.5%) were born at term and 49 (14.5%) were born preterm.Of the 354 women who completed V2, 66 (12.4%) decided to not deliver at PRHB but near their home.
The sociodemographic characteristics of the study population are outlined in Supplementary Information S1.
Near the half of pregnant women (48.4%) were symptomatic, i.e., suffered from an abnormal vaginal discharge, vaginal itching, burning sensation after sexual intercourse, and/or a foul smell from the vagina.

Prevalence of vaginal Candida colonization and speciation
Table 1 documents the prevalence assessed in both symptomatic and asymptomatic pregnant women according to the different diagnostic tests.The overall prevalence of CC by means of microscopy of Gram-stained smears was 28.1%, in symptomatic and asymptomatic women, this was 48.1% and 8.3%, respectively.Using wet mount microscopy, the overall CC prevalence was 27.9%, in symptomatic and asymptomatic women, the prevalence was 55.6% and 1.2%, respectively.The overall CC prevalence assessed by qPCR (38.2%) was significantly higher compared to the microscopic methods, and similar in symptomatic and asymptomatic women (38.9% and 37.5%, respectively).

Diagnostic performance of microscopic and molecular tests
The diagnostic performance parameters are shown in Table 2.The highest sensitivity was found for qPCR (performed on CVLs) (99.0%), followed by microscopic examination of Gram-stained

Independent risk factors for vaginal Candida colonization
The univariate and multivariate associations are shown in Supplementary Information S1 and Table 3, respectively.CC was significantly and independently associated with the use of a pit toilet compared to a flush toilet (aOR: 1.85; 95% CI: 1. 22

Symptomatology of vaginal Candida colonization
The univariate associations of clinical signs and symptoms with CC are shown in Supplementary Information S1.The multivariate model stratified by the VMB status is shown in Table 4. Vaginal itching was statistically significantly associated with CC across strata (aOR: 1.98, 95% CI: 1.02-3.83;aOR: 3.26, 95% CI: 1.30-8.17

Discussion
We aimed to investigate the prevalence, species, risk factor and clinical correlates of CC, as well as the association with pregnancy outcomes.To be best of our knowledge, no studies on Candida in pregnancy have been performed in DRC.The prevalence of CC was high, and associated with other genital tract infections, hygiene, and preterm birth.

The prevalence of vaginal Candida colonization in pregnant women from Bukavu was high
We found a prevalence of 28.1% of CC in second trimester of pregnancy, as assessed by microscopy of Gram-stained smears; wet mount gave very similar results.Not surprisingly, in our study, the prevalence of CC as assessed by qPCR was much higher (38%).Interestingly, there was a clear difference in prevalence between symptomatic (48% and 56% for Gram-stain and wet mount, respectively) and asymptomatic women (8% and 1% for Gramstain and wet mount, respectively) for the microscopic methods only.Such as difference was not seen for the qPCR method (symptomatic and asymptomatic women 39% and 38%, respectively).This might be due to the higher sensitivity of qPCR.The concentration of Candida is likely higher when the patient is symptomatic and it is also possible that qPCR is detecting DNA from dead/lysed Candida cells, hence, not visible by microscopy.
An overview of studies from other countries in sub-Saharan Africa reporting CC rates in pregnant women is given in Supplementary Information S5.In Sub-Saharan Africa in pregnancy, prevalence ranges from 14% (Burkina Faso) to 57% (South Africa) in the normal pregnant population.Differences in prevalence between all studies may be explained by different populations and the use of different diagnostic tests, which have different sensitivities, as has been shown in current study and by others (37)(38)(39).Also, the different mean gestational age between populations can impact Candida prevalence (36), although some studies from sub-Saharan Africa found no statistically significant association with gestational age (10,40,41).
We found that C. albicans-the most virulent Candida species colonizing the vagina (42-44)-was the most prevalent yeast (91.0%).Other studies from sub-Saharan Africa documented C. albicans as the most yeast species in pregnant women, ranging from 40.4% to 90.7% (Supplementary Information S3).The nonalbicans species (C.glabrata, C. dubliensis, C. krusei, and C. tropicalis) in our study all have been isolated previously from cases of vaginitis, with C. glabrata being considered the most prevalent cause of non-albicans vaginitis.Other species found in our study, C. famata, C. inconspicua, C. kefyr and S. cerevisiae, are less common to (very) rare causes of vaginitis (45,46).

