## Abstract

BACKGROUND: Overgrowth of bacteria in the birth canal is associated with an increased risk of late miscarriage, preterm labour, post-partum endometritis and low birthweight. Conception rates in assisted reproduction treatments (ART) remain frustratingly low. We examined whether the nature of bacterial flora, found in the uterine cervical canal at embryo transfer, is associated with the rate of conception in ART. METHODS: We sampled for bacteriological culture the cervical canal of 204 patients who underwent embryo transfer. Of these, 139 (68%) were of fresh embryos, following recent vaginal oocyte retrieval and prophylactic antibiotic therapy, and 65 (32%) of frozen–thawed embryos, without any vaginal intervention in the preceding days. Bacteriological work-up included identification, colony count and antibiotic susceptibility profile. Conception was correlated with bacterial type and colony count. RESULTS: In 75 patients (36.8%) sterile cervical cultures or lactobacillus were recorded. Of these 75 patients, 23 (30.7%) conceived, whereas among the 129 in whom any pathogenic micro-organism was recovered only 21 (16.3%) conceived (P = 0.002). No difference in colonization was found between women who underwent frozen–thawed versus fresh embryo transfer (57 and 67% respectively). Any Gram-negative colonization was associated with no conception. All Gram-positive, and 90% of the Gram-negative bacteria, were sensitive to augmentin. CONCLUSIONS: Failure to conceive in ART is significantly associated with bacterial colonization of the uterine cervix.

## Introduction

Bacterial vaginosis is associated with an increased risk of late miscarriage, preterm labour, post-partum endometritis and low birthweight (Hay et al., 1994; McCoy et al., 1995; Carey et al., 2000). Several reports indicate a higher risk for the pregnancy if bacterial vaginosis is present during early stages of pregnancy (Kurki et al., 1992; Hay et al., 1994, McGregor et al., 1995). Assuming a putative role for bacterial overgrowth in all the above gestational pathologies, one is tempted to speculate an even earlier effect, i.e. in embryonic implantation. In this regard, it is of note that assisted reproductive technology (ART) has been used for almost two decades, but conception rates consistently run lower than expectations (Fanchin et al., 1998), especially in the light of the high costs involved. Several studies have addressed the possible association of bacterial colonization of the vagina and cervix with success in ART, but their conclusions seem to contradict each other. Whereas some studies (Egbase et al., 1996; Fanchin et al., 1998) that sampled the cervical canal during embryo transfer found an association between the type of flora and the chances for conception, another study (Ralph et al., 1999) that sampled the vagina prior to oocyte retrieval did not.

In this study we assessed whether the nature of bacterial flora, found in the uterine cervical canal at embryo transfer, affects the rate of conception in ART cycles.

## Materials and methods

From June 1 to October 31, 1999, 204 embryo transfers were performed on days Sunday through to Thursday, from which samples could be seeded and processed. Of these transfers, 139 were of fresh embryos and 65 of frozen–thawed embryos. The protocols used for controlled ovarian hyperstimulation were as previously described (Shalev et al., 1995; Ben-Shlomo et al., 1997). HCG (Pregnyl; Organon, Os, The Netherlands) 10 000 IU i.m. was administered when at least three follicles reached a mean diameter of 18 mm. Oocyte retrieval was performed transvaginally under ultrasonographic guidance (6.5 MHz probe, Elcsint 1000; Elscint, Haifa, Israel) 32–38 h later. For prophylaxis against infection, a single i.v. dose of cefazoline 1 g was administered prior to retrieval. Women allergic to penicillin (five women in this study) were given i.v. erythromycin 300 mg. Oocytes were either inseminated with partners' spermatozoa or were subjected to ICSI, as applicable.

The protocols for embryo transfer of frozen–thawed embryos were two: (i) sonographic and biochemical (serum LH and progesterone) detection of spontaneous ovulation; (ii) endometrial estrogenic build-up, commencing at day 2 of menstruation and consisting of 4–8 mg/day oral micronized estradiol until endometrial thickness reached 9 mm, when 50 mg/day progeterone in oil was added. The diagnosis of conception was made on the basis of positive serum HCG testing and demonstration of a gestational sac in the uterine cavity. Patients were monitored for signs and symptoms of genital infection during the study and up to 2 weeks after embryo transfer.

