Our data showed that 15.5% of Vietnamese pregnant women had AV in the third trimester of pregnancy, and majority of these cases represented AV light form (13%), while the moderate AV was diagnosed only in 2.5% of pregnant women. However, despite of the light/moderate form, the AV was significantly associated with an increased risk of postpartum infections that is considered to be one of the leading causes of maternal death worldwide.
Etiopathogenesis, diagnosis and treatment of AV
Healthy vaginal microbiota is dominated by lactobacilli while imbalance of this community may result in AV, BV or candidiasis. In case of AV, lactobacilli counts are decreased while that of several enteric bacteria are increased, including
S. agalactiae,
E. faecalis,
E. coli, CoNS and
S. aureus [
2]
. During last years, some studies have applied next-generation sequencing to AV patients. These studies have indicated increased number of species [
10], elevated proportions of
Actinobacteria and
Bacteroidetes [
11], and surprisingly also high abundance of
Gardnerella vaginalis and
Prevotella bivia, in addition to that of
S. agalactiae [
8]. Not only the species composition, but also their microbial concentration may be of importance. Moreover, it is not finally clear whether this imbalanced community is of the causative factor of the AV or the associated feature present in AV. It has also been proposed that AV may be an immunological disorder with an influence on the vaginal microbiota [
12]. This idea is supported by variable degrees of inflammation, not only measured by the number of leucocytes, but also by the proportions of so-called ‘toxic leucocytes’ (swollen leucocytes containing granules showing lysosomal activity), imbalance of other immune markers and a variable presence of parabasal or immature epithelial cells. Other possible pathogenetic factors include low local estrogen level, presence of sialidase-producing bacteria and deficiency of vitamin D [
3], while also presence of some sexually transmitted infections [
4] has also been linked to AV. Intrauterine devices, being unmarried, long term use of antibiotics and frequent vaginal douching have been considered to contribute to the AV etiology [
13].
Diagnosis of AV can be made using wet mount phase contrast microscopy. An AV score is calculated according to lactobacilli grade (LBG), presence of inflammation, proportion of toxic leukocytes, microbiota characteristics, and parabasal epithelial cells. A score ranging from 0 to 2 is assigned to each of the above-mentioned characteristics, and AV is diagnosed according to the composite score: 3–4 indicates light AV, 5–6 represents moderate AV and 7–10 represents severe AV (Table
1). Cultures can give additional information about the microbiota composition but the diagnosis cannot rely on the cultures only. To circumvent the hurdle of microscopic investigation, some attempts have been made to develop nucleic-acid-based and enzymatic diagnostic tests, but the detailed information obtained with microscopy, especially cellular pattern, has remained irreplaceable [
3,
14]. The AV diagnostic criteria among previous studies have not been unanimous. Choice of diagnostic approach depends on the suspected diagnosis and the readiness of the health care system to recognize, diagnose and treat AV. Clinical picture is often nonspecific, affecting the initial diagnostic approach as well as prescribing inaccurate medication. Most clinicians deem that the expected AV rate may be higher and that the majority of AV cases remain undiagnosed. Though, many researchers and clinicians increasingly focus on AV in case of vaginitis symptoms. Majority of studies have used Gram stain microscopy to reveal lactobacilli and other bacteria, rate of leukocytes and presence of parabasal epitheliocytes [
15,
16]. We have applied similar diagnostic method on the group of Vietnamese pregnant women. In addition to microscopy, we performed also cultures that revealed common list of bacteria that is characteristic to AV.
S. agalactiae, enterococci and
S. aureus were the three predominant bacteria in our AV patients being somewhat different from the previous studies [
6,
9], where
E. coli made up 80—90% of AV cases in pregnant women. To date, there is not much information on the consistency between AV diagnostic criteria by microscopy and culture, thus explaining the discrepancies between the results of the current and previous studies.
One component of the AV-complex, the group B streptococci (GBS) or
S. agalactiae is a common causative agent of neonatal infections, therefore routine screening of this species at the end of pregnancy has been recommended [
17‐
19]. Its prevalence is about 10–30% being fairly higher than the reported AV during pregnancy in the present study. It must be pointed that the presence of this species does not mean AV diagnosis. Our previous study revealed 6% and 9.5% when a culture or SYBR Green real-time PCR technique was applied to 116 pregnant women to detect GBS, respectively [
20]. In the present study, culture revealed this species from 10 women (3%) from the 323 cohort of pregnant women.
