Skip Navigation

Molecular Quantification of Gardnerella vaginalis and Atopobium vaginae Loads to Predict Bacterial Vaginosis

  1. Jean-Pierre Menard1,3,
  2. Florence Fenollar1,2,
  3. Mireille Henry2,
  4. Florence Bretelle3, and
  5. Didier Raoult1,2
  1. 1Unité des rickettsies, IFR 48, CNRS UMR 6020, Faculté de Médecine, Université de la Méditerranée, Marseille, France
  2. 2Pôle de Maladies Infectieuses, Marseille, France
  3. 3Service de Gynécologie Obstétrique, Marseille, France
  1. Reprints or correspondence: Dr. Didier Raoult, Unité des rickettsies, IFR 48, CNRS UMR 6020, Faculté de Médecine, Université de la Méditerranée, 27 Blvd. Jean Moulin, 13385 Marseille Cedex 5, France (didier.raoult{at}gmail.com).
 
Next Section

Abstract

Background. Bacterial vaginosis (BV) is a poorly detected public health problem that is associated with preterm delivery and for which no reliable diagnostic tool exists.

Methods. Molecular analysis of 231 vaginal samples, classified by Gram stain–based Nugent score, was used to propose molecular criteria for BV; these criteria were prospectively applied to 56 new samples. A quantitative molecular tool targeting 8 BV-related microorganisms and a human gene was developed using a specific real-time polymerase chain reaction assay and serial dilutions of a plasmid suspension. The targeted microorganisms were Gardnerella vaginalis, Lactobacillus species, Mobiluncus curtisii, Mobiluncus mulieris, and Candida albicans (which can be identified by Gram staining), as well as Atopobium vaginae, Mycoplasma hominis, and Ureaplasma urealyticum (which cannot be detected by Gram staining).

Results. With use of the Nugent score, 167 samples were classified as normal, 20 were classified as BV, and 44 were classified as intermediate. Except for U. urealyticum, M. mulieris, and Lactobacillus species, DNA of the tested bacteria was detected more frequently in samples demonstrating BV, but the predictive value of such detection was low. The molecular quantification of A. vaginae (DNA level, ⩾108 copies/mL) and G. vaginalis (DNA level, ⩾109 copies/mL) had the highest predictive value for the diagnosis of BV, with excellent sensitivity (95%), specificity (99%), and positive (95%) and negative (99%) predictive values; 25 (57%) of the samples demonstrating intermediate flora had a BV profile. When applied prospectively, our molecular criteria had total positive and negative predictive values of 96% and 99%, respectively.

Conclusions. We report a highly reproducible, quantitative tool to objectively analyze vaginal flora that uses cutoff values for the concentrations of A. vaginae and G. vaginalis to establish the molecular diagnosis of BV.

Bacterial vaginosis (BV) is an important public health problem that affects the normal vaginal flora; BV of unknown etiology has been reported in 8%–40% of fertile women [1, 2]. Mounting evidence has associated BV with susceptibility to sexually transmitted diseases, premature labor and delivery, and low–birth-weight infants [1, 2]. In the United States, where preterm birth is the leading cause of infant morbidity and mortality, the overall rate of preterm delivery is ∼11%; among the population at risk of BV, the rate of preterm delivery is estimated to be 30%, and the associated cost is $1 billion per year [2]. Consequently, medical management of BV may help to reduce the rate of preterm delivery. A recent literature review, however, has revealed conflicting results for the efficiency of screening and treating BV in pregnant women [3, 4]. The results were limited by the significant heterogeneity of the included studies. One factor contributing to the inconsistencies in these studies may be differences in screening methods because of the lack of a reliable standardized and objective diagnostic tool [3, 4].

The BV syndrome was first defined by the clinical criteria of Spiegel et al. [5]. However, a clinical diagnosis is limited when attempting to evaluate an asymptomatic population, as observed for one-half of pregnant women with BV [6, 7]. Therefore, this approach is not routinely used. Most of the published articles have based the diagnosis of BV on the Nugent score (NS) [7]. With this technique, the diagnosis of BV is based on a quantitative estimation of the number of Lactobacillus, Gardnerella vaginalis, and Mobiluncus morphotypes by Gram stain [8]. Currently, the NS requires a fastidious method that is not standardized, even if the κ value is high [9–13]. Thus, NS results in the artificial category of intermediate flora correspond to 8%–22% of samples; however, this category remains uncharacterized [6, 14–16]. Finally, this method does not permit the identification of several bacteria implicated in BV, such as Mycoplasma species, which lack a cell wall, and Atopobium vaginae, which presents a variable morphology, leading to misidentification [17, 18]. The involvement of BV in pregnancy outcome supports the urgent need for accurate diagnosis.

