Preterm birth continues to be a major challenge in obstetrics 1. Nearly 50% of spontaneous preterm births are associated with ascending genital tract infection 2. Some studies demonstrated that an elevated vaginal pH by itself or combined with an elevated Nugent score was associated with preterm delivery and low birth weight 3. Normal vaginal microflora was associated with a four-fold decreased risk for spontaneous preterm birth 4. Women without abnormalities of the vaginal flora in the first trimester had a 75% lower risk for delivery before 35 weeks compared with women with abnormal vaginal flora 5.
The pH of vaginal secretion varies between 3.8 and 4.5 6. A pH less than 4.5 signifies the absence of vaginitis, whereas a pH of 4.5 or more is classified as vaginitis 7. Some investigators consider a vaginal pH value of 5.0 or higher to be the strongest predictor of bacterial vaginosis (BV) 8 and found a significantly higher incidence of delivery before the 37th week among pregnant women with a vaginal pH of 5.0 or higher. Hauth et al. 9 noted that a vaginal pH of 5.0 or higher in early pregnancy was significantly associated with a higher incidence of preterm labor. Bacterial flora is predominantly of the lactobacillary type (normal) 10. They produce lactic acid from carbohydrate metabolism to maintain a low vaginal pH (<4.5) 11. They generate hydrogen peroxide, which plays a key role in maintaining the health of the vagina 12. When Lactobacillus spp. is deficient, numerous infectious conditions such as BV and aerobic vaginitis (AV) may develop 13. The lactobacillary grades (LBGs) are: LBGI – lactobacilli are the only organism detected; LBGIIa – mixed flora but predominantly lactobacilli; LBGIIb – mixed flora but proportion of lactobacilli severely decreased; LBGIII – lactobacilli severely depressed or absent because of overgrowth of other bacteria 14. The LBGI and LBGIIa are known to be the normal vaginal flora. The LBGIII and to a lesser extent LBGIIb are said to be abnormal vaginal flora 13.
The absence of lactobacilli (LBGIII) at the first visit was associated with increased risks for preterm birth (PTB) and miscarriage. Patients presenting with an absence of lactobacilli had a 7.4-fold increased risk for spontaneous preterm birth 5.
There is a massive overgrowth of anaerobic and facultative organisms such as Gardnerella vaginalis, Atopobium vaginae, Bacteroides spp., Mobiluncus spp., and genital mycoplasmas 15.
The normally present Lactobacillus spp. are replaced with aerobic organisms, predominantly enteric commensals or pathogens 16. Microscopically, it shows a paucity of lactobacilli, increased number of leukocytes, and presence of parabasal cells as well as cocci or coliform bacteria 17.
The aim of the study was to evaluate the correlation of increased vaginal pH of 4.5 or higher with abnormal vaginal flora and their relationship with cervical length measurement in the prediction of preterm birth at 37 weeks’ gestation or less and early preterm birth at 34 weeks’ gestation or less.
Patients and methods
The study was a prospective study on a cohort of 100 pregnant women attending the Obstetrics Outpatient Clinic at Cairo University from March 2011 to September 2011. The patients were informed about the study and consent was obtained for their enrollment in the study. Local Ethical Committee approval for the study design was obtained.
The studied patients were subjected to the following: complete history taking, physical examination, and vaginal sterile speculum examination including assessment of vaginal pH and vaginal smear slide; three swabs for aerobic and anaerobic culture were taken. Vaginal ultrasound was performed for accurate dating, first trimester routine scanning, and cervical length measurement.
The vaginal fluid pH was measured with phenolphthalein paper. A cutoff point of pH 4.5 was selected. One cotton-tipped swab was used for preparing a vaginal smear slide that was Gram stained, whereas the other two vaginal swabs were transferred on Amies Transport Medium (Oxoid Ltd, Basingstoke, Hampshire, UK) and transported to the microbiology laboratory to be cultured both aerobically and anaerobically.
Vaginal Gram-stained slides
Direct Gram-stained slides were examined under oil immersion at a magnification of 1000 and were classified according to the LBG, Nugent score, and the AV score.
