Vaginal Microflora Associated with Bacterial Vaginosis in Japanese and Thai Pregnant Women

Transcription

1 7 Vaginal Microflora Associated with Bacterial Vaginosis in Japanese and Thai Pregnant Women Supaporn Puapermpoonsiri, Naoki Kato, Kunitomo Watanabe, Kazue Ueno, Chompilas Chongsomchai, and Pisake Lumbiganon From the Department of Clinical Microbiology (Faculty of Associated Medical Sciences) and the Department of Obstetrics and Gynecology (Faculty of Medicine), Khon Kaen University, Khon Kaen, Thailand; and the Gifo University School of Medicine (Institute of Anaerobic Bacteriology) and Gifu College of Medical Technology, Gifo, Japan The vaginal flora of 11 Japanese and Thai pregnant women were investigated for the presence of bacterial vaginosis (BV), BV-associated organisms, and BV-associated enzyme. A similar prevalence ofbv was found among the Japanese (.%) and Thai women (1.%). The microbial flora of women with BV were complex; the mean number of isolates recovered in the BV group was ~ times more than that in a group of healthy women. Prevotella species, Porphyromonas species, Peptostreptococcus species, Mobiluncus species, Gardnerella vaginalis, and H r nonproducing lactobacilli were significantly associated with BV. These organisms were less associated with H - producing lactobacilli, which were predominant in women with normal flora, suggesting that H r producing lactobacilli have antibacterial activity against BV -associated organisms. The vaginal sialidase assay by means of a filter-paper spot test was not proved to be a useful screening aid for diagnosis of BV because of the low sensitivity (.%) of this test. Bacterial vaginosis (BV), the most common vaginal disorder in women of reproductive age, is now regarded as a risk factor for complications of pregnancy, including chorioamnionitis and prematurity [1]. Before now, no extensive studies concerning BV and vaginal microflora have been conducted in Asian communities. The purpose of this study was to determine among Japanese and Thai pregnant women the prevalence of BV and the vaginal microflora as well as the vaginal sialidase activity, which may serve as a good indicator for the presence of this syndrome. Materials and Methods A total of pregnant women were enrolled in this study, including 11 Japanese women who visited the prenatal clinic at Iwasa Hospital (Gifu, Japan) and Thai women who visited the clinic at the Srinagarind Hospital of Khon Kaen University (Khon Kaen, Thailand). The ages of the Japanese women ranged from 1 to 7 years (mean ± SD, ±. years), and those of the Thai women ranged from 1 to years ( ±. years). Women were excluded from the study if they had used antibiotics within the previous weeks. Informed consent to undergo examination was obtained verbally. Received January 1; revised May 1. Financial support: This work was supported in part by the Japanese Society for the Promotion of Science (JSPS) and the National Research Council of Thailand. S. P. is a fellow of the JSPS. Reprints or correspondence: Dr. Kunitomo Watanabe, Institute of Anaerobic Bacteriology, Gifu University School of Medicine, Tsukasa-machi, Gifu, Japan; facsimile: Clinical Infectious Diseases 1;:7-1 by The University of Chicago. All rights reserved. 1-//-1$. Three swab specimens were obtained from the posterior vaginal fornix with a polyester fiber-tipped applicator (Becton Dickinson, Franklin Lakes, NJ) and a nonlubricated speculum. These swabs were placed into anaerobic transport medium (Kenki porter; Clinical Supply, Gifu, Japan) and taken to the laboratory within hours. For culture, the first swab was placed into a test tube containing. ml of prereduced anaerobic broth, subjected to vortex, and then removed. Single drops of the vaginal suspension were placed onto agar plates for aerobes and anaerobes, as described elsewhere []. Each agar plate was streaked into four quadrants to obtain semiquantitative counts of organisms. Media for the isolation of anaerobes were incubated for -7 days at 7 C in an anaerobic plastic bag with an oxygen absorbent and CO generator (Anaeromate; Nissui, Tokyo). All other media were incubated at 7 C in % CO for -7 hours. The second vaginal swab was suspended in 1. ml of PBS (PH, 7.). The suspensions were stored at -7 C for subsequent analysis of vaginal