Our aim was to propose objective molecular cutoff values for the definition of BV in a study with 2 phases: (1) development and (2) validation. For this purpose, a quantitative molecular tool targeting 8 microorganisms and human albumin was developed using a specific quantitative real-time PCR (qPCR) assay and serial dilutions of a plasmid suspension. This strategy was previously used by our team to detect potential agents of bioterrorism [19] and to quantify human papillomavirus [20]. The quantitative molecular criteria for predicting BV were determined using the NS as a reference.

Previous SectionNext Section

Patients, Materials, and Methods

Patients

Development phase. A total of 231 vaginal samples were prospectively obtained from 204 women during a pregnancy follow-up period at Conception University Hospital of Marseille (France) from June 2005 through April 2006. Twenty-one women had >1 sample obtained. After informed consent, vaginal samples were collected from symptomatic and asymptomatic patients. Vaginal samples were used either for screening for BV during the first trimester for patients who had a prior preterm delivery or for screening for Streptococcus agalactiae during the third trimester. After placement of a nonlubricated speculum into the vaginal vault, sampling was performed with 2 sterile cytobrushes rotated against the vaginal wall (scrinet, 5.5 mm; Laboratory CCD International). A sample from 1 cytobrush was rolled onto a glass slide for Gram staining. The second cytobrush was transferred to a sterile tube containing 500 μL of BME Baral Medium (Gibco) for DNA extraction and was stored at −80°C until use.

Validation phase. From a new cohort of 56 pregnant women at the North University Hospital of Marseille (France), 56 vaginal samples were collected prospectively from May through June 2007 with the technique described above.

Nugent Score

Gram staining was performed on the vaginal smear samples using an automated stainer (Model 7320 Aerospray Gram Slide Stainer); and the samples were graded, according to the NS [8], as demonstrating normal flora, intermediate flora, or BV. The assessment was performed independently by 2 investigators who were unaware of each other's results. Discordant results were reviewed until a consensus was reached.

Construction of Molecular Tools

Nine DNA sequences targeting 8 microorganisms (Lactobacillus species, G. vaginalis, Mobiluncus curtisii, Mobiluncus mulieris, Ureaplasma urealyticum, A. vaginae, Candida albicans, and Mycoplasma hominis) and a human gene were analyzed by qPCR. Human albumin gene quantification was used as an internal control to provide evidence for DNA presence, DNA quality, and PCR inhibitors. The DNA targets, primers, probes sequences, and oligonucleotides used for qPCR are shown in table 1. Primers and oligonucleotides were synthesized by Eurogentec, and probes were synthesized by Applied Biosystems. Duplex qPCR reactions were performed in double fluorescence (FAM-labeled probe and VIC-labeled probe) for 8 targets (table 2). Only 1 target (M. hominis) was quantified alone. The specificity of primers and probes was controlled by BLAST analysis, and the primers and probes were tested against specific strains of each microorganism analyzed in this study and against 40 additional bacterial species (table 3). The construction of the nucleotidic fragment containing the 9 DNA targets and the quantification plasmid is shown in the [Appendix (online only)[Appendix.

Figure 1
View larger version:
Figure 1

Scattergram showing microbial concentration according to real-time PCR quantification for 20 samples demonstrating bacterial vaginosis (left) and 167 samples demonstrating normal vaginal flora (right), as determined by Nugent score.

Figure 2
View larger version:
Figure 2

Percentage of flora with Gardnerella vaginalis (dotted line), Atopobium vaginae (continuous line), and Lactobacillus species (dotted line), according to threshold quantifications in 167 samples demonstrating normal vaginal flora (gray) and 20 samples demonstrating bacterial vaginosis (black).

Figure 3
View larger version:
Figure 3

Receiver operating characteristic (ROC) curves for bacterial molecular counts used to predict bacterial vaginosis. The closer the area under the curve is to 1.0, the better the bacterial counts predict bacterial vaginosis. Atopobium vaginae (A) and Garnerella vaginalis (B) counts have the best predictive power for bacterial vaginosis. C, Lactobacillus species ROC curve. D, Mobiluncus curtisii ROC curve. E, Mycoplasma hominis ROC curve. F, Ureaplasma urealyticum ROC curve.