According to the Nugent criteria, the amount of three morphotypes characteristic of BV was quantified and scored: Lactobacillus spp., Mobiluncus spp., and G. vaginalis. The range of scores for these morphotypes was used in the current analysis. For Lactobacillus spp., the scores ranged from 0 to 4, with 0 indicating that 30 or more organisms were found and 4 indicating that no organisms were found in the sample. In contrast, for G. vaginalis, a score of 0 indicated that no organisms were found and the highest score, 4 indicated that 30 or more organisms were found. For Mobiluncus spp., the scores ranged from 0 to 2, with a score of 2 indicating that five or more organisms were identified in the sample and 0 indicating that no organisms were found 18.
The criteria for the microscopic diagnosis of AV are provided in Table 116.
Vaginal swabs were inoculated on the following 20:
- Blood agar plates (Oxoid Ltd), which were incubated aerobically at 37°C for 24–48 h to isolate aerobic and facultative anaerobic bacteria.
- Columbia human blood bilayer agar plates (Oxoid Ltd), which were incubated at 37°C and read after 48–72 h for G. vaginalis isolation.
- Mycoplasma agar with Mycoplasma supplement G (Oxoid Ltd), which was incubated both aerobically and anaerobically at 37°C for 3–5 days.
- De Mann, Rogosa, Sharpe medium (Oxoid Ltd), which was used for the isolation of Lactobacillus spp. after incubation in CO2 at 37°C for 48 h.
Data management and analysis were performed using SigmaStat Program (version 3.5; SigmaStat, San Jose, California, USA). Graphs were drawn using Microsoft Excel (Microsoft, Remington, Washington, USA). Data were statistically described in terms of range, mean±SD, and median. Categorical data were summarized as percentages. Comparisons between categorical variables were made using the χ2-test. All P-values were considered significant when P-values were less than 0.05.
In our cohort, the mean age was 26.79±0.533 years; 83% (n=83) were multipara, whereas 17% (n=17) were primipara. The minimum parity was 0 and the maximum was 5 with a mean parity of 1.65±0.166. The mean parity for patients who had a preterm delivery was 2.07±0.385, whereas the mean parity for patients who had a term delivery was 1.56±0.126. In our cohort, 8% of the patients had a history of preterm birth.
On measuring the cervical length, the minimum length was 2 cm and the maximum was 6 cm. The mean cervical length was 4.02±0.070 cm. The minimum vaginal pH was 4 and the maximum pH was 6. The mean vaginal pH was 4.29±0.049. Three patients had missed abortion (3%). Seven patients had urinary tract infection (7.3%) evidenced by complete urine analysis. Thirteen patients had threatened abortion (13.5%). Symptomatic vaginitis was present in 26% of the patients in their first visit. The minimum gestational age at delivery was 28 weeks and the maximum was 41 weeks, with a mean of 38.17 weeks (SD=2.311). Delivery at 37 weeks or less was 14.6% and delivery at 34 weeks or less was 6.3%.
On analysis of our results using the LBG, we found that the prevalence of LBGI was 8.3%, LBGIIa was 20.8%, LBGIIb was 34.4%, and LBGIII was 36.5%. Using the Nugent score, the prevalence of the normal flora was 19.9%, intermediate flora was 43.8%, and BV was 36.8%. Using the AV score, the prevalence of mild AV was 18.8%, moderate AV was 24%, and severe AV was 21%. The prevalence of Mycoplasma hominis was 60.4%.
Increased vaginal pH (≥4.5) among our cohort was 31%; 16.6% of them had normal flora, 43.3% had intermediate flora, and 43.3% had BV. The mean vaginal pH in the preterm delivery at 37 weeks or less was 4.64±0.613, which was higher than that in the term delivery at more than 37 weeks, which was 4.26±0.309. The mean pH in the preterm delivery at 34 weeks or less was 4.83±0.683, which was higher than that in the term delivery at more than 34 weeks, which was 4.28±0.34. A statistically significant higher difference in the mean vaginal pH was found in the preterm delivery at 37 weeks or less than in the term delivery at more than 37 weeks (P=0.025). A statistically significant higher difference in the mean vaginal pH was found in the preterm delivery at 34 weeks or less than in the term delivery at more than 34 weeks (P=0.03).
