Frequency of microorganisms in vaginal discharges of high-risk pregnant women from a hospital in Caruaru , Pernambuco , Brazil

Introduction: During pregnancy, estrogen and progesterone levels may favor vaginal colonization by pathogenic microorganisms that can be associated with obstetric complications and neonatal infections. These pathogens include Candida spp., Trichomonas vaginalis, Gardnerella vaginalis and Streptococcus agalactiae. Objective: To determine the frequency of such microorganisms amongst high-risk pregnant women receiving prenatal care at a hospital in Caruaru, Pernambuco. Material and method: Analytical study in which samples of vaginal discharge were collected from women experiencing high-risk pregnancies, without restriction of age or gestational period. The samples were submitted to Gram-stained direct smear, as well as to fresh wet-mount examination and to cultures in Sabouraud and blood agar. Results: From May to December 2018, 92 patients were selected for sampling. The frequency of Candida spp. was found to be 31.52%. Concerning Gardnerella vaginalis, the frequency was 1.25%. The rate of colonization by Streptococcus agalactiae was 3.23%. No cases of Trichomonas vaginalis were found in this study. Discussion: The frequency of Candida that was found corroborates the literature, as it is a common infection during pregnancy due to hormonal increase. However, the frequencies of colonization by G. vaginalis, T. vaginalis and S. agalactiae were lower than those found in other studies. Conclusion: Due to the risks that these microorganisms can bring to a pregnant woman and a fetus, health professionals should be alert to signs and symptoms, requesting the screening of these pathogens, as well as treating gestating women when necessary.

introduction During pregnancy, estrogen and progesterone hormone levels induce changes in the genital tract of women. Those changes favor vaginal colonization by pathogenic microorganisms. Vaginitis is common during pregnancy, being associated with adverse perinatal outcomes, such as premature rupture of membranes and preterm delivery (1) . It is normally detected by the presence of odorless or abnormal smelling vaginal discharge (2) .
Vulvovaginal candidiasis (VVC) is an infection that affects more than 75% of women at least once in a lifetime, causing symptoms such as pruritus, dyspareunia and whitish curdy discharge. The Candida albicans species is responsible for around 80%-90% of the cases and proliferates in acidic environments, as in decreased vaginal pH (2,3) . Pregnancy is considered a risk factor for VVC, because the high estrogen and progesterone levels facilitate adherence and multiplication of the yeast and predispose pregnant women to primary candidiasis and recurrences (4) . Other risk factors are diabetes, immunosuppression, and the use of antibiotics (5) . In the latest years, studies verified that colonization by Candida spp. in early pregnancy is associated with increased risk of preterm delivery and low birth weight (6,7) . For treatment of VVC in pregnant women, topical nystatin is recommended. The use of systemic antifungal agents is contraindicated, especially in the first trimester of pregnancy (2) .
Other causes of vaginitis also associated with adverse perinatal outcomes are trichomoniasis and bacterial vaginosis (BV). Trichomoniasis, caused by the protozoa Trichomonas vaginalis, appears as an asymptomatic infection in 10%-50% of the cases; common symptoms are yellow or green foul-smelling vaginal discharge, pruritus, dysuria, and abdominal pain. There is evidence of association between infection by this microorganism during pregnancy and premature delivery and low birth weight (8,9) . Infection by T. vaginalis was also associated with postpartum maternal sepsis (10) . Trichomoniasis during pregnancy can be treated with a single 2-g dose of oral metronidazole (2) .
BV is characterized by increase in vaginal pH, reduction in normal microbiota, and overgrowth of anaerobic bacteria, particularly Gardnerella vaginalis. The main feature of this infection is the presence of a discharge with foul odor similar to "rotten fish", of variable color. This infection has been associated with preterm delivery, premature rupture of membranes, and postpartum endometritis (2,11) . Larsson et al. (2016) (12) demonstrated that, even after treatment with clindamycin, pregnant women with prior diagnosis of BV had higher risk of spontaneous preterm delivery than pregnant women with normal microbiota (12) .
Group B streptococcus (GBS), or Streptococcus agalactiae, is a major causing agent of gestational complications and neonatal infections. The gastrointestinal tract is a natural reservoir for GBS, which can colonize the vagina asymptomatically. Neonates born to colonized mothers can develop sepsis, pneumonia, and meningitis (13) . After the onset of labor and rupture of membranes, GBS can invade the amniotic fluid, although the microorganism also goes through intact membranes. Thus, fetal aspiration of GBS can lead to bacteremia. The bacterium can also be acquired during passage through the birth canal (14) .
According to the guidelines for prevention of perinatal GBS disease, issued by the Centers for Disease Control and Prevention (CDC), universal screening is recommended for pregnant women between the 35 th and the 37 th gestational weeks. GBS-colonized women must be given antibiotics during labor, and penicillin G crystalline is the drug of choice (13)(14)(15) . In Brazil, no consensus exists about prophylactic measures to reduce the incidence of neonatal GBS infection, so that GBS investigation is not part of the protocol of antenatal care proposed by the Ministry of Health (16) .
objectives Determine the prevalence of microorganisms in the vaginal discharge of high-risk pregnant women receiving prenatal care at a reference hospital in the municipality of Caruaru, Pernambuco, Brazil.