Table 1
View larger version:
Table 1

DNA targets according to GenBank and nucleotide sequences of primers and probes used for quantitative realtime PCR.

Table 2
View larger version:
Table 2

Optimal volumes and concentrations of primers and probes used for quantitative real-time PCR, to obtain the same sensitivity and quantification in simplex and duplex quantitative real-time PCR.

Table 3
View larger version:
Table 3

Microorganisms used to test sensitivity and specificity of each quantitative real-time PCR targeting the 8 microorganisms.

Vaginal Samples Analysis Using Molecular Tools

DNA extraction. Two hundred microliters of each homogenized vaginal suspension were used for DNA purification with the QIAamp DNA Mini Kit (Qiagen), which was modified as described elsewhere [20].

qPCR assay. The qPCR assay was performed using a Stratagene MX 3000P. The amplification program was run at 50°C for 2 min and at 95°C for 15 min, followed by 45 cycles at 95°C for 30 s and at 60°C for 1 min. Five microliters of (1) a pure undiluted DNA sample, (2) a DNA sample diluted to 1×10 μL, (3) a DNA sample diluted to 1×100 μL, or (4) the serially diluted plasmid suspension was added to the 20-μL PCR mix that contained the Quantitect Probe PCR Kit mix (Qiagen), the 2 couples of primers, the 2 probes (1 labeled “FAM” and 1 labeled “VIC”), and 100 U Uracil DNA glycosylase (Sigma-Aldrich). Negative controls were introduced in each reaction plate.

The final results were expressed as copies of microorganism DNA per 1 mL of vaginal suspension. The quantification was obtained by multiplying the number of copies per 5 μL of extracted DNA by 100. Indeed, 5 μL of DNA sample were obtained from 100 μL of eluded DNA, which correspond to the DNA contained in 200 μL of vaginal sample suspension.

Statistical Analysis

The quantitative distribution of the microorganisms in flora demonstrating BV and normal flora was analyzed using the Mann-Whitney U test in EpiInfo, version 6.04a (Centers for Disease Control and Prevention). Differential expression was considered to be statistically significant when P<.05. To propose objective molecular cutoff values for the definition of BV, the sensitivity and specificity of the NS were determined. For the most important results, 95% CIs were calculated using the UBC Bayesian Calculator Type 2 [22].

Previous SectionNext Section

Results

Validity of the PCR method. The conditions of the duplex qPCR were adjusted. Optimal quantities of primers and probes are shown in table 1. The cycle threshold values for the standard curve varied from 17 (107 copies) to 37 (1 copy), and the slope varied from −3.2 to −3.5. The linearity remained excellent. All plasmid scale solutions were tested in duplicate. The results were highly reproducible. They enabled a lower detection limit for all microorganisms of 10 copies per 5 μL of sample (data not shown). In each reaction plate, all DNA samples were tested undiluted and diluted to 1×10 μL and 1×100 μL. The CT values had excellent reproducibility and linearity. Finally, the interpretation of microbial quantification was performed only if there was a narrow range of values for the number of albumin copies in the vaginal samples. Among the 231 samples, the CT values for albumin level varied from 19 to 22 copies per 5 μL of sample (106–105 copies per 5μL of sample). Thus, none of the samples were excluded from the analysis.

NS. Seventeen discrepancies (7.4%) among the 231 samples were observed for low Lactobacillus count because of the difficulty in assigning a small gram-positive rod to the Lactobacillus morphotypes. After consensus was reached, of the 231 samples, 167 (72%) were categorized as demonstrating normal flora, 20 (9%) were categorized as demonstrating BV, and 44 (19%) were categorized as demonstrating intermediate flora.

Qualitative PCR comparison between BV and normal flora. PCR results for samples demonstrating BV and normal flora, as determined by NS, are shown in figures 1 and 2 and in table 4. PCR was positive for A. vaginae more often (P<.001) for samples demonstrating BV (19 of 20 samples) than for samples demonstrating normal flora (116 of 167), with a low positive predictive value (PPV) of 14% for BV diagnosis. The same profile of results was observed for PCR for G. vaginalis (PPV, 19%; 19 of 20 samples demonstrating BV vs. 79 of 167 samples demonstrating normal flora; P<.001), PCR for M. curtisii (PPV, 38%; 13 of 20 samples vs. 21 of 167 samples; P<.001), and PCR for M. hominis (PPV, 31%; 10 of 20 samples vs. 22 of 167 samples; P<.001). In contrast, results of PCR for Lactobacillus species were positive less frequently (P=.007) for samples demonstrating BV (12 of 20 samples) than for samples demonstrating normal flora (141 of 167), with a low PPV (8%). No statistically significant difference was observed between samples demonstrating BV and samples demonstrating normal flora with use of PCR for C. albicans (P=.11) and PCR for U. urealyticum (P=.14). No samples demonstrating BV or normal flora were positive for M. mulieris by PCR.

