Pathogenesis of abnormal vaginal bacterial flora

Gilbert G.G. Donders, MD, PhD,a Eugene Bosmans, MD,b Alfons Dekeersmaecker,b
Annie Vereecken, Pharm,b Ben Van Bulck, MD,c and Bernard Spitz, MD, PhDa
Leuven and Antwerp, Belgium
OBJECTIVE: This study was undertaken to determine the relationships between microscopy findings on wet
mounts, such as lactobacillary grade or vaginal leukocytosis, and results of vaginal culture, lactate and succinate content of the vagina, and levels of selected cytokines.
STUDY DESIGN: In a population of 631 unselected women seeking treatment at an obstetrics and gynecology
outpatient clinic, vaginal fluid was obtained by wooden Ayre spatula for wet mounting and pH measurement, by
high vaginal swab for culture, and by standardized vaginal rinsing with 2 mL 0.9% sodium chloride solution for
measurements of lactate, succinate, interleukin 1, interleukin 8, leukemia inhibitory factor, and interleukin 1 receptor antagonist concentrations. Lactate and succinate levels were measured by gas-liquid chromatography
and the cytokine concentrations were measured by specific immunoassays. Both univariate analysis (Student t
test, Welch test, 2 test, and Fisher exact test) and multivariate regression analysis (Cox analysis) were used.
RESULTS: Increasing disturbance of the lactobacillary flora (lactobacillary grades I, IIa, IIb, and III) was highly
correlated with the presence of Gardnerella vaginalis, Trichomonas vaginalis, enterococci, group B streptococci,
and Escherichia coli. Vaginal pH and interleukin 8 and interleukin 1 concentrations increased linearly with increasing lactobacillary grade, whereas lactate concentrations and the presence of epithelial cell lysis decreased.
A similar pattern of associations with increasing leukocyte count was clear, but in addition there was an increase
in leukemia inhibitory factor concentration. Multivariate analysis of vaginal leukocytosis, lactobacillary grades,
and the presence of positive vaginal culture results showed that interleukin 1 concentration was most closely
related to the lactobacillary grade, leukemia inhibitory factor concentration was most closely related to the lactobacillary grade and positive culture results, interleukin 8 concentration was most closely related to positive culture results, and interleukin 1 receptor antagonist concentration was most closely related to vaginal leukocytosis
and positive culture results. The concentration ratio of interleukin 1 to interleukin 1 receptor antagonist remained stable, except when vaginal leukocytosis increased. In its most severe form, with >10 leukocytes per epithelial cell present, a decompensation of the vaginal flora with a collapse in interleukin 1 and interleukin 1 receptor antagonist concentrations was seen, but there was a concurrent sharp increase in leukemia inhibitory
factor concentration. This pattern was completely different from the course of the cytokine concentrations associated with a lactobacillary grade increase.
CONCLUSION: Both disturbed lactobacillary grade and the presence of increasing vaginal leukocytosis were
correlated with lactobacillary substrate (lactate) concentration, pH, and the concentrations of a variety of cytokines. There was a remarkably linear increase in these cytokines as either leukocytosis or lactobacillary grade
became more severe. In circumstances in which leukocytosis was extreme, however, interleukin 1 was no
longer produced but leukemia inhibitory factor concentrations increased. We speculate that in extreme inflammation the body tries to limit the damage that can be done by exaggerated cytokine production. (Am J Obstet
Gynecol 2000;182:872-8.)

Key words: Bacterial vaginosis, cytokines, interleukin, lactobacillary grades, lactobacilli, wet mount
microscopy

The absence of vaginal lactobacillary morphotypes is

associated with anaerobic bacterial vaginosis,1 with cerFrom the Department of Obstetrics and Gynecology, Gasthuisberg
Hospital, Katholieke Universiteit Leuvena and the Laboratory of
Clinical Pathology, Division of Microbiology,b and the Department of
Obstetrics and Gynecology, Erasmus Hospital.c
Received for publication March 3, 1999; revised November 19, 1999;
accepted December 16, 1999.
Reprint requests: Gilbert G.G. Donders, MD, PhD, University Hospital
Gasthuisberg, Herestraat 49, Leuven 3000, Belgium.
Copyright 2000 by Mosby, Inc.
0002-9378/2000 $12.00 + 0 6/1/105195
doi:10.1067/mob.2000.105195