Study design and population
Analytical cross-sectional study, in which all women seen at the study period were included, following inclusion and exclusion criteria. Collections were carried out from May to December 2018. Inclusion criterion: being a high-risk pregnant woman under prenatal care at a hospital in the city of Caruaru, with no restriction of age or gestational period. Exclusion criteria: pregnant women suffering any clinical condition that impaired collection, such as miscarriage or bleeding risk, or that did not accept to have samples collected.

Data collection
The patients answered a questionnaire, prior to collection, in which it was possible to evaluate the following variables: age, marital status, level of schooling, income, occupation, race, origin (urban or rural area), presence of diseases, occurrence of miscarriage or preterm delivery in previous pregnancies, presence of discharge, pruritus, and dyspareunia.

Collection of vaginal and rectal specimens
Samples of vaginal and rectal discharge were collected using sterile swabs. Firstly, four vaginal swabs were collected from the distal third of the vagina and, later, a rectal swab. The first vaginal swab sample was used for smear preparation over a microscope slide for later staining with the Gram method. The second swab was placed into a tube containing 1-ml normal saline for fresh wet-mount examination. The third and fourth swabs were seeded onto Petri dishes containing Sabouraud dextrose agar 4% (Merck, Darmstadt, Germany) and sheep blood agar 5%, respectively. The rectal swab was also seeded onto a blood agar plate. After collection, the samples were transported up to the microbiology laboratory of Centro Universitário Tabosa de Almeida (Asces-Unita) to be analyzed.

Sample processing
The heat-fixed Gram-stained slides were microscopically visualized for detection of clue cells, characteristic of the infection by Gardnerella vaginalis. For the investigation of Trichomonas vaginalis and yeasts, a drop of the saline containing the discharge was deposited on the slide and looked under on a microscope. The plates containing Sabouraud agar and blood agar were kept in an incubator at 37°C for 24 hours.
In the plates where there was bacterial growth, colony morphology was observed. With the colonies suggestive of Candida spp. (white/beige, presenting creamy texture), Gramstained smears were prepared to confirm the presence of oval yeast-like structures of purple color.
With the colonies suggestive of Streptococcus agalactiae (small, grayish and beta-hemolytic), streaking was done over a new blood agar plate, to obtain pure culture. For final identification of the possible bacterium, the Christie, Atkins and Munch-Petersen (CAMP) test was conducted. For each GBS isolated found, the profile of susceptibility to antimicrobials was determined as recommended by the manual of the Clinical and Laboratory Standards Institute (CLSI) 2018 (17) . The test was carried out on blood agar; the following antibiotic discs were used: ampicillin (10 µg), cefepime (30 µg), vancomycin (30 µg), erythromycin (15 µg), clindamycin (2 µg), chloramphenicol (30 µg), tetracycline (30 µg) and azithromycin (15 µg).

Data analysis
Data, obtained from the questionnaire answers and laboratory results, were entered into an Excel spreadsheet (version 1810), to analyze prevalence and profile of pregnant women. In order to identify associations between clinical variables and laboratory results, Fisher exact test was used, with significance level of 0.05.