Table 4
View larger version:
Table 4

Quantitative molecular results and analysis for 167 samples demonstrating normal vaginal flora (NVF) and 20 samples demonstrating bacterial vaginosis (BV), as determined by the Nugent score.

qPCR comparison between samples demonstrating BV and samples demonstrating normal flora. Median concentrations of A. vaginae, G. vaginalis, M. curtisii, and M. hominis were significantly higher in samples demonstrating BV than in samples demonstrating normal flora, whereas the median concentration of Lactobacillus species was significantly lower in samples demonstrating BV than in samples demonstrating normal flora (figure 1). To optimize the molecular techniques for routine practice, an adjusted quantification was required (figures 2 and 3). A threshold quantification of A. vaginae DNA level ⩾108 copies/mL had excellent sensitivity (90%), specificity (99%), negative predictive value (99%), and PPV (95%) for the diagnosis of BV. For a threshold quantification of G. vaginalis DNA, a concentration ⩾109 copies/mL was required, and the sensitivity and specificity were 50% and 100%, respectively. However, the negative predictive value (94%) and PPV (100%) remained high. For a threshold quantification of M. curtisii DNA, a concentration ⩾105 copies/mL was required, and the sensitivity and specificity were 45% and 100%, respectively. For a threshold quantification of M. hominis DNA, a concentration ⩾106 copies/mL was required, and the sensitivity and specificity were 30% and 99%, respectively. In contrast, an increasing concentration of Lactobacillus species was predictive of normal flora. Indeed, the threshold quantification of Lactobacillus DNA concentration ⩾108 copies/mL had a high specificity (100%) and moderate sensitivity (44%) for normal flora. Among the 20 samples demonstrating BV, 19 had either an A. vaginae DNA level ⩾108 copies/mL or a G. vaginalis DNA level ⩾109 copies/mL. In addition, 9 had both an A. vaginae DNA level ⩾108 copies/mL and a G. vaginalis DNA level ⩾109 copies/mL.

Predictive quantitative criteria for BV. The combination of an A. vaginae DNA level ⩾108 copies/mL and a G. vaginalis DNA level ⩾109 copies/mL demonstrated the best predictive criteria for BV diagnosis, with an excellent sensitivity (95%), specificity (99%), negative predictive value (99%), and PPV (95%).

Molecular profile of intermediate flora. Thirty-one (70%) of the 44 samples demonstrating intermediate flora were positive for G. vaginalis by PCR, and 12 (27%) of 44 samples had a G. vaginalis DNA level ⩾109 copies/mL. Thirty-seven (84%) of the 44 samples demonstrating intermediate flora were positive for A. vaginae by PCR, and 24 (55%) of the 44 samples had an A. vaginae DNA level ⩾108 copies/mL. Using both quantification assays, 25 (57%) of the 44 samples had an A. vaginae DNA level ⩾108 copies/mL and/or a G. vaginalis DNA level ⩾109 copies/mL.

Validation phase. According to the NS of the 56 pregnant women, 39 (69.6%) were considered to have normal flora, 7 (12.5%) were considered to have BV, and 10 (17.9%) were considered to have intermediate flora. The quality and reproducibility of PCR results were validated as described in Patients, Materials, and Methods. According to the quantitative criteria for BV, 11 of 56 women were considered to have BV. On the basis of the NS of these 11 women, 7 were considered to have BV, and 4 were considered to have intermediate flora. None of the samples demonstrating normal flora were considered to demonstrate BV on the basis of the molecular criteria.

Previous SectionNext Section

Discussion

Development of a rational tool for the diagnosis of BV is critical, because there is a lack of reliable, standardized, and objective diagnostic tools [3, 4]. The Amsel criteria and the NS are the 2 currently available diagnostic methods and are often used in published studies. However, they are not routinely used by physicians. In a preliminary survey detailing the techniques used for diagnosis of BV among 79 interviewed gynecologists, 16%, 20%, and 6% used Amsel criteria, NS, or both techniques, respectively ([Appendix (online only)[Appendix).