sialidase activity by a filterpaper spot test, as described below. A direct vaginal smear was prepared with the third swab for BV-diagnostic gram staining []. Anaerobic bacteria growing in the third and fourth streak zones were identified by a standard method and with use of the RapID/ANA II (Innovative Diagnostic Systems, Atlanta) and gas-liquid chromatography for the identification of fatty acids []. Aerobic and facultative anaerobic isolates were identified by means of the API Identification Systems (Japan bio Merieux, Tokyo). All lactobacilli isolated were tested for production of H by means of a qualitative method []. For the purposes of statistical analysis, women carrying both Hz-producing and H -nonproducing lactobacilli were considered to be positive for the former. Sialidase activity of anaerobic isolates was measured by a filter-paper spot test []. Downloaded from on 1 May 1

2 em 1; (October) Vaginal Flora and Bacterial Vaginosis 7 Table 1. Numbers of isolates from Japanese and Thai pregnant women whose gram-stained vaginal flora had normal, intermediate, and bacterial vaginosis (BY) patterns. No. of isolates per vaginal flora pattern: Japanese women (n) No. of isolates per vaginal flora pattern: Thai women (n) Nonnal Intennediate BV Total Nonnal Intennediate BV Total Isolates (n = 7) (n = ) (n = 1) (n = 11) (n = 1) (n = ) (n = ) (n = ) Total no Mean ± SD (per specimen). ± 1.. ± ± 1..7 ± 1.. ± 1.1. ± ± 1.. ± 1. NOTE. The gram-stain scoring system was as follows: nonnal vaginal flora, score of -; intennediate, -; and BV, 7-1. The Mantel-Haenszel test [], Fisher's exact test, and X tests were used for comparative analysis of the data obtained. Results By the gram-stain scoring system for vaginal smears, pregnant women were categorized as follows: 7 (.1 %) of the 11 Japanese and 1 (.%) of the Thai had normal vaginal flora (scores, -), (.%) of the Japanese and (.%) of the Thai had intermediate vaginal flora (scores, -), and 1 (.%) of the Japanese and (1.%) of the Thai had BV (scores, 7-1). The numbers of isolates from Japanese and Thai pregnant women, classified by vaginal flora patterns, are shown in table 1. The mean number of isolates per specimen from women with BV (both Japanese and Thai) was ~ times more than that from those with normal flora. Frequent vaginal isolates from Japanese and Thai pregnant women, categorized according to the vaginal flora patterns, are listed in tables and, respectively. Prevotella species (mainly Table. Frequency of microflora in the vagina of 11 Japanese pregnant women, stratified by the patterns of their vaginal flora. Microorganisms isolated* No. of isolates recovered Percentage of women (n) with indicated vaginal flora patternt from whom isolates were recovered Nonnal (n = 7) Intennediate (11 = ) BV (11 = 1) P value! Prevotella bivia Prevotella melaninogenica Other Prevotella species Porphyromonas species Mobiluncus species Eubacterium species Clostridium species Bifidobacterium species Actinomyces species Peptostreptococcus asaccharolyticus Peptostreptococcus magnus Other Peptostreptococcus species Gemella morbillorum Gardnerella vagina lis Lactobacillus species HzO,-producing H -nonproducing Group B Streptococcus species Enterococcus species Candida albicans o <. <.1 <. * Microorganisms isolated from <% of the Japanese women tested are not listed (except for Mobiluncus species). t See footnote to table 1 for definitions of the three vaginal flora patterns. j The Mantel-Haenszel test was used to analyze the linear association of the three vaginal flora groups. Downloaded from on 1 May 1

3 7 Puapermpoonsiri et al. eid 1; (October) Table. Frequency of microflora in the vagina of Thai pregnant women, stratified by the patterns of their vaginal flora. Percentage of women (n) with indicated vaginal flora pattern t from whom isolates were recovered No. of isolates Normal Intermediate BV Microorganisms isolated* recovered (n = 1) (n = ) (n = ) P valuel Prevotella bivia or disiens Other Prevotella species Porphyromonas species 1 Mobiluncus species 1 Actinomyces species 11 Peptostreptococcus asaccharolyticus 7 Peptostreptococcus magnus Other Peptostreptococcus species 1 Gardnerella vaginalis 7 Lactobacillus species 1 HO,-producing 1 H O,-nonproducing Group B Streptococcus species Enterococcus species Candida albicans 1 < <. 