The mean cervical length for patients who had a preterm delivery at 37 weeks was 3.97±0.708 cm, whereas the mean cervical length in the term group was 4.11±0.702 cm. This was statistically nonsignificant (P=0.54). The cervical length was directly related to the delivery week but was statistically nonsignificant (P=0.189). No statistically significant difference between M. hominis-positive and M. hominis-negative patients was found with respect to delivery at 37 weeks or less or at 34 weeks or less.
The vaginal pH can be used for prediction of preterm labor (≤37 weeks) [odds ratio (OR) 9.545, 95% confidence interval (CI) 2.25–40.48, P=0.002, overall accuracy 88.5%, sensitivity 21.4%, and specificity 100%]. It can also be used to predict early preterm labor (≤34 weeks) (OR 19.6, 95% CI 2.48–154.88, P=0.005, overall accuracy 95.8%, sensitivity 33.3%, and specificity 100%) (Figs 1 and 2).
Many studies suggest that the presence of abnormal vaginal flora early in pregnancy increases the risk for PTB, premature rupture of the membranes, and low birth weight 5,21. In our study, we found that the prevalence of increased vaginal pH (≥4.5) among our cohort is 31%; 16.6% of them had normal flora, 43.3% had intermediate flora, and 43.3% had BV. This pH is slightly higher than that in other studies conducted during the first trimester of pregnancy in which they found a prevalence rate of 19.4% 22. This may be due to racial difference between the two groups.
Our vaginal pH results are consistent with the study performed by Hauth et al. 9 who noted that a vaginal pH of 5.0 or higher in early pregnancy was significantly associated with a higher incidence of preterm birth. This is also in agreement with the prospective study conducted by Jazayeri et al. 8 who found a significantly higher incidence of delivery before the 37th week among pregnant women with a vaginal pH of 5.0 or higher (OR 9.6, 95% CI 2.0–45.5). In our study, the vaginal pH was directly proportional to the Nugent score, LBG, AV score, and cervical length (r=0.205, 0.244, 0.245, and 0.192 and P=0.045, 0.017, 0.055, and 0.061, respectively; P<0.05, significant). This correlates with the results of the study conducted by Zodzika et al. 22 on factors related to increased vaginal pH in early pregnancy.
The prediction for preterm birth at 37 weeks or less using vaginal pH of at least 4.5 among our cohort had an OR 0.25, 95% CI 0.08–0.83, P=0.023, and overall accuracy 85.4%. This is consistent with the study conducted by Matijevic et al. 23 in which they found that an elevated vaginal pH was a good predictor of preterm labor at 37 weeks or less (LR 3.7, 95% CI 1.3–10.4) and early preterm labor at 34 weeks or less (LR 1.7, 95% CI 1.1–3.1).
In our study, the cervical length was directly related to the delivery week but was statistically nonsignificant. This is in agreement with the prospective study conducted by Carvalho et al. 24 involving 529 pregnant women; they found that the mean cervical length at 11–14 weeks was 42.4 mm. Cervical length at 11–14 weeks was not significantly different between the groups that delivered at term (42.7 mm) and preterm (40.6 mm). The study conducted by Donders et al. 5 also correlates with our result in which they found that the mean cervical length was 4.6±1.09 cm at 10–14 weeks visit (n=790) and was not a valuable tool to predict preterm delivery.
In our study comparing the LBGI with LBGIII for delivery at 37 weeks or less as against delivery at more than 37 weeks, there was a statistically significant difference with P=0.051. This correlates with the study conducted by Donders et al.5 in which they found that the absence of lactobacilli (LBGIII) at the first trimester was associated with increased risks for preterm birth at 37 weeks or less. This is also in agreement with the study by Usui et al. 25 who prospectively examined the vaginal flora in a sample of 1958 pregnant women during the first and second trimesters. They found that 11% delivered before the 37th week (n=224) and in these patients lactobacilli were significantly decreased.