We developed a quantitative molecular tool targeting BV-related microorganisms and a human gene, using specific qPCR and a standard plasmid scale solution. We first compared the molecular analysis of the samples that had been classified by the NS as demonstrating normal flora or BV to define the molecular criteria for BV. The accuracy of standard quantification was assessed on the basis of the linearity of DNA amplification of samples with known concentrations and the reproducibility of the quantification in each PCR run. In addition, we validated our molecular criteria in an independent cohort.

One of the important aspects of our method was the ability to identify false-negative samples that result either from an insufficient amount of DNA or the presence of PCR inhibitors. Indeed, human albumin was used as an internal control to provide evidence for the DNA load in vaginal samples to check sampling quality. This manipulation ensured accurate sample comparisons. Thus, samples with insufficient cells could be detected and excluded from the analysis; this was the major missing element in previous studies that attempted to estimate bacterial load in vaginal samples [10, 23, 24]. Furthermore, we used the N-glycosylase dUTP system to avoid false-positive PCR results attributable to technical factors, such as carry-over contamination with amplicon DNA generated from prior PCR [25]. Thus, our method allows reproducible and accurate bacterial quantification.

A. vaginae has recently been implicated in BV [17, 26]. The bacterium was detected by PCR in 96% of patients with BV and in only 12%–19% of those with normal flora [17, 23, 27]. Our data confirm the implication of A. vaginae in BV; however, we detected this bacterium in a greater proportion (69%) of samples classified as demonstrating normal flora [23, 27]. This suggests that mere detection of A. vaginae has a poor predictive value; however, our results also suggest that quantification of A. vaginae is a good predictor of BV. This contradicts the results of the study by Bradshaw et al. [23], in which detection of A. vaginae was more predictive of BV than was the determination of a threshold bacterial load (which reduced sensitivity). This discrepancy between our study and that of Bradshaw et al. [23] may be linked to differences between epidemiological characteristics and/or PCR assays. Geographical and/or ethnic origin, pregnancy status, risks of STDs, and prevalence of BV in the population studied by Bradshaw et al. [23] were shown to influence the rate of A. vaginae in the vaginal flora. PCR technical parameters, such as DNA target, directly influence the sensitivity of PCR assays. For example, it has been clearly demonstrated that reducing the length of the same DNA target by ∼250 base pairs increases the sensitivity of the PCR assay from 24% to 39% [28, 29]. Because our 16S rDNA target for A. vaginae (length, 86 base pairs) was shorter than that used by Bradshaw et al. [23] (length, 430 base pairs), we can hypothesize that our molecular tool is more sensitive.

G. vaginalis is a well-known bacterium that has been implicated in BV. Our results are concordant with previous molecular approaches that reported a high load of G. vaginalis in most patients with BV [10, 23, 24]. Normal flora is dominated by a limited number of Lactobacillus species, such as Lactobacillus gasseri, Lactobacillus crispatus, Lactobacillus jensenii, and Lactobacillus iners, which can be detected by our assay [17, 27]. The decrease in the quantity of Lactobacillus species is one of the major characteristics described in BV [8, 30]; this finding has been confirmed by molecular approaches [10, 24]. M. curtisii and M. mulieris have also been implicated in BV and are included in the NS [30–32]. In our study, M. curtisii was found to be associated with BV, whereas the quantity of M. mulieris was not increased even in normal flora. M. hominis is not included in the NS. However, an increase in the concentration of this bacterium, in contrast to U. urealyticum, is associated with BV [10, 30, 33].

In our study, we demonstrated that the quantification of several microorganisms allows for a molecular diagnosis of BV. In fact, there was a noticeable balance between the Lactobacillus load and the A. vaginae and G. vaginalis loads. The coexistence of A. vaginae and G. vaginalis was recently documented in 78%–96% of samples demonstrating BV, whereas this association was only present in 5%–10% of samples demonstrating normal flora [17, 23, 26]. We also found high concentrations of both bacteria in samples demonstrating BV. To note, A. vaginae is even more specific than G. vaginalis in BV, as suggested by Bradshaw et al. [23]. Finally, we demonstrated that the combination of an A. vaginae DNA level ⩾108 copies/mL and a G. vaginalis DNA level ⩾109 copies/mL was the best diagnostic definition of BV.