1 1 < * Microorganisms isolated from <% of the Thai women tested are not listed. t See footnote to table 1 for definitions of the three vaginal flora patterns. I The Mante1-Haenszel test was used to analyze the linear association of the three vaginal flora groups. Prevotella bivia), Porphyromonas species, and Peptostreptococcus species were significantly associated with BV in both Japanese and Thai women (P <.). Mobiluncus species were more prevalent in Thai women with BV (%) than in Japanese women with BV (%). Gardnerella vaginalis was recovered from all Japanese and Thai women with BV except one Thai woman, but it was found in only 1%-1% of those with normal vaginal flora (P <.1). H r producing lactobacilli were the most prevalent organisms in women with normal flora (P <.1), whereas H r nonproducing lactobacilli were the most prevalent in women with BV (P <.1). Group B streptococci, enterococci, and Candida albicans were not significantly associated with BV. The associations between vaginal microorganisms frequently recovered from Japanese and Thai women and the presence of H r producing, H -nonproducing, or no lactobacilli in the vagina are shown in table. All anaerobic gram-negative rods (including P. bivia or Prevotella disiens, Prevotella melaninogenica, other Prevotella species, and Porphyromonas species), Mobiluncus species, Actinomyces species, Peptostreptococcus species, and G. vaginalis were significantly negatively associated with H "'Producing lactobacilli (P <.). Sialidase activity was tested for in the anaerobic vaginal microflora. All strains of Bacteroidesfragilis (), P. bivia (), P. melaninogenica (1), and Prevotella loescheii (), as well as of the 11 isolates of Actinomyces species, were positive for sialidase activity. All strains of Prevotella corporis (11), Porphyromonas asaccharolytica (), Veillonella species (), Mobiluncus species (1), Clostridium species (1), Peptostreptococcus species (), Gemella morbillorum (1), and Streptococcus constellatus () were negative for such activity. Vaginal sialidase activity was detected in (.%) of the women with BV but in only (.%) of the 77 women without BV. All vaginal sialidase-positive, BV-negative women had intermediate vaginal flora. Overall, the vaginal sialidase test for diagnosis ofbv had a sensitivity of.% and a specificity of.%, and the positive and negative predictive values ofthis assay were 7.1 % and.7%, respectively. Sialidase-positive bacteria were recovered from.% of women with BV and 1.% of those without BV. Discussion To our knowledge, this is the first extensive study of the vaginal microflora of pregnant women in Asian countries, which focused on the characterization of women with BV with use of the most recent and advanced techniques and nomenclature for anaerobic bacteria. The prevalences of BV in our populations were comparable to or ~ times lower than those in previous studies [, 7], a circumstance suggesting that BV is roughly as prevalent in the two Asian countries studied as in western countries. Although we were unable to demonstrate the quantitative differences between women with BV and healthy controls, as Downloaded from on 1 May 1

4 eid 1; (October) Vaginal Flora and Bacterial Vaginosis 71 Table. Associations between vaginal microorganisms recovered from 11 Japanese and Thai pregnant women and the presence or absence of H -producing lactobacilli. Percentage of women with or without lactobacilli from whom organism was recovered With Hz-producing With HzOz-nonproducing lactobacilli or no lactobacilli Microorganisms recovered* (n = ) (n = ) P valuet Prevotella bivia or disiens Prevotella melaninogenica Other Prevotella species Porphyromonas species Eubacterium species Mobiluncus species Clostridium species Bifidobacterium species Actinomyces species Peptostreptococcus asaccharolyticus Peptostreptococcus magnus Other Peptostreptococcus species Gemella morbillorum Gardnerella vaginalis Group B Streptococcus species Enterococcus species Candida albicans 7 17 <.1 < <. 1 1 <.1 1 <.1 <. 