The best treatment for AV patients is not yet fully determined, but it must be tailored according to the microscopic findings and the patient's needs. There is a role for estrogen therapy, corticosteroids, local antiseptics, antibiotics (clindamycin, moxifloxacin, carbapenems, combinations of beta-lactam with beta-lactamase inhibitor and others) and probiotics, depending on the leading symptoms and comorbidity. Metronidazole that is commonly used to treat BV is not effective in case of AV [
3]. It cannot be forgotten that AV can co-occur with other genital tract infections, such as BV, candidiasis and sexually transmitted infections, therefore these diseases need to be ruled out in AV patients.
Prevalence of AV in pregnancy
The prevalence of AV is largely unknown, both in pregnant and unpregnant women, since routine diagnostics in case of vaginitis includes detection of BV, candidiasis and trichomoniasis but not AV. Since reliable tests are not commercially available and wet mounts are not routinely performed in many settings, the AV is likely underreported. Therefore, it has been recommended to include AV in the diagnostic workup for patients seeking medical care for vaginitis symptoms [
10].
In our study, the prevalence of AV in late pregnancy was 15.5%, mainly representing the light form of the disease (13.0%) and lesser extent (2.5%) the moderate form of AV. No severe cases were found in our study. These numbers are generally comparable to the previous reports where the prevalence of AV has been reported to be 4–8% in pregnant women and 7–13% in non-pregnant women [
3]. Hence, the AV is less common than BV, and it tends to be slightly less represented among pregnant than non-pregnant women, being probably associated with physiological changes in the vaginal environment. However, according to some studies the AV in pregnant women may be even higher in prevalence than in non-pregnant women [
21]. Still, the available data is not yet sufficient to draw the final conclusions, and moreover, the used methods have varied between the studies.
Impact of AV on pregnancy outcome
Previous studies have revealed that AV can be related to different pregnancy complications like chorioamnionitis, fetal infection, premature birth, PROM and pPROM as well as neonatal infections [
3,
4], less data is available about postpartum infections [
4]. Thus, the profile of complications tends to be quite similar to that of BV.
In our study, 6% of pregnant women diagnosed with AV had a puerperal sepsis, whereas in the group without AV, the rate was only 0.7%. Thus, the AV patients had almost nine-fold increased risk (OR = 8.65) of postpartum infection compared to the controls. Postpartum infections account for a significant, and often preventable, portion of the global healthcare burden. They are one of the top five causes of maternal deaths worldwide and account for 10–15% of deaths in the postpartum period. Puerperal sepsis causes more than 75,000 maternal deaths annually [
22]. Abnormal vaginal microbiota forms an important reservoir for causative agents of these infections. This is a strong indication that AV needs to be diagnosed and treated in pregnant women. The reasons why AV can be associated with different neonatal and maternal adverse pregnancy outcomes are not finally known but the bacterial pattern and load may likely have its impact.
The most prevalent bacteria in our AV patients were
S. agalactiae, enterococci and
S. aureus, which may cause puerperal and neonatal infections. The frequency of other pregnancy complications and caesarean section was not different in women with and without AV in our study, and the preterm birth rate was low both groups (2% vs 3.3%). Yet, to the contrary of the current study, some previous studies have revealed association of AV with preterm labor and PROM [
3,
4]. In our research, we analyzed the rate of amniotic fluid rupture at all the gestational ages, without seeing the effect of AV on either PROM or pPROM.
Even though screening and prevention policies for
S. agalactiae have been systematically introduced, early neonatal infection caused by this species continues to be a problem. However, since the turn of the millennium, the neonatal infections have remained on the plateau [
19]. The screening and prevention of neonatal Group B streptococcus infections have currently been primarily successfully applied in term infants, while the current pattern of neonatal infection and mortality caused by
S. agalactiae has still remained the serious risk factor for preterm neonates [
17]. In our study, both groups had similar preterm birth rates, but prevalence of
S. agalactiae was only 3% in the full cohort, being lower than in the general reports. In the current study, the neonatal sepsis was found in 36 neonates (11.1% out of all 323 neonates), while the neonatal sepsis occurred slightly, but insignificantly more frequently in AV (16%) than in non-AV (10.3%) groups of pregnant women. Still, we have no data and solid clue about the etiology of the neonatal infections in our study groups.
Our study has several limitations. Majority of pregnant women were admitted to our hospital during the final weeks before delivery, therefore we have no information about the previous weeks. The women did not get treatment for AV, hence we have no data about the possible effect of treatment. Since the puerperal sepsis was recorded according to clinical criteria and by using stain method, we have no information about the causative agents. In addition, the study was not adequately powered to find outcome differences such as puerperal or neonatal sepsis, instead, the sample size was calculated for detecting the prevalence of AV in pregnant women.