872

tain sexually transmitted diseases,2 and with overgrowth

of facultative pathogenic commensal bacteria of intestinal origin.3 The presence of abnormal flora (bacterial
vaginosis) has been associated with elevated concentrations of selected bacteria, an elevated sialidase level, a
high pH, and elevated cytokine levels in the cervix and
vagina.4, 5 Abnormal vaginal flora and, in particular bacterial vaginosis, are related to preterm delivery, amniotic
fluid infection, and histologic chorioamnionitis.6-8
Treatment of bacterial vaginosis has reduced preterm
delivery among many subsets of women, but not in all
studies and not among all women, especially low-risk

Volume 182, Number 4

Am J Obstet Gynecol

women.4, 8-11 The mediators thought to be responsible

for the complications at issue are proinflammatory products, such as cytokines, that originate from leukocytes
and from several sites in the female genital tract.
Interleukin 1 (IL-1) is liberated first, followed by interleukins 6 (IL-6) and 8 (IL-8) and the prostaglandins. By
regulating prostaglandin release these cytokines play a
role not only in normal parturition but also in preterm
labor and preterm delivery.12 It remains obscure, however, why most women with abnormal flora are not delivered preterm.
To further examine the possible role of infection in
preterm delivery, we studied other characteristics of vaginal fluid related to vaginal flora. This study represents an
attempt to begin to understand the characteristics that
could potentially represent independent factors for
upper genital tract invasion and infection.
Microscopic evaluation of the vaginal flora is a cheap
and easy technique for assessing infection-related risks in
women.2-4, 13-15 We examined the relationships between
several microscopic characteristics of the vaginal flora
and the vaginal culture results, vaginal lactate concentration, and concentrations of the cytokines IL-8, IL-1,
leukemia inhibitory factor (LIF), and IL-1 receptor antagonist (IL-1RA).
Material and methods
Patient population. After oral consent was given, vaginal samples were obtained from 631 unselected women
who sought treatment at the obstetrics and gynecology
clinic at Gasthuisberg University Hospital, Leuven,
Belgium. Among these women, 263 were undergoing examination for a pregnancy follow-up visit and 368 were
examined for a contraceptive checkup or because of genital infection.
Menopausal women who were not receiving hormonal
replacement therapy, women with genital prolapse, and
women with overt genital bleeding were excluded, because these conditions would alter the results from vaginal rinsing fluid. Antibiotic therapy, recent sexual intercourse, and recent use of vaginal products were not
exclusion criteria, because they would form part of the
normal setting during screening.
Sampling and laboratory procedures. An unmoistened
sterile speculum was inserted before any other vaginal examination was performed. A sample of vaginal fluid was
taken from the lateral vaginal vault with a wooden Ayre
spatula and spread onto a glass slide. Sodium chloride solution was applied and covered with a glass slip for immediate microscopic evaluation. Vaginal pH was measured
by use of color strips with a pH range of 4.0 to 7.0 (Merck
& Co, Inc, Whitehouse Station, NJ). A standardized vaginal rinsing with 2 mL 0.9% sodium chloride solution was
performed by flushing and reaspirating the fluid through
a 0.5-mm-wide and 6-cm-long needle in the left, central,