Previous vaginal infections and concurrent VM disturbances were independent risk factors for vaginal Candida colonization
Women with a history of vaginal infection in the last six months had a nearly five-time higher odds of Candida colonization.Although we have no specific information on the type of these previous vaginal infections, not unlikely, these were (i) (recurrent) vulvovaginal candidiasis, which has a global prevalence of 9% in the age group 25-34 years (47), (ii) and/or a previous STI, a risk factor for vulvovaginal candidiasis (48), (iii) and/or (recurrent) intermediate VM/BV, which were also risk factors for CC in our study.Indeed, women with an intermediate VMB or BV had a two-to three-fold higher odds for CC.The protective role of (probiotic) lactobacilli against Candida albicans is controversial (49) and clinical studies have shown conflicting findings on the association between the VM and vaginal Candida (50,51).It has been demonstrated different Lactobacillus species have different probiotic effects against Candida in vitro (52).

Improved sanitation might reduce the prevalence of vaginal Candida colonization
Women using a pit toilet had a higher risk for CC compared to women using a flushing toilet.Commonly, pit toilets lack (sufficient) running water and/or soap for a proper cleaning of hands and/or perineum.As the gut has been documented as a reservoir for vaginal Candida (53), poor cleansing could enhance the spread Candida from perineum to the vagina and everywhere hand-touched in the toilet.Pit toilets have also been associated with other urogenital infections such as trichomoniasis (54,55).
Our findings further stress the need of improving sanitation facilities-also in urban areas such as in our study-to reduce urogenital infections, especially among pregnant women.We further found that pregnant women who practiced labia elongation had a lower risk of CC.In several central African countries, lengthening of the labia is done as an esthetic improvement, to increase sexual pleasure, and/or to be accepted by the tribe (56).

Vaginal Candida colonization was independently associated with preterm birth
Prevention of PTB remains one of the most important challenges in maternity and neonatology care (16,17).Although the etiology of PTB is multifactorial, both local and systemic infections are a major cause (57).
Here, we found that CC during the second trimester of pregnancy was independently associated with PTB, with women who carried Candida having a two-fold higher odds for PTB.Importantly, these odds were nearly four-fold in women colonized with high concentrations of Candida, defined in our study as >2E05 Candida cells/ml by means of qPCR.
A high concentration of Candida might increase the possibility of an ascending infection leading to PTB.In some cases, vaginal pathogen concentrations are indeed important in reproductively health.For example, Goodfellow and colleagues showed that the vaginal concentration of L. iners was associated with early PTB recurrence (58).Also, in Group B Streptococcus (GBS) neonatal sepsis, GBS is transmitted from the maternal vagina to the neonate during birth or after breaking of the fetal membranes.It is known that there is a direct relationship between the maternal vaginal GBS concentration, the risk of vertical transmission and the likelihood of serious disease in neonates (59).
Future studies investigating the role of the VMB in pregnancy should consider absolute concentrations.This is currently a shortcoming of the widely used 16S rRNA gene metagenomic next-generation sequencing studies that report in relative abundances.Himschoot et al. (2024, under revision) showed that across samples, the same relative abundances of vaginal key species can represent a large range (several logs) of absolute concentrations.
Two meta-analyses (one considering only asymptomatic CC (14) and one considering both symptomatic and asymptomatic CC (15)) reported no association between (asymptomatic) vaginal Candida and PTB.Several factors might explain why our findings are contrasting.
First, only two studies (from Kenya) in these meta-analyses were performed in sub-Saharan Africa.Both PTB and CC, common among black women, are influenced by maternal ethnicity (60,61).Also, the VMB differs between African and Caucasian populations (62).Furthermore, there is an important difference in the management (including treatment) of vaginal Candida between high-income countries and low-and-middle income countries (see further), and treatment differences between studies are present.Second, we tested for CC in the second trimester of pregnancy, but the timepoint(s) of testing varied in other studies.Candida later in pregnancy might have less impact on PTB.Third, diagnostic methods differ between studies.

Candida and BV management in antenatal care
In DRC, self-reported vaginal symptoms remain the backbone of the syndromic management of genital infections, and clinical examination or microscopy are not often used.This approach has been shown to perform poorly, for example, an evaluation in Kenya showed that only 50% of women presenting with vaginal complaints received appropriate treatment (63).Currently, the WHO guidelines for the management of genital infections have included recommendations for the use of molecular assays or more simple point-of-care tests, but implementing this remains a challenge/unfeasible in most resource-limited settings.We showed that microscopic examination of wet mount preparations and/or Gram-stained smears-even with a sub-optimal sensitivity compared to qPCR-only fails to detect Candida if present at low concentrations (as assessed by qPCR), which might be clinically less relevant and was not associated with PTB in our study.
In our study, 45% of women had CC and/or BV (19% CC, 17% BV and 9% CC/BV coinfection).Often, symptoms of these conditions are similar, complicating clinical diagnosis and effective treatment (47,64).In our study, the only symptom associated with CC across VMB categories of healthy (OR 1.98), intermediate (3.26) and BV (OR 3.40) was vaginal itching, similar to findings reported by Gloria and colleagues (OR 2.20, 95% CI 1.40-3.46)(63).
WHO recommends to screen women with previous spontaneous abortion, stillbirth and/or preterm delivery for BV and trichomoniasis (65), and our findings call for further investigation if Candida should be included.Microscopy could be considered as routine screening of women in antenatal care.Although microscopy is an affordable and rather easy to implement technique, it's use often seems to be arduous and tedious owing to the lack of a relevant practitioner's training.Development and implementation of cheap and quick point-ofcare tests should be intensified as a means for the reduction of PTB in resources constrained countries.These assays should include multiple targets (Candida, BV, and Trichomonas vaginalis).