In patients who delivered at 34 weeks or less, all had LBGIII. Comparison of LBGI with LBGIII for delivery at 34 weeks or less as against delivery at more than 34 weeks among our cohort showed a statistically nonsignificant difference with P=0.3. This result is contradictory to the results of the study conducted by Donders et al.5 in which they found that the absence of lactobacilli (LBGIII) at the first trimester was associated with increased risks for preterm birth at 34 weeks or less. This may be possibly due to the small number of patients delivering at 34 weeks or less in our study.
In our study using the Nugent score, 19.8% had normal flora, 36.5% of the patients had BV, and 43.8% had intermediate flora. This is partially comparable with the study conducted by Trabert and Misra 26 in which they found that nearly 50% of pregnant African–American women had BV in National Health and Nutrition Examination Survey. This prevalence rate was higher than that in the study conducted by Donders et al.5 on 1026 pregnant women in which they had 81.4% of the patients with normal flora, 8.5% had BV, and 11.3% had intermediate flora. This difference may be due to smaller number of participants in our study.
In our study, comparison of the patients with normal flora as against those with BV with respect to delivery at 37 weeks or less as against after 37 weeks showed statistical significance (P=0.01).
The Nugent score was inversely related to the delivery week (r=−0.425, P=0.000). This is in agreement with the study by Leitich and Kiss 27 who recently updated a previously published meta-analysis of 32 studies including a total of 30 518 pregnant women and evaluated both BV and intermediate vaginal flora as risk factors for adverse pregnancy outcome. They found that the presence of BV more than doubled the risk for preterm delivery in both asymptomatic pregnant women (OR 2.16, 95% CI 1.56–3.00) and pregnant women with symptoms of preterm labor (OR 2.38, 95% CI 1.02–5.58). In addition, Donders et al.5 found that BV at 10–14 weeks visit was associated with preterm birth (OR 2.4, 95% CI 1.1–4.7) and early preterm birth (OR 5.3, 95% CI 2.1–12.9).
In contrast, Figueroa et al. 28 found that BV by either the Nugent score or pH was not associated with PTB at less than 28 weeks (P=0.92 and 0.55), less than 35 weeks (P=0.13 and 0.09), or less than 37 weeks (P=0.36 and 0.74).
The prevalence of severe and moderate AV among our cohort was 50.9%. This prevalence is much higher than that in other studies. In a study conducted by Donders et al.29 on Belgian low-risk population of 759 pregnant women, they found 8.3% women with coccoid AV. In addition, in a recent study conducted by Zodzika et al. 22, only one pregnant woman of the 139 examined had severe AV, whereas moderate AV was present in 14 (10%) women. In our study, we found that the AV score was inversely related to the delivery week (r=−0.300, P=0.018, significant). This correlates with the studies conducted by Donders et al.5,16 in which they found that microflora suggestive for aerobic bacteria was associated with mid-trimester pregnancy losses and early preterm delivery at less than 35 weeks. The prevalence of M. hominis among our cohort was 60.4%. This was higher than that in the study conducted by Lee et al. 30 in which the prevalence of a positive vaginal culture for genital mycoplasmas was 30%. This may be due to different study population. This was also higher than the prevalence in other studies conducted by Kataoka et al. 31 and Donders et al. 21, which was 11.2 and 1.1%, respectively. This difference may be due to the use of other identification method using PCR. In our study, we also found that 30% of M. hominis-positive patients were associated with BV. This also is higher than the results found by Donders et al. 5 in which they found that 17% of M. hominis-positive patients were associated with BV.
In our cohort, we compared M. hominis-positive with M. hominis-negative patients with respect to delivery at 37 weeks or less as against delivery at more than 37 weeks; it was statistically nonsignificant (P=0.749). We also compared M. hominis-positive with M. hominis-negative patients with respect to delivery at 34 weeks or less as against delivery at more than 34 weeks; it was statistically nonsignificant (P=0.746). This coincides with the study conducted by Lee et al. 30 in which they found that a positive vaginal culture for genital mycoplasmas was not associated with an increased risk for spontaneous preterm birth. However, this disagrees with the study conducted by Donders et al.5 in which nearly half of the patients harboring M. hominis delivered preterm, which, despite low numbers, was significantly different from the number of women with normal flora delivering preterm (43 vs. 8.1%, respectively, OR 8.1, 95% CI 2.8–25.5, P=0.0006).