Donders et al 873

and right upper areas of the vaginal vault for centrifugation and measurements of lactate, succinate, IL-1, IL-8,
LIF, and IL-1RA concentrations in the supernatant.
Lactate and succinate levels were measured by gas-liquid
chromatography, and the cytokine concentrations were
measured by specific immunoassays. All samples were collected and partially processed by one person (Gilbert
G.G. Donders, MD, PhD); they arrived at the laboratory
for further processing within 6 hours after sampling.
Samples were transported at room temperature and then
frozen at 18C for batch analysis at a later time.
Immune assays of cytokine levels. IL-1 concentration
was determined with a Cistron (catalog No. 03-HS96;
Cistron Biotechnology, Pine Brook, NJ) high-sensitivity enzyme-linked immunosorbent assay (ELISA) kit, a monoclonal antibodybased sequential enzyme immunoassay
that was specific for human IL-1 with a dynamic range of
2 to 1000 pg/mL. IL-8 concentration measurements were
performed with a Eurogenetics (catalog No. E01-18-1300;
Eurogenetics, Tessenderlo, Belgium) ELISA kit, a sandwich ELISA that is based on a monoclonal and polyclonal
antibody combination covering a dynamic range of 25 to
1500 pg/mL. LIF levels were determined with a Eurogenetics ELISA kit (catalog No. E04-18-1290), a monoclonal antibodybased two-step sandwich ELISA with a dynamic range of 25 to 1000 pg/mL in which 25 pg/mL is
the detection limit. The IL-1RA ELISA used for this study
was a Eurogenetics kit (catalog No. E01-18-1370) with a
calibration curve ranging from 50 pg/mL to 10 ng/mL.
For each of these immunologic methods the vaginal fluid
washings were diluted such that most of the resulting cytokine concentrations were located in the middle of the
calibration curves for each cytokine.
Grading of the lactobacillary morphotypes. The wet
mount grading system was modified14 from Schrders
original classification.15 The numbers of lactobacillary
morphotypes were estimated in comparison with the
numbers of other bacteria present, leading to 4 groups.
Normal flora (lactobacillary grade I) comprised a predominance of lactobacillary morphotypes, with few coccoid bacteria present. Care must be taken not to misdiagnose the cellular debris from lysed epithelial cells as
coccoid bacteria. Also, the excessive acidification by normal lactobacilli may cause lysis of epithelial cells, leaving
bare nuclei behind. These must not be confused with
leukocytes, which have approximately the same size.
Intermediate flora (lactobacillary grade II) comprised a
diminished lactobacillary flora that was mixed with other
bacteria. A mixed flora with lactobacilli still outnumbering the other bacteria was categorized as IIa; a more abnormal mixed flora with fewer lactobacilli than other
bacteria was labeled as IIb. Finally, abnormal flora (lactobacillary grade III) comprised numerous other bacteria
with no lactobacilli present.
This grading system was chosen because of its value in

874 Donders et al

April 2000
Am J Obstet Gynecol

Table I. Characteristics of vaginal fluid (pH, wet mount microscopy, and biochemical marker and cytokine concentrations) as a function of lactobacillary grade

pH (mean SD)
Lactate (mg/dL, geometric mean titer SD)
Succinate (mg/dL, geometric mean titer SD)
Succinate >1 mg/dL (No.)
Coccoid background (No.)
Lysis of epithelial cell (No.)
Parabasal cells (No.)
IL-8 (pg/mL, mean SD)
LIF (pg/mL, mean SD)
IL-1 (pg/mL, mean SD)
IL-1RA (pg/mL, mean SD)

Grade I
(n = 241)

Grade IIa
(n = 194)

Grade IIb
(n = 58)

Grade III
(n = 112)

Statistical
significance

4.5 0.5
126 2.5
1.2 2.9
123 (51%)
0 (0%)
81 (34%)
1 (0.4%)
2.2 5.2
53 162
127 367
91 81

4.8 0.7
110 2.6
1 3.6
86 (44%)
26 (13%)
36 (19%)
2 (1%)
4.5 9.1
39 71
243 512
84 85

5.6 0.8
56 2.4
0.95 3.5
23 (40%)
25 (45%)
1 (1.7%)
3 (5.2%)
6.9 11
42 56
821 1056
87 75

6.1 0.8
24 3.2
0.79 4.8
43 (38%)
60 (54%)
2 (1.8%)
9 (8.0%)
8.1 13
37 76
1033 1870
84 82

P < .0001
P < .0001
P = .0220
P < .0001
P < .0001
P < .0001
P < .0001

The t test was used for normally distributed analogous data, and the Welch test was used for data not normally distributed (unequal
variances). For discrete numbers the 2 test for trend was used or, when numbers were too low, lactobacillary grades I and IIa were compared with lactobacillary grades IIb and III by Fisher exact test.