Intermediate flora were considered to be either a transitional step between normal flora and BV [26] or a heterogeneous flora including BV and normal flora [11, 15, 16]. Recently, PCR assays have shed some light on intermediate flora, suggesting a profile more similar to that of BV [23]. Indeed, the majority of samples demonstrating intermediate flora had G. vaginalis (92% of samples), A. vaginae (78%), or both organisms (75%) detected [23]. In our study, nearly one-half of the samples demonstrating intermediate flora satisfied quantitative criteria used to define BV, confirming the heterogeneous character of intermediate flora. These data also suggest that the NS severely lacks sensitivity, compared with the qPCR assay used in our study. This gap may be explained by the fact that A. vaginae, which is significantly associated with BV, could not be detected using Gram staining and NS.

There were a few limitations in our study. The spectrum of targeted microorganisms—even if those tested are significant in BV—was narrow. The population recruited was from only 2 locations in France. The results of our study might depend on the studied patients and may not apply to pregnant women from other ethnic origins or to nonpregnant women. Finally, this molecular approach of BV diagnosis might seem complex; however, a similar molecular strategy already exists for point-of-care diagnosis of group B Streptococcus infection at delivery [34, 35].

We report a reliable molecular method to objectively analyze vaginal flora. Currently, no standardized tool exists for BV diagnosis. Clearly defining BV will enable international randomized, therapeutic, clinical studies to be performed.

Previous SectionNext Section

Acknowledgments

We thank Melanie Ihrig, for reviewing the manuscript; the Crédits Ministériels Projet Hospitalier de Recherche Clinique 2006 (Caractérisation et Impact des Anomalies de la Flore Aaginale Chez la Femme Enceinte); and Drs. Antoine, Baytur, Bongain, Boulvain, Carbonne, Chauveaux, Closset, Coulon, Decebal, Denoual, Fernandez, Fournie, Jurouk, Gamad, Garbin, Gatterer, Haddad, Horovitz, Hourdin, Ceausu, Fouda, Kalis, Langer, Lavergne, Lopez, Madelenat, Magnin, Maillet, Maisonneuve, Marchesini, Megid Ali, Ohl, Peintinger, Perrotin, Perschler, Picone, Pierre, Prettenhofer, Raudrand, Schaub, Souames, Subtil, Van Mieghem, Vayssiere, and Vandittelli, for their contributions in explaining their current strategy used for the diagnosis of bacterial vaginosis.

Financial support. Programme Hospitalier de Recherche Clinique 2006.

Potential conflicts of interest. The Université de la Méditerranée deposed a patent on this study for which the authors are the inventors.

  • Received September 20, 2007.
  • Revision received February 20, 2008.
Previous Section
 