Profile of pregnant women
During the cited period, 92 samples from high-risk pregnant women were obtained. Patients' ages ranged from 13 to 43 years, with an average of 29.3 years. The predominant age group was that of 25-29 years (33.7%). A large group of patients had not completed primary school (38.04%). Common-law marriage was the most prevalent marital status (44.57%), and most women worked just at home (68.48%). The white race was the most frequent (45.65%), as well as the patients from the urban areas (69.57%) ( Table 1).
The found prevalence of Candida spp. and Gardnerella vaginalis was, respectively, 31.52% and 1.25%. The rate of colonization by GBS was 3.23%. No case of Trichomonas vaginalis was found in this study. The patient who had a result compatible with Gardnerella vaginalis was in the second trimester of gestation and reported a case of miscarriage; however, she did not present symptoms suggestive of BV.
Regarding the detected cases of GBS colonization, patients presented positivity just in vaginal culture. About the gestational period, a pregnant woman was in the 18 th week and another, in the 29 th week. One of the GBS isolates presented susceptibility just to vancomycin and chloramphenicol, being resistant to all the other tested antibiotics, including ampicillin. The other GBS isolate showed resistance to clindamycin, cefepime and tetracycline ( Table 2).  In the present study, the most prevalent microorganism in vaginal discharge samples of pregnant women was Candida spp., with a prevalence rate of 31.52%.
The VVC epidemiology varies a lot from region to region, depending on the population profile. Climate also interferes with the prevalence of this microorganism, as candidiasis is the most common type of vulvovaginitis in tropical countries (18,19) . The results found in this study are close to those obtained by Bonfanti and Gonçalves (2010) (20) , who discovered a prevalence of 33.75% during analysis of cytopathological reports of pregnant women from Rio Grande do Sul, Brazil. Lower rates were found in Argentina (28%) (21) , United Kingdom (12.5%) (22) , Malaysia (17.2%) (23) , Nigeria (25%) (24) and India (4,13%) (1) .
Pregnancy is considered a risk factor for VVC, given the hormone alterations and the increased deposit of glycogen and other substrates in the vagina during that period (25,26) . In many cases, the infection is asymptomatic, being called colonization (27) . In the present study, 62.07% of the women with positive culture for Candida spp. presented at least one of the symptoms suggestive of VVC, while 37.93% were asymptomatic. Another important risk factor for candidiasis is a situation of hyperglycemia, when excessive glycogen facilitates the establishment of the microorganism (25) . In this study, seven patients were diabetics and among them, three presented positive cultures for Candida spp.
Some authors, such as Olowe et al. (2014) (5) and Sangaré et al. (2017) (27) , stated that the incidence of candidiasis increases with gestational age. There are divergences in the literature, however. In the study conducted by Masri et al. (2015) (23) , pregnant women in the first and second trimesters of gestation had higher risk of acquiring candidiasis than those in the third trimester. Parveen et al. (2008) (28) and Brandão (2017) (3) , in their turn, did not find association between gestational period and VVC. In this study, the highest prevalence of Candida spp. colonization occurred in the second trimester of pregnancy. Kanagal et al. (2014) (29) also found higher prevalence of Candida spp. in the second trimester.

Gardnerella vaginalis
A case of BV was found in this study, what corresponds to a rate of 1.25% -low, compared to the literature. Akinbiyi et al. (2008) (22) also found a low prevalence (3.54%) among asymptomatic pregnant women in the United Kingdom. Rao and Chandini (2017) (11) found a rate of 19.2% among pregnant women of a hospital in India, without statistically significant association with gestational period. In Nigeria, Olowe et al. (2014) (5) found a prevalence of 38%, associated with recent use of antibiotics. In the study by Monteiro et al. (2017) (30) , the found prevalence of BV was 19% among pregnant women treated at a maternity hospital in the city of Natal, Rio Grande do Norte, Brazil (30) .

GBS
It is estimated that 10%-30% of pregnant women globally are colonized asymptomatically by GBS (14) . This prevalence varies a lot among studies due to several factors, such as characteristics of the studied population, collection method and laboratory tests.
In this study, the prevalence of GBS colonization was 3.23%, a percentage close to the work conducted by Shirazi et al. (2014), which verified prevalence of 4.8% (31) . A lower prevalence was found by Sharmila et al. (2014) (32) -2.3% among Indian pregnant women who were in the 35 th -37 th weeks of gestation. A factor that might have contributed to this result is the high number of pregnant women that underwent recent treatment for urinary tract infections, what led to decolonization upon collection time (13) .
In Brazil, the prevalence of GBS colonization among pregnant women varies widely according to the region. Dias (2014) (33) found a 13.95% rate in pregnant women in Cuiabá, Mato Grosso. In Niterói, Rio de Janeiro, a 6.1% prevalence was found (34) . Nunes et al. (2015) (35) analyzed 1,425 records of pregnant women treated in Florianópolis who underwent GBS investigation; 16.5% were colonized. In the study by Senger et al. (2016) (36) , 22.5% of the pregnant women were colonized, in the municipality of Santo Ângelo, Rio Grande do Sul (36) . Another study, also in Rio Grande do Sul, demonstrated a rate of 8.8% (37) .
Penicillin is the drug of choice for intrapartum antibiotic prophylaxis of neonatal infection. Ampicillin can be used as an alternative. Allergic patients are suggested to undergo susceptibility tests to clindamycin and erythromycin, antibiotics that can also be used. In case of resistance, vancomycin is recommended (13)(14)(15) .
conclusion Candida spp. was the most prevalent microorganism in the samples from vaginal discharge of high-risk pregnant women.
As several studies point the association of this microorganism (although common during pregnancy) with preterm deliveries and low birth weight, it is important to confirm the agent and provide the treatment, because the picture may be asymptomatic. Low prevalence of Gardnerella vaginalis, Trichomonas vaginalis and Streptococcus agalactiae was found. All these microorganisms deserve attention by health professionals because of the possible risk they pose to pregnant women and fetuses.

acknowledgements
The authors thank Hospital Jesus Nazareno for the availability of the service, and Asces-Unita for the support in infrastructure and inputs. e2062020 Letícia Fernanda Q. Freitas; Larissa Rayanne S. Maia; Maria Rafaela A. R. de Deus; Sibele R. Oliveira; Adrya Lúcia Peres