clinical practice.2, 3 Although this grading system has

some similarities to the lactobacillary scoring in the classification for diagnosis of bacterial vaginosis of Nugent et
al,16 it differs from the latter in that our system is meant
to diagnose abnormality of the flora in the first place,
without attempting a specific diagnosis, whereas the
Nugent score is designed to discover anaerobic bacterial
vaginosis alone. A Nugent score of 7 is diagnostic for
bacterial vaginosis. However, intermediate scores of 5 or
6 are difficult to interpret, and it is not known whether
such scores indicate an intermediate stage between normal flora and bacterial vaginosis or rather represent a
separate entity. Consequently, this intermediate group
has been misused by being joined either with the bacterial vaginosis group or with the normal flora group,
whichever produced a better-looking result. Use of lactobacillary grades allows a progressive degree of abnormality. This permits high-risk groups to be detected by comparison of the number of lactobacillary morphotypes
with the number of other bacteria present, without any
requirement for a refined diagnosis on its etiology. If
treatment is envisaged, further diagnostic workup in
women with abnormal flora is a prerequisite.
Vaginal leukocytes were counted and semiquantitatively graded as follows: <10 leukocytes per high-power
field (400 magnification) and >10 leukocytes per highpower field combined with either <5 leukocytes per epithelial cell, 5 to 10 leukocytes per epithelial cell, or >10
leukocytes per epithelial cell.
Statistics. Both univariate analysis and multivariate regression analysis (Cox analysis) were used. For univariate
analysis the Student t test and the 2 test were replaced
with the Welch test and the Fisher exact test when there
were unequal variances between the groups or the numbers in both groups were rather small. P values are only
given when the difference between groups was significant
at a level of P < .05.

Results
Lactobacillary grades were normal (lactobacillary
grade I) in 256 of 631 cases (40%), slightly disturbed
(lactobacillary grade IIa) in 201 (32%), moderately disturbed (lactobacillary grade IIb) in 58 (9%), and severely
disturbed (lactobacillary grade III) in 116 (18%). Vaginal
leukocytosis was <5 leukocytes per high-power field in 21
of 593 (3.5%), >10 leukocytes per high-power field but <5
leukocytes per epithelial cell in 381 (64%), 5 to 10 leukocytes per epithelial cell in 141 (25%), and more than 10
leukocytes per epithelial cell in 44 (7.4%).
With increasing lactobacillary grade (decreasing lactobacilli), pH increased from a mean of 4.5 to a mean of
6.1 (P < .0001 for trend; Table I), whereas lactate concentration decreased from a geometric mean titer of 126
mg/dL to a geometric mean titer of 24 mg/dL (P < .0001
for trend). IL-8 and IL-1 concentrations increased 4- to
10-fold with increasing lactobacillary grade (P < .0001),
but LIF and IL-1RA levels remained stable.
With increasing vaginal leukocytosis the pH and lactate
concentration curves showed a bimodal pattern. The pH
was 5.4 when no leukocytes were present, decreased to
4.9 in the group with moderate vaginal leukocytosis, and
increased to 5.9 when leukocytosis was excessive (Table
II). Lactate levels followed a similar bimodal pattern,
only reciprocal. The lactate concentration was low (45
mg/dL) when leukocytosis was low, fell within the reference range (87-89 mg/dL) when leukocytosis was intermediate, and was again low (59 mg/dL) when leukocytosis was extreme.
All the cytokines studied, including IL-1RA, showed a
tendency toward an increase in concentration as the degree of vaginal leukocytosis increased. IL-8 and LIF concentrations showed a steady rise up to the group with severe leukocytosis of >10 leukocytes per epithelial cell.
IL-1 and IL-1RA concentrations increased to maximum
levels in the group with moderate leukocytosis of 5 to 10

Donders et al 875

Volume 182, Number 4

Am J Obstet Gynecol

Table II. Characteristics of vaginal fluid (pH, wet mount microscopy, and biochemical marker and cytokine concentrations) as a function of vaginal leukocytosis

pH (mean SD)
Lactate (mg/dL, geometric mean titer SD)
Succinate (mg/dL, geometric mean titer SD)
Succinate >1 mg/dL (No.)
Coccoid background (No.)
Lysis of epithelial cells (No.)
Parabasal cells (No.)
IL-8 (pg/mL, mean SD)
LIF (pg/mL, mean SD)
IL-1(pg/mL, mean SD)
IL-1RA (pg/mL, mean SD)