References

  1. ↵
    1. Marrazzo JM
    . A persistent(ly) enigmatic ecological mystery: bacterial vaginosis. J Infect Dis 2006;193:1475-7.
    FREE Full Text
  2. ↵
    1. Koumans EH,
    2. Kendrick JS
    . Preventing adverse sequelae of bacterial vaginosis: a public health program and research agenda. Sex Transm Dis 2001;28:292-7.
    CrossRefMedlineWeb of Science
  3. ↵
    1. Varma R,
    2. Gupta JK
    . Antibiotic treatment of bacterial vaginosis in pregnancy: multiple meta-analyses and dilemmas in interpretation. Eur J Obstet Gynecol Reprod Biol 2006;124:10-4.
    CrossRefMedlineWeb of Science
  4. ↵
    1. McDonald HM,
    2. Brocklehurst P,
    3. Gordon A
    . Antibiotics for treating bacterial vaginosis in pregnancy. Cochrane Database Syst Rev 2007;1:CD000262.
    Medline
  5. ↵
    1. Spiegel CA,
    2. Amsel R,
    3. Holmes KK
    . Diagnosis of bacterial vaginosis by direct gram stain of vaginal fluid. J Clin Microbiol 1983;8:170-7.
  6. ↵
    1. Goffinet F,
    2. Maillard F,
    3. Mihoubi N,
    4. et al
    . Bacterial vaginosis: prevalence and predictive value for premature delivery and neonatal infection in women with preterm labour and intact membranes. Eur J Obstet Gynecol Reprod Biol 2003;108:146-51.
    MedlineWeb of Science
  7. ↵
    1. Joesoef M,
    2. Schmid G
    . Bacterial vaginosis. Clin Evid 2002;8:1592-600.
    Medline
  8. ↵
    1. Nugent RP,
    2. Krohn MA,
    3. Hillier SL
    . Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin Microbiol 1991;29:297-301.
    Abstract/FREE Full Text
  9. ↵
    1. Forsum U,
    2. Jakobsson T,
    3. Larsson PG,
    4. et al
    . An international study of the interobserver variation between interpretations of vaginal smear criteria of bacterial vaginosis. APMIS 2002;110:811-8.
    CrossRefMedlineWeb of Science
  10. ↵
    1. Sha BE,
    2. Chen HY,
    3. Wang QJ,
    4. Zariffard MR,
    5. Cohen MH,
    6. Spear GT
    . Utility of Amsel criteria, Nugent score, and quantitative PCR for Gardnerella vaginalis, Mycoplasma hominis, and Lactobacillus spp. for diagnosis of bacterial vaginosis in human immunodeficiency virus-infected women. J Clin Microbiol 2005;43:4607-12.
    Abstract/FREE Full Text
  11. ↵
    1. Ison CA,
    2. Hay PE
    . Validation of a simplified grading of Gram stained vaginal smears for use in genitourinary medicine clinics. Sex Transm Infect 2002;78:413-5.
    Abstract/FREE Full Text
    1. Thomason JL,
    2. Anderson RJ,
    3. Gelbart SM,
    4. et al
    . Simplified Gram stain interpretive method for diagnosis of bacterial vaginosis. Am J Obstet Gynecol 1992;167:16-9.
    MedlineWeb of Science
  13. ↵
    1. Hilmarsdottir I,
    2. Hauksdottir GS,
    3. Johannesdottir JD,
    4. Danielsdottir T,
    5. Thorsteinsdottir H,
    6. Olafsson JH
    . Evaluation of a rapid Gram stain interpretation method for diagnosis of bacterial vaginosis. J Clin Microbiol 2006;44:1139-40.
    Abstract/FREE Full Text
  14. ↵
    1. Guerra B,
    2. Ghi T,
    3. Quarta S,
    4. et al
    . Pregnancy outcome after early detection of bacterial vaginosis. Eur J Obstet Gynecol Reprod Biol 2006;128:40-5.
    CrossRefMedlineWeb of Science
  15. ↵
    1. Larsson PG,
    2. Carlsson B,
    3. Fahraeus L,
    4. Jakobsson T,
    5. Forsum U
    . Diagnosis of bacterial vaginosis: need for validation of microscopic image area used for scoring bacterial morphotypes. Sex Transm Infect 2004;80:63-7.
    Abstract/FREE Full Text
  16. ↵
    1. Libman MD,
    2. Kramer M,
    3. Platt R
    . Comparison of Gram and Kopeloff stains in the diagnosis of bacterial vaginosis in pregnancy. Diagn Microbiol Infect Dis 2006;54:197-201.
    CrossRefMedlineWeb of Science
  17. ↵
    1. Verhelst R,
    2. Verstraelen H,
    3. Claeys G,
    4. et al
    . Cloning of 16S rRNA genes amplified from normal and disturbed vaginal microflora suggests a strong association between Atopobium vaginae, Gardnerella vaginalis and bacterial vaginosis. BMC Microbiol 2004;4:16.
    CrossRefMedline
  18. ↵
    1. Ferris MJ,
    2. Masztal A,
    3. Martin DH
    . Use of species-directed 16S rRNA gene PCR primers for detection of Atopobium vaginae in patients with bacterial vaginosis. J Clin Microbiol 2004;42:5892-4.
    Abstract/FREE Full Text
  19. ↵
    1. Charrel RN,
    2. La SB,
    3. Raoult D
    . Multi-pathogens sequence containing plasmids as positive controls for universal detection of potential agents of bioterrorism. BMC Microbiol 2004;4:21.