When the patient was prepared for the embryo transfer procedure, the vaginal portion of the cervix was cleaned with dry gauze, a sham catheter was introduced into the cervical canal and the distal 10 mm piece was cut into a tube containing 1 ml of sterile saline solution for sampling. Embryo transfer was performed immediately after this. Within 1 h after sampling, the contents of the tube with the 10 mm catheter piece were shaken vigorously. Ten μl aliquots of this suspension were seeded onto the following plates: two plates of blood agar (Geloise du Sang, Sanofi Pasteur, Marnes-La-Coquetter, France, supplemented with 5% human blood), one plate of MacConkey agar (Oxoid, Basingstoke, UK), chocolate agar (Eugonagar; Becton and Dickinson, USA, supplemented with chocolatized human blood, haemin and nicotinamide adenine dinucleotide), potato dextrose agar (Difco, Detroit, USA), Centre for Disease Control (CDC) anaerobic agar (Geloise du Sang) and modified SP4 agar for mycoplasma/ureaplasma (Clarke, 1992). C.trachomatis was not sought, because the prevalence of positive cultures in asymptomatic women in Israel is very low (Ghinsberg and Nitzan, 1994). The rest of the fluid was centrifuged and the bottom 50 μl was seeded on the same plates as above. All plates were incubated for 48 h aerobically at 37°C, but the CDC anaerobic agar plates and one of the plain blood agar were incubated in an anaerobic jar for 72 h and then examined.

All bacterial growth was quantified as follows: every single colony growing in the 10 μl inoculated plates represented 100 colony forming units (CFU)/cm in the original 1 cm tip. Any bacteria growing in the centrifuged inoculation, but not in the first inoculation, were included in the category <100 CFU/cm'. Samples showing more than three different organisms in similar counts were defined as mixed culture'.

Bacterial isolates were identified by Gram stain, conventional biochemical methods, latex agglutination and automatic identification by Microscan system (Dade, Behrung, West Sacramento, USA). The adequate routine method was performed in each case.

Antibiotic susceptibility testing for aerobic bacteria was performed by disk diffusion test (Kirby and Bauer method) according to the standards of the National Committee for Clinical Laboratory Standards (NCCLS) (M2-A6, 1999). Antibiotic susceptibility testing for anaerobic bacteria was performed by E-test method (AB-Biodisk, Sweden) as recommended by NCCLS guidelines (M2-A6, 1999).

### Statistical analysis

Statistical analysis was conducted with the t-test, Mann–Whitney U-test, χ2 and regression analysis as applicable. No power calculation was done because the study was observational.

## Results

Overall, in 75 patients either a sterile cervical culture was recorded or only lactobacilus was recovered. The rate of potentially pathogenic colonisation (any bacteria other than lactobacillus) in women who underwent embryo transfer of frozen–thawed embryos was 57% compared with 67% in cycles of fresh embryo transfer (not significantly different). None of the patients reported clinical symptoms related to overt genital infection, neither was such an infection suggested on examination before embryo transfer. There were 36 clinical pregnancies in the 139 cycles of fresh embryo transfer (25.9%) and nine (13.8%) in the 65 frozen–thawed embryo transfer cycles. Of the 75 patients with either sterile or lactobacilli-positive cultures, 23 (30.7%) conceived, whereas among the 129 in whom any other micro-organism was recovered only 21 (16.3%) conceived. This difference was statistically significant (P = 0.002). As shown in Table I, these two groups of patients did not differ in background characteristics. Table IIreveals that the two groups did not differ in parameters of response to treatment. A further breakdown by fresh versus frozen–thawed embryo transfer revealed that fresh embryo transfer in women who had a sterile culture or lactobacillus resulted in a conception rate of 35.5%, whereas fresh embryo transfer in women who had positive cultures for other organisms resulted in a conception rate of 20%. This difference was statistically significant (P = 0.045).