<5 leukocytes per

high-power field
(n = 21)

<5 leukocytes per

epithelial cell
(n = 381)

5.4 1.0
45 2.5
1.25 4.4
10 (48%)
2 (9.5%)
2 (9.5%)
0 (0%)
2.9 7.1
15 7.7
93 744
67 58

4.9 0.8
87 3.6
1.0 3.46
86 (45%)
47 (12%)
100 (26%)
3 (1%)
3.3 7.4
36 90
260 587
89 87

5-10 leukocytes per >10 leukocytes per

epithelial cell
epithelial cell
Statistical
(n = 147)
(n = 44)
significance
5.0 0.9
89 3.2
1.15 3.7
70 (48%)
35 (24%)
15 (12%)
6 (4.5%)
7.5 13
53 162
1844 2640
98 80

5.9 1.0
59 2.7
10 5.13
18 (41%)
23 (52%)
0 (0%)
6 (14%)
8.2 12
96 111
243 683
49 51

P < .0001
P < .0001
P < .0001
P < .0001
P < .0001
P < .0001
P = .0036
P < .0001
P = .0040

Statistical methods are the same as for Table I.

Table III. Age-adjusted multiple regression analysis on relationship of IL-1, LIF, IL-8, and IL-1RA concentrations with
lactobacillary grade, vaginal leukocytosis, and positive vaginal culture result (any pathogen except Candida species)
IL-1

Vaginal leukocytosis
Lactobacillary grade
Positive vaginal
culture result

R2

Statistical
significance

0.22
0.26
0.23

0.7
1.9
0.3

P = .50
P = .06
P = .80

LIF
R2

0.06 0.8
0.04 2000
0.09 2.8

leukocytes per epithelial cell and then sharply decreased

in the group with severe leukocytosis (Table III).
The ratio of IL-1 to IL-1RA showed a steady increase
as a function of the increasing lactobacillary grade, leading to a 5-fold increased ratio among the women with severely disturbed flora (Fig 1, A). Increasing vaginal leukocytosis also corresponded with a similar rise in the ratio
of IL-1 to IL-1RA, a sharp 10-fold increase in the ratio in
the group with moderate leukocytosis (5-10 leukocytes
per epithelial cell). When the leukocytosis increased further to >10 leukocytes per epithelial cell, however, the IL1/IL-1RA ratio dropped to baseline levels (Fig 1, B).
LIF concentration was not influenced by the bacterial
flora but increased gradually with the extent of leukocytosis present. In contrast to IL-1 and the IL-1/IL-1RA
ratio, LIF further increased as leukocytosis rose to >10
leukocytes per epithelial cell.
To exclude possible mutual interactions of lactobacillary grade, vaginal leukocytosis, and positive culture results, age-adjusted multiple regression analysis was applied
to study the individual correlations between these three
markers of vaginal infection and the individual cytokine
levels (Table III). After multivariate analysis IL-1 concentration correlated best with the lactobacillary grade (R2 =
0.26; P = .06) but did not correlate significantly with either
positive vaginal culture results or vaginal leukocytosis. LIF