Limitations to the study
This study was conducted on a small cohort; hence, a need for larger study is required. As a result of the splitting of abnormal flora into several categories, the numbers are rather too small to draw firm conclusions, but as they showed significant differences further investigation is warranted.
Conclusion and recommendations
We conclude that an elevated vaginal pH in early pregnancy was the best predictor of preterm labor in this cohort of pregnant women. We also conclude that different patterns of abnormal vaginal flora can be recognized at the beginning of pregnancy. Specific markers of aerobic vaginal flora, such as the absence of lactobacilli, signs of AV, or partial BV, seem to be better predictors of infection-related PTB than formerly used criteria, such as full BV. The cervical length was not related to abnormalities in the vaginal flora nor had a relationship with the vaginal pH.
Conflicts of interest
There are no conflicts of interest.
1. Hill JL, Campbell MK, Zou GY, Challis JRG, Reid G, Chisaka H, Bocking AD.Prediction of preterm birth in symptomatic women using decision tree modeling for biomarkers.Am J Obstet Gynecol2008;198:468.e1–468.e9.
2. Klein LL, Gibbs RS.Infection and preterm birth.Obstet Gynecol Clin North Am2005;32:397–410.
3. Cauci S, McGregor J, Thorsen P, Grove J, Guaschino S.Combination of vaginal pH
with vaginal sialidase and prolidase activities for prediction of low birth weight and preterm birth.Am J Obstet Gynecol2005;192:489–496.
4. Verstraelen H, Verhelst R, Roelens K, Claeys G, Weyers S, De Backer E, et al..Modified classification of Gram-stained vaginal smears to predict spontaneous preterm birth: a prospective cohort study.Am J Obstet Gynecol2007;196:528.e1–528.e6.
5. Donders GG, Van Calsteren K, Bellen G, Reybrouck R, Van Den Bosch T, Riphagen I, Van Lierde S.Predictive value for preterm birth of abnormal vaginal flora, bacterial vaginosis and aerobic vaginitis during the first trimester of pregnancy.BJOG2009;116:1315–1324.
6. Larsson P-G, Platz-Christensen J-J.The vaginal pH
and leucocyte/epithelial cell ratio vary during normal menstrual cycles.Eur J Obstet Gynecol Reprod Biol1990;38:39–41.
7. Thinkhamrop J, Lumbiganon P, Thongkrajai P, Chongsomchai C, Pakarasang M.Vaginal fluid pH as a screening test for vaginitis.Int J Gynecol Obstet1999;66:143–148.
8. Jazayeri A, Arnold H, Jazayeri MK, Spellacy WN.A prospective study of vaginal pH
as a predictor of preterm delivery.J Matern Fetal Neonatal Med2002;11:30–33.
9. Hauth JC, MacPherson C, Carey JC, Klebanoff MA, Hillier SL, Ernest JM, et al..Early pregnancy threshold vaginal pH
and Gram stain scores predictive of subsequent preterm birth in asymptomatic women.Am J Obstet Gynecol2003;188:831–835.
10. Donders GGG, Desmyter J, Vereecken A, Culver BD, Smally AJ, Sobel JD.Vaginitis.N Engl J Med1998;338:1548–1549.
11. Hawes SE, Hillier SL, Benedetti J, Stevens CE, Koutsky LA, Wølner-Hanssen P, Holmes KK.Hydrogen peroxide-producing lactobacilli and acquisition of vaginal infections.J Infect Dis1996;174:1058–1063.
12. Muench DF, Kuch DJ, Wu H, Begum AA, Veit SJ, Pelletier M-E, et al..Hydrogen peroxide-producing lactobacilli inhibit gonococci in vitro but not during experimental genital tract infection.J Infect Dis2009;199:1369–1370.
13. Donders GGG.Definition and classification of abnormal vaginal flora.Best Pract Res Clin Obstet Gynaecol2007;21:355–373.