After calculation of the number of colony forming units (CFU), which were counted after the seeding as described above, a preliminary association with conception was found. Since the presence of any Gram-negative colonisation was associated with no conceptions at all, these cases were deleted for the purpose of further calculation. No statistically significant correlation was found between the number of CFUs and conception rate. However, out of 41 patients with >10 000 CFUs of any bacterial type other than lactobacillus, only four (9.7%) conceived, whereas of 89 who had <10 000 colonies, 17 (19.1%) conceived.

The specific susceptibility to antibiotics among the group of pathogens, as observed in the 129 positive cases, reveals that all were resistant to cefazoline. All Gram-positive and 90% of the Gram-negative bacteria were sensitive to amoxicillin/clavulanic acid compound.

## Discussion

A host of elements during IVF treatment predict poorer success, including age, duration and type of infertility, the number of previous attempts at IVF, quality of the embryos, the number of oocytes and the number of resulting embryos available for transfer (Isenberg and D'Amato, 1995; Templeton et al., 1996). In this prospective study we found that the nature of cervical colonization is an independent and significant factor in the determination of success in ART.

Bacterial colonization of the uterine cervix has been suspected to influence conception rate. Possible causes of this could be an association between a cervical-positive culture and a concomitant, pre-existing uterine infection (Templeton and Morris, 1998); or direct inoculation of the endometrium or the embryo as a result of passage through the colonized cervix (Czernobilsky, 1978; Paulson et al., 1990; Tabibzadeh and Babaknia, 1995). The settlement of this issue is beyond the scope of clinical studies. The above notwithstanding, the incomplete and differing results of the existing studies do not provide enough ground for concrete recommendations on the preferred mode of intervention in this regard, which would be expected to improve conception rates (Fanchin et al., 1998). In order to collect data which will enable a firm conclusion and recommendation, we examined not only the type of bacteria present in the uterine cervix, but also quantified it and performed detailed antibiotic susceptibility tests. Although there seems to be a relationship between the colony count and chances for conception, there could not be a defined cut-off count, beyond which it would be preferable to cancel embryo transfer. We also added a `natural' control group in the form of cycles in which frozen–thawed embryos were transferred. This in turn allowed evaluation of the effect of our oocyte retrieval and the routine antibiotic preventive treatment on cervical findings. Interestingly, the colonization in women both after recent oocyte retrieval and in those scheduled for embryo transfer of frozen–thawed embryos was not susceptible to cefazoline, the drug that we used for prophylaxis during oocyte retrieval. This may explain the absence of difference in colonisation between the fresh and the frozen cycle patients.

Whereas our study results delineate a significant association between cervical colonization and the chance for conception, they suggest but do not provide statistically significant evidence for an association with either a bacterial type or the number of CFUs. Thus, our findings of E.coli in 8.5% of the positive cultures do not agree with another published study (Fanchin et al., 1998) that described a significant predominance of E.coli (68%) among the positive cultures, implying an association with conception rates. It has been suggested (Egbase et al., 1996) that there is a correlation between the number of colonies and the type of bacteria and the chance for conception. In this regard, it should be pointed out, that unlike ours, both these studies examined only aerobic bacterial colonization. An additional aspect of the results of these two studies and this one is the differing bacterial colonization profiles among different patient populations. This may indicate different future interventions for these respective populations.

Regarding the true nature of the association that we found, it is unclear whether there is a causative role for bacterial colonization in decreasing conception, or whether these two phenomena are merely the result of an obscure common factor. Nevertheless, our findings on the profile of antibiotic susceptibility pave the way for an interventional study, which will examine this issue in our patient population.

Table I.

Background characteristics in two groups of women from whom cervical cultures were obtained before embryo transfer