IL-8

Statistical
significance
P = .500
P = .040
P = .007

R2

0.09 0.4
0.08 100
0.13 2.4

IL-1RA

Statistical
significance
P = .700
P = .300
P = .018

R2

0.4 2.4
0.4 1.10
0.5 3.9

Statistical
significance
P = .0200
P = .3000
P = .0001

concentration showed a slight but significant negative correlation with lactobacillary grade (LIF decreased with increasing disruption of the vaginal flora). There was also a
strong positive correlation between LIF and vaginal colonization with pathogens (R2 = 0.09; P = .007) but not between LIF and vaginal leukocytosis.
IL-8 levels correlated with positive vaginal culture results (R2 = 0.13; P = .02) but showed no independent relationship with the lactobacillary grade. 1L-1RA concentration was positively associated with both vaginal
leukocytosis (R2 = 0.4; P = .02) and vaginal colonization
(R2 = 0.5; P = .0001).
Comment
This study demonstrated a good correlation between
lactobacillary grades and pH, lactate content, and increased cytokine production in the vagina. Lactobacillary
grade III corresponds to a complete disruption of the
normal lactobacillary resistance of the vagina, with no
lactobacillary morphotyes present on microscopic examination of the vaginal fluid. Although it is tempting to
equate lactobacillary grade III with bacterial vaginosis,
the two conditions should be seen as overlapping but separate markers of vaginal infectious disease.14 It has been
shown before that lactobacillary grade III flora may harbor a large number of aerobic organisms, organisms

876 Donders et al

April 2000
Am J Obstet Gynecol

Fig 1. Vaginal IL-1/IL-1RA ratio (open triangles) and concentrations of IL-1 (open bars), LIF (striped bars), and IL-1RA
(dotted bars) as function of lactobacillary (Lb) grade (A) and vaginal leukocytes (B). hpf, High-power field.

causing sexually transmitted diseases, and anaerobic organisms.2, 3, 17, 18 Grossly abnormal flora may lead either
to anaerobic bacterial vaginosis or to aerobic vaginitis.
We use the term aerobic vaginitis for a condition in which
disrupted vaginal flora has been replaced by pathogenic
aerobic microorganisms, usually Escherichia coli, group B
streptococci, or enterococci of intestinal origin, which
elicits a severe local immune response in the host as indicated by increased vaginal leukocytosis (unpublished observations).14, 19
In this study we used the wet mount of fresh vaginal
fluid as the diagnostic test of choice for lactobacillary
grade. Although the Papanicolaou test,1 the Gram stain,2
and probably many other techniques may be used, wet
mounting is at least as useful as the Gram stain.19, 20 The
close correlation between lactobacillary grade and vaginal lactate content and pH further indicates that wet
mount is an acceptable method to use for the assessment
of lactobacillary flora.21
As vaginal leukocytosis increased, the pH changed in a
manner different from that of the lactate concentration.
Both very high and very low vaginal leukocytosis levels
were associated with low lactate levels and high pH.
A probable explanation could be that a large number of
women in the group with low vaginal leukocytosis harbored bacterial vaginosis. Bacterial vaginosis is typically
devoid of vaginal host immune response but is associated
with decreased lactobacillary morphotypes and thus with
low lactate and high pH.
Cytokines are released from macrophages and possibly
also from the genital tissues in response to microbial invasion, which explains the rises in IL-8, LIF, and IL-1RA
concentrations associated with colonization. In multivariate analysis, however, positive vaginal culture results were
not associated with IL-1. With decreasing lactobacilli
both IL-1 and IL-8 concentrations increased, whereas
LIF and IR-1RA concentration remained unchanged. It

therefore appears that the way in which the cytokines

react to microbial colonization is different from the way
in which they respond to the absence of vaginal lactobacillary resistance.
Increasing vaginal leukocytosis was found to be associated with yet another set of cytokines. With increasing
leukocyte levels all cytokines were produced in significant amounts, but this seems not to be confirmed by a
multivariate regression analysis.
Extreme care is warranted in the interpretation of this
logistic regression analysis. The relationships between
vaginal leukocytosis and most cytokine concentrations,
lactate concentration, and pH follow a bimodal pattern.
The concentration of IL-1 rose steadily as leukocytosis
increased up to 10 leukocytes per epithelial cell but declined sharply when leukocytosis rose further beyond 10
leukocytes per epithelial cell.
To a lesser degree the same pattern was also seen for
IL-1RA, which resulted in a rising ratio of IL-1 to IL-1RA
with increasing leukocytosis followed by a sudden decrease in the value when leukocytosis increased to >10
leukocytes per epithelial cell. In contrast, both IL-8 and
LIF concentrations continued to rise, even when vaginal
leukocytosis was extremely intense.
In most septic conditions IL-1 is liberated first, followed by IL-6 and IL-8. In the general circulation the latter two cytokines have profound effects on the human
immune system, with release of vasodilatory leukotrienes
and toxins leading to septic shock. IL-1RA is believed to
constitute a natural negative feedback mechanism by
competitive inhibition at the level of the IL-1 receptor.
The function of LIF is less well understood, but it may
exert negative action on the attraction and function of
leukocytes.
Fig 2 shows a diagram summarizing these findings. First,
when the lactobacillary flora deteriorates, as witnessed by
decreased lactate levels and increased pH, pathogenic mi-