14. Donders GGG, Vereecken A, Bosmans E, Dekeersmaecker A, Salembier G, Spitz B.Aerobic vaginitis: abnormal vaginal flora entity that is distinct from bacterial vaginosis.Int Congr Ser2005;1279:118–129.
15. Rampersaud R, Randis TM, Ratner AJ.Microbiota of the upper and lower genital tract.Semin Fetal Neonatal Med2012;17:51–57.
16. Donders GGG, Bellen G, Rezeberga D.Aerobic vaginitis in pregnancy.BJOG2011;118:1163–1170.
17. Donders GGG, Vereecken A, Bosmans E, Dekeersmaecker A, Salembier G, Spitz B.Definition of a type of abnormal vaginal flora that is distinct from bacterial vaginosis: aerobic vaginitis.BJOG2002;109:34–43.
18. Nugent RP, Krohn MA, Hillier SL.Reliability of diagnosing bacterial vaginosis is improved by a standardized method of Gram stain interpretation.J Clin Microbiol1991;29:297–301.
19. Newbern EC, Foxman B, Leaman D, Sobel JD.Desquamative inflammatory vaginitis: an exploratory case–control study.Ann Epidemiol2002;12:346–352.
20. Demba E, Morison L, van der Loeff MS, Awasana AA, Gooding E, Bailey R, et al..Bacterial vaginosis, vaginal flora patterns and vaginal hygiene practices in patients presenting with vaginal discharge syndrome in the Gambia, West Africa.BMC Infect Dis2005;5:12.
21. Donders GG, Van Calsteren C, Bellen G, Reybrouck R, Van den Bosch T, Riphagen I, Van Lierde S.Association between abnormal vaginal flora and cervical length as risk factors for preterm birth.Ultrasound Obstet Gynecol2010DOI: 10.1002/uog.7568 [Epub ahead of print].
22. Zodzika J, Rezeberga D, Jermakova I, Vasina O, Vedmedovska N, Donders G.Factors related to elevated vaginal pH
in the first trimester of pregnancy.Acta Obstet Gynecol Scand2011;90:41–46.
23. Matijevic R, Grgic O, Knezevic M.Vaginal pH
versus cervical length in the mid-trimester as screening predictors of preterm labor in a low-risk population.Int J Gynecol Obstet2010;111:41–44.
24. Carvalho MHB, Bittar RE, Brizot ML, Maganha PPS, Borges Da Fonseca ESV, Zugaib M.Cervical length at 11–14 weeks’ and 22–24 weeks’ gestation evaluated by transvaginal sonography, and gestational age at delivery.Ultrasound Obstet Gynecol2003;21:135–139.
25. Usui R, Ohkuchi A, Matsubara S, Izumi A, Watanabe T, Suzuki M, Minakami H.Vaginal lactobacilli and preterm birth.J Perinat Med2002;30:458–466.
26. Trabert B, Misra DP.Risk factors for bacterial vaginosis during pregnancy among African American women.Am J Obstet Gynecol2007;197:477.e1–477.e8.
27. Leitich H, Kiss H.Asymptomatic bacterial vaginosis and intermediate flora as risk factors for adverse pregnancy outcome.Best Pract Res Clin Obstet Gynaecol2007;21:375–390.
28. Figueroa D, Mancuso MS, Szychowski JM, Paden MM, Owen J.Does mid-trimester Nugent score or high vaginal pH
predict gestational age at delivery in women at risk for recurrent preterm birth?Am J Obstet Gynecol2011;204:46.e1–46.e4.
29. Donders GGG, Odds A, Vereecken A, Van Bulck B, Caudron J, Londers L, et al..Abnormal vaginal flora in the first trimester, but not full-blown bacterial vaginosis, is associated with preterm birth.Prenat Neonat Med1998;3:588–593.
30. Lee SE, Romero R, Kim E-C, Yoon BH.A high Nugent score but not a positive culture for genital mycoplasmas is a risk factor for spontaneous preterm birth.J Matern Fetal Neonatal Med2009;22:212–217.
31. Kataoka S, Yamada T, Chou K, Nishida R, Morikawa M, Minami M, et al..Association between preterm birth and vaginal colonization by mycoplasmas in early pregnancy.J Clin Microbiol2006;44:51–55.