7 * Microorganisms isolated from <% of the total women tested are not listed. t Per Fisher's exact or X test. shown previously [7 -], Prevotella species and Porphyromonas species-both of which were previously classified as Bacteroides species-and Peptostreptococcus species were isolated frequently and were significantly associated with BV. The high recovery rate of Mobiluncus species, from only Thai pregnant women With BV, was associated with high gram-stain scores (of or 1); such scores were obtained with none of the Japanese women's specimens. These results indicate that Mobiluncus species are positively associated with the higher gram-stain scores. Recovery rates oflactobacilli (%-%) from women with BV in this study were strikingly high, compared with those of previous studies (%-7%) [1]. The culture methods used in this study may have partially contributed to this high recovery rate. Culture on human blood bilayer Tween (HBT) agar facilitated isolation of lactobacilli, because this organism grew very well on the medium; its colony morphology was distinct from G. vaginalis. HBT agar was primarily used for isolation of G. vagina lis. In pregnant women with BV, %-7% oflactobacilli were H r nonproducers. Conversely, >% of lactobacilli were H r producers in women with normal flora patterns. These results strongly support previous studies showing that only H r producing (vs. nonproducing) lactobacilli play an important role in the regulation of vaginal colonization by potential pathogens associated with BV [, 7]. H r producing lactobacilli are bactericidal to G. vaginalis and Mobiluncus species [11]. However, further study is required to reveal the exact spectrum of H -producing lactobacilli and their protective effect against BY. In this study, the vaginal sialidase test for diagnosis of BV had relatively low sensitivity. This may be due to the dilution effect, since we suspended a vaginal swab in 1. ml of diluent and used J.LL of the suspension for the test. A spot sialidase test with nondiluted or less diluted vaginal swabs may improve the sensitivity of the assay enough for it to be a rapid and simple method of screening for BV. We found vaginal sialidase activity in three of the 1 specimens in the absence of sialidaseproducing bacteria (except for Actinomyces species, of which the growth is easily overlooked because of its fastidious and slow-growth characteristics). Actinomyces species may also account for sialidase activity, to some extent, in vaginal fluid. Acknowledgments The authors thank Dr. S. Iwasa, Director of the Iwasa Hospital, and Drs. Kanok Seejoin and Yuthapong Werawatrakul, of the Srinagarind Hospital, for their cooperation in collecting specimens. They also thank Dr. N. Kawakami for assistance with statistical analysis. Downloaded from on 1 May 1

5 7 Puaperrnpoonsiri et ai. em 1; (October) References 1. Hillier SL, Martius J, Krohn M, Kiviat N, Holmes KK, Eschenbach DA. A case-control study of chorioamnionic infection and histologic chorioamnionitis in prematurity. N Engl J Med 1;1:7-.. Summanen P, Baron EJ, Citron DM, Strong C, Wexler HM, Finegold SM. Wadsworth anaerobic bacteriology manual. th ed. Belmont, California: Star Publishing Company, 1.. Nugent RP, Krohn MA, Hillier SL. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of gram stain interpretation. J Clin MicrobioI11;:7-1.. Eschenbach DA, Davick PR, Williams BL, et al. Prevalence of hydrogen peroxide-producing Lactobacillus species in normal women and women witjt bacterial vaginosis. J Clin MicrobioI1;7:1-.. MoncJa BJ, Braham P. Detection of sialidase (Neuraminidase) activity in Actinomyces species by using ' -(-methylumbelliferyl) l'-d-n-acetylneuraminic acid in a filter paper spot test. J Clin Microbiol 1;7: 1-.. Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Nat! Cancer Inst 1;: Hillier SL, Krohn MA, Rabe LK, Klebanoff SJ, Eschenbach DA. The normal vaginal flora, H,orproducing lactobacilli, and bacterial vaginosis in pregnant women. Clin Infect Dis 1; 1(suppl ):S7-1.. Wilks M, Thin RN, Tabaqchali S. Quantitative bacteriology of the vaginal flora in genital disease. J Med Microbiol1; 1: Masfari AN, Duerden BI, Kinghorn GR. Quantitative studies of vaginal bacteria. Genitourin Med 1; :-. 1. Hill GB. The microbiology of bacterial vaginosis. Am J Obstet Gynecol 1; 1:-. II. Skarin A, Sylwan J. Vaginal lactobacilli inhibiting growth of Gardnerella vaginalis, Mobiluncus and other bacterial species cultured from vaginal content of women with bacterial vaginosis. Acta Pathol Microbiol Immunol Scand [B]1;:-. Downloaded from on 1 May 1