Donders et al 877

Volume 182, Number 4

Am J Obstet Gynecol

Fig 2. Schematic presentation of sequence of events in case of disruption of vaginal lactobacillary microflora. LBG,
Lactobacillary grade; IL-1b, IL-1; I-RA, IL-1RA.

crobial flora is allowed to develop, liberating IL-1. IL-1

and the presence of microbial opsonins from the
pathogens stimulate macrophages to produce IL-8, which
is a vasodilator and a powerful attractant for leukocytes.
The presence of pathogenic microflora also stimulates
liberation of LIF, which had been suppressed as long as
the lactobacillary resistance was still present. In addition,
excessive IL-1 production and degranulation of leukocytes stimulate IL-1RA production, causing the IL-1/IL1RA ratio to decrease and the IL-8 production to fade.
Together with LIF liberation, this decrease in IL-8 concentration halts further attraction of leukocytes, so that
severe local destruction after massive degranulation of
leukocytes is prevented.
This scheme implies that the loss of lactobacillary resistance precedes the pathogenic bacterial overgrowth. For
many years there has been argument as to whether
Gardnerella vaginalis and the associated anaerobic bacteria proliferate because there is a lack of lactobacillary
resistance (eg, after menstrual blood loss) or whether facultative pathogenic bacteria simply overgrow the indigenous flora and overwhelm the vagina.
Levels of IL-1, the first cytokine released, increase
steadily with decreasing lactobacillary resistance, as measured by lactobacillary grade, lactate concentration, and
to a lesser extent pH. Once microbial colonization becomes the most prominent feature IL-8 concentration increases, but at the same time some negative feedback
mechanisms come into action. The production of IL-1
in large quantities even before vaginal microbial invasion
is prominent may well precede vaginal leukocytosis and
may therefore possibly be caused by cells other than
those from the immune system.
The critical factors in the vaginal flora that place a
woman at risk for preterm delivery, amniotic fluid infection, postpartum endometritis, and possibly pelvic inflammatory disease are still largely unknown. The first

step in elucidating them is to determine a subset of

women with and without abnormal flora who have characteristics of the vaginal discharge that differ from those
of others and that could potentially be used to identify a
group at particularly high risk for infection and possibly
to explain the pathogenesis of these infections.
REFERENCES

1. Spiegel CA, Amsel R, Holmes KK. Diagnosis of bacterial vaginosis by direct Gram stain of vaginal fluid. J Clin Microbiol
1983;18:170-7.
2. Donders GG, De Wet GH, Hooft P, Desmyter J. Lactobacilli in
Papanicolaou smears, genital infections and pregnancy. Am J
Perinatol 1993;10:358-61.
3. Donders GG. Bacterial vaginosis in pregnancy: screen and treat?
[editorial] Eur J Obstet Gynecol Reprod Biol 1999;83:1-4.
4. McGregor JA, French JI, Jones W, Milligan K, McKinney PJ,
Patterson E, et al. Bacterial vaginosis is associated with prematurity and vaginal fluid mucinase and sialidase: results of a controlled trial of topical clindamycin cream. Am J Obstet Gynecol
1994;170:1048-59.
5. Hay PE, Lamont RF, Taylor-Robinson, Morgan DJ, Ison C,
Pearson C. Abnormal bacterial colonization of the genital tract
and subsequent preterm delivery and late miscarriage. BMJ
1994;308:295-8.
6. Platz-Christensen JJ, Mattsby-Baltzer I, Thomsen P, Wiqvist N.
Endotoxin and interleukin-1 alpha in the cervical mucus and
vaginal fluid of pregnant women with bacterial vaginosis. Am J
Obstet Gynecol 1993;169:1161-6.
7. Martius J, Eschenbach DA. The role of bacterial vaginosis as a
cause of amniotic fluid infection, chorioamnionitis and prematurity: a review. Arch Obstet Gynecol 1990;247:1-13.
8. McDonald HM, OLoughlin JA, Jolley PT, Vigneswaran R,
McDonald PJ. Changes in vaginal flora during pregnancy and
association with preterm birth. J Infect Dis 1994;170:724-8.
9. Rosenstein IJ, Morgan DJ, Sheehan M, Lamont RF, TaylorRobinson D. Effect of topical clindamycin on bacterial vaginosis
during pregnancy [abstract]. Int J Gynaecol Obstet 1999;67:S43.
10. Carey JC, Klebanoff MA, Hauth JC, Hillier SL, Thom EA, Ernest
JM, et al. Metronidazole to prevent preterm delivery in pregnant women with asymptomatic bacterial vaginosis. National
Institute of Child Health and Human Development Network of
Maternal-Fetal Medicine Units. N Engl J Med 2000;342:534-40.
11. Hauth JC, Goldenberg RL, Andrews WW, DuBard MB, Copper
RL. Reduced incidence of preterm delivery with metronidazole
and erythromycin in women with bacterial vaginosis. N Engl J
Med 1995;333:1732-6.