Normal Colonized Significance
(n = 75) (n = 129)
NS = not significant.
Age in years (SD) 31 (6.1) 32 (5.5) NS
Gravidity (SD)  1.01 (1.61)  1.05 (1.24) NS
Parity (SD)  0.39 (0.63)  0.40 (0.58) NS
Infertility factor
Male (%) 43 (57.3) 61 (47.2) NS
Female (%) 11 (14.6) 25 (19.3) NS
Anovulation (%)  7 (9.3) 11 (8.5) NS
Tubal (%)  3 (4) 13 (10) NS
Endometriosis (%)  1 (1.3)  1 (0.8) NS
Mixed (%) 16 (21.3) 28 (21.7) NS
Unexplained (%)  5 (6.6) 15 (11.6) NS
Infertility duration (years) (SD)  5.96 (4.26)  6.61 (4.39) NS
Early follicular FSH IU/l (SD)  6.76 (5.24)  6.50 (4.96) NS
Previous treatment cycles (SD)  1.85 (1.0)  1.91 (1.0) NS
Normal Colonized Significance
(n = 75) (n = 129)
NS = not significant.
Age in years (SD) 31 (6.1) 32 (5.5) NS
Gravidity (SD)  1.01 (1.61)  1.05 (1.24) NS
Parity (SD)  0.39 (0.63)  0.40 (0.58) NS
Infertility factor
Male (%) 43 (57.3) 61 (47.2) NS
Female (%) 11 (14.6) 25 (19.3) NS
Anovulation (%)  7 (9.3) 11 (8.5) NS
Tubal (%)  3 (4) 13 (10) NS
Endometriosis (%)  1 (1.3)  1 (0.8) NS
Mixed (%) 16 (21.3) 28 (21.7) NS
Unexplained (%)  5 (6.6) 15 (11.6) NS
Infertility duration (years) (SD)  5.96 (4.26)  6.61 (4.39) NS
Early follicular FSH IU/l (SD)  6.76 (5.24)  6.50 (4.96) NS
Previous treatment cycles (SD)  1.85 (1.0)  1.91 (1.0) NS
Table II.

Response to treatment in two groups of women from whom cervical cultures were obtained before embryo transfer

Normal Colonized Significance
(n = 75) (n = 129)
NS = not statistically significant; PR = pregnancy rate.
Gonadotrophin ampoules (SD) 36.9 (13.8) 39.2 (17.3) NS
Days of stimulation (SD)  9.8 (2.0) 10.1 (2.3) NS
Oocytes retrieved (SD) 15.1 (10.1) 13.7 (8.4) NS
Fertilized (SD)  8.4 (5.9)  7.7 (5.8) NS
Number of embryos transferred (SD)  2.6 (1.2)  2.7 (1.3) NS
Hours to embryo transfer (SD) 61.1 (12.0) 63.0 (11.7) NS
Cells/embryo (SD)  4.9 (1.7)  5.1 (1.6) NS
Implantation rate (%) 34/195 (17.3) 28/348 (8.0) P = 0.001
PR/embryo transfer (%) 23/75 (30.7) 21/129 (16.3) P = 0.002
Normal Colonized Significance
(n = 75) (n = 129)
NS = not statistically significant; PR = pregnancy rate.
Gonadotrophin ampoules (SD) 36.9 (13.8) 39.2 (17.3) NS
Days of stimulation (SD)  9.8 (2.0) 10.1 (2.3) NS
Oocytes retrieved (SD) 15.1 (10.1) 13.7 (8.4) NS
Fertilized (SD)  8.4 (5.9)  7.7 (5.8) NS
Number of embryos transferred (SD)  2.6 (1.2)  2.7 (1.3) NS
Hours to embryo transfer (SD) 61.1 (12.0) 63.0 (11.7) NS
Cells/embryo (SD)  4.9 (1.7)  5.1 (1.6) NS
Implantation rate (%) 34/195 (17.3) 28/348 (8.0) P = 0.001
PR/embryo transfer (%) 23/75 (30.7) 21/129 (16.3) P = 0.002
Table III.

The distribution of micro-organism types in cervical cultures taken from women before embryo transfer during assisted reproduction treatment

Microorganism Total positive Conceptions per embryo transfer Rate (%)
aTwelve cases of E.coli
Gram-negative 19a
Anaerobic pathogens 38 10.5
Gram-positive 46 10 21.7
Candida species 18 22.2
Mixed 21 14.0
Microorganism Total positive Conceptions per embryo transfer Rate (%)
aTwelve cases of E.coli
Gram-negative 19a
Anaerobic pathogens 38 10.5
Gram-positive 46 10 21.7
Candida species 18 22.2
Mixed 21 14.0
4
To whom correspondence should be addressed at: Department of Obstetrics and