    CrossRefMedline
  20. ↵
    1. Carcopino X,
    2. Henry M,
    3. Benmoura D,
    4. et al
    . Determination of HPV type 16 and 18 viral load in cervical smears of women referred to colposcopy. J Med Virol 2006;78:1131-40.
    CrossRefMedlineWeb of Science
    1. Misener S,
    2. Krawetz S
    1. Rozen S,
    2. Skaletsky HJ
    . Primer3 on the WWW for general users and for biologist programmers. In: Misener S, Krawetz S, editors. Bioinformatics methods and protocols: methods in molecular biology. Totowa, NJ: Humana Press; 2000. p. 365-86.
  22. ↵
    UBC Bayesian Calculator Type 2. Available at: http://www.healthcare.ubc.ca/calc/bayescalc2.html. Accessed 21 May 2008.
  23. ↵
    1. Bradshaw CS,
    2. Tabrizi SN,
    3. Fairley CK,
    4. Morton AN,
    5. Rudland E,
    6. Garland SM
    . The association of Atopobium vaginae and Gardnerella vaginalis with bacterial vaginosis and recurrence after oral metronidazole therapy. J Infect Dis 2006;194:828-36.
    Abstract/FREE Full Text
  24. ↵
    1. Zariffard MR,
    2. Saifuddin M,
    3. Sha BE,
    4. Spear GT
    . Detection of bacterial vaginosis-related organisms by real-time PCR for Lactobacilli, Gardnerella vaginalis and Mycoplasma hominis. FEMS Immunol Med Microbiol 2002;34:277-81.
    CrossRefMedlineWeb of Science
  25. ↵
    1. Rys PN,
    2. Persing DH
    . Preventing false positives: quantitative evaluation of three protocols for inactivation of polymerase chain reaction amplification products. J Clin Microbiol 1993;31:2356-60.
    Abstract/FREE Full Text
  26. ↵
    1. Verhelst R,
    2. Verstraelen H,
    3. Claeys G,
    4. et al
    . Comparison between Gram stain and culture for the characterization of vaginal microflora: definition of a distinct grade that resembles grade I microflora and revised categorization of grade I microflora. BMC Microbiol 2005;5:61.
    CrossRefMedline
  27. ↵
    1. Fredricks DN,
    2. Fiedler TL,
    3. Marrazzo JM
    . Molecular identification of bacteria associated with bacterial vaginosis. N Engl J Med 2005;353:1899-911.
    CrossRefMedlineWeb of Science
  28. ↵
    1. Honma M,
    2. Ohara Y,
    3. Murayama H,
    4. Sako K,
    5. Iwasaki Y
    . Effects of fixation and varying target length on the sensitivity of polymerase chain reaction for detection of human T-cell leukemia virus type I proviral DNA in formalin-fixed tissue sections. J Clin Microbiol 1993;31:1799-803.
    Abstract/FREE Full Text
  29. ↵
    1. Chakravorty S,
    2. Pathak D,
    3. Dudeja M,
    4. Haldar S,
    5. Hanif M,
    6. Tyagi JS
    . PCR amplification of shorter fragments from the devR (Rv3133c) gene significantly increases the sensitivity of tuberculosis diagnosis. FEMS Microbiol Lett 2006;257:306-11.
    CrossRefMedlineWeb of Science
  30. ↵
    1. Spiegel CA
    . Bacterial vaginosis. Clin Microbiol Rev 1991;4:485-502.
    Abstract/FREE Full Text
    1. Obata-Yasuoka M,
    2. Ba-Thein W,
    3. Hamada H,
    4. Hayashi H
    . A multiplex polymerase chain reaction-based diagnostic method for bacterial vaginosis. Obstet Gynecol 2002;100:759-64.
    CrossRefMedlineWeb of Science
  32. ↵
    1. Schwebke JR,
    2. Lawing LF
    . Prevalence of Mobiluncus spp among women with and without bacterial vaginosis as detected by polymerase chain reaction. Sex Transm Dis 2001;28:195-9.
    CrossRefMedlineWeb of Science
  33. ↵
    1. Thorsen P,
    2. Jensen IP,
    3. Jeune B,
    4. et al
    . Few microorganisms associated with bacterial vaginosis may constitute the pathologic core: a population-based microbiologic study among 3596 pregnant women. Am J Obstet Gynecol 1998;178:580-7.
    CrossRefMedlineWeb of Science
  34. ↵
    1. Honest H,
    2. Sharma S,
    3. Khan KS
    . Rapid tests for group B streptococcus colonization in laboring women: a systematic review. Pediatrics 2006;117:1055-66.
    Abstract/FREE Full Text
  35. ↵
    1. Bergeron MG,
    2. Ke D,
    3. Menard C,
    4. et al
    . Rapid detection of group B streptococci in pregnant women at delivery. N Engl J Med 2000;343:175-9.
    CrossRefMedlineWeb of Science
| Table of Contents

This Article

  1. Clin Infect Dis. (2008) 47 (1): 33-43. doi: 10.1086/588661
  1. AbstractFree
  2. » Full Text (HTML)Free
  3. Full Text (PDF)Free
  4. Supplementary File

Classifications

Share

  1. Email this article

Navigate This Article

Current Issue

  1. March 1, 2012 54 (5)
  1. Clinical Infectious Diseases
  1. Alert me to new issues

Most