Limitations and future perspectives
Our study was limited by the drop-out of approximately one third of the study participants during follow-up due to socioeconomic and political reasons.Because of this, our study might be underpowered to assess the association between CC and PTB, hence, results should be interpreted with care.
Considering we found a substantial part of pregnant women, not only from our study but also other Sub-Saharan African countries (see Supplementary Information S3), were colonized by Candida, and CC was associated with increased risk for PTB, further research should be done.A meta-analysis of two randomized controlled trials indicated that clotrimazole treatment for vaginal Candida reduced the risk for spontaneous preterm birth compared to standard of care (risk ratio 0.36, 95% CI 0.17-0.75)(60).Similar trials should investigate if a similar approach also can reduce the incidence of PTB in DRC.To guide such interventions, research on the development and implementation of cheap point-of-care tests and/or microscopy to diagnose Candida, BV and sexually transmitted infections in antenatal care in LMIC should be intensified.

Conclusion
In this first report on vaginal Candida in women from DRC, we document a high prevalence of CC in pregnancy, mainly C. albicans.CC was associated with microbiological and modifiable risk factors.Vaginal Candida, especially when present at high concentrations, was an independent risk factor for PTB.

Key message
Vaginal Candida colonization is associated with preterm birth in a population from the Democratic Republic of the Congo.
).As positive controls in each run, DNA from Candida dubliniensis, C. famata, C. glabrata, C. guillermondi, C. inconspicua, C. kefyr, C. krusei, C. lipolytica, C. lusitania, C. metapsilosis, C. nivariensis, C. norvegensis, C. metapsilosis, C. ortopsilosis, C. parapsilosis, C. tropicalis and Saccharomyces cerevisiae was used.In each run, ten negative template controls were used, and each sample was analyzed in duplicate.High resolution melting curves peaks of positive samples were compared with those of reference standards for identification of the yeast.

FIGURE 1 Flowchart
FIGURE 1Flowchart of the study.WGA, weeks of gestational age.

FIGURE 2
FIGURE 2 Positivity rate of vaginal Candida colonization as assessed by microscopic techniques according to Candida concentration measured by qPCR.The different classes of Candida concentration (assessed by means of qPCR) are categorized into low, moderate and high.The black part of each column represents the proportion of samples found positive by Gram-stain or wet mount in the different categories, the grey part the samples found negative.The number in the columns are absolute numbers, the numbers on the Y-axis are percentages.GS, microscopic examination of Gram-stained smears; WM, microscopic examination of wet mounts.

TABLE 2
Diagnostic performance (with 95% confidence interval) of the different tests for Candida vaginal colonization (wet mount, gram-stain, qPCR) based on 330 women.For the comparison of the diagnostic performance of the different Candida assays (i.e., wet mount microscopy, Gram-stain microscopy, and qPCR), a composite reference standard (CRS) was considered as the gold standard and a positive CC case was defined if at least two of these assays were positive. a

TABLE 1
Prevalence (with 95% confidence interval) of vaginal Candida colonization among symptomatic and asymptomatic pregnant women assessed by three diagnostic methods.
a Defined as abnormal vaginal discharge, vaginal itching, burning sensation after sexual intercourse, and/or a foul smell from the vagina.

TABLE 3
Multivariate risk factor associations with vaginal Candida colonization.
aOR, adjusted odds ratio; CC, vaginal Candida colonization; cOR, crude odds ratio; Ref., reference category; 95% CI, 95% confidence interval.Bold values represent values that are statistically significant at the P < 0.05 level.a In the last six months.b An old practice to lengthen outer lips by using herbs during adolescence.

TABLE 4
Clinical symptoms associated with vaginal Candida colonization stratified by the vaginal microbiota status., adjusted odds ratio; Ref., reference category; 95% CI, 95% confidence interval.Bold values represent values that are statistically significant at the P < 0.05 level. aOR

TABLE 5
Multiple regression model for preterm birth and vaginal Candida colonization.As assessed by means of the Nugent score. a