878 Donders et al

April 2000
Am J Obstet Gynecol

12. Dudley DJ, Trautman MS. Infection, inflammation and contractions: the role of cytokines in the pathophysiology of preterm
labor. Semin Reprod Endocrinol 1994;12:263-72.
13. Larsson PG, Platz-Christensen JJ, Thejls H, Forsum U, Phlson
C. Incidence of pelvic inflammatory disease after first-trimester
legal abortion in women with bacterial vaginosis after treatment
with metronidazole: a double-blind, randomized study. Am J
Obstet Gynecol 1992;166:100-3.
14. Donders GG. Microscopy of bacterial flora on fresh vaginal
smears. Infect Dis Obstet Gynecol 1999;7:126-7.
15. Schrder K. Zr pathogenese und Klinik des vaginalen
Vaginalbiocoenose auf sechs grundbilder. Zentralbl Gynekol
1921;45:1350-61.
16. Nugent RP, Krohn MA, Hillier SA. Reliability of diagnosing bacterial vaginosis is improved by a standardized method of Gram
stain interpretation. J Clin Microbiol 1991;29:297-301.

17. Donders GG, 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 premature birth. Prenat Neonat Med 1998;3:588-93.
18. Donders GG, Moerman P, De Wet GH, Hooft P, Goubau P.
The association between Chlamydia cervicitis, chorioamnionitis and neonatal complications. Arch Gynecol Obstet 1991;
249:79-85.
19. Donders GG, Vereecken A, Salembier G, Van Bulck B, Spitz B.
Assessment of vaginal lactobacillary flora in wet mount and fresh
or delayed Gram stain. Infect Dis Obstet Gynecol 1996;4:2-6.
20. Donders GG, Vereecken A, Dekeersmaecker A, Van Bulck B,
Spitz B. Wet mount reflects functional vaginal lactobacillary
flora better than Gram stain. J Clin Pathol. 2000 In press.
21. Donders GG, Desmyter J, Vereecken A. Vaginitis [letter]. N Engl
J Med 1998;338:1548-9.

N THE MOVE?
Send us your new address at least six weeks ahead

Dont miss a single issue of the journal! To ensure prompt service when you change your address,
please photocopy and complete the form below.
Please send your change of address notification at least six weeks before your move to ensure continued service.
We regret we cannot guarantee replacement of issues missed due to late notification.
JOURNAL TITLE:
Fill in the title of the journal here.

OLD ADDRESS:

NEW ADDRESS:

Affix the address label from a recent issue of the journal here.

Clearly print your new address here.

Name
Address
City/State/ZIP

COPY AND MAIL THIS FORM TO:

Journal Subscription Services
Mosby
11830 Westline Industrial Dr
St. Louis, MO 63146-3318

OR FAX TO:
314-432-1158

OR PHONE:
1-800-453-4351
Outside the U.S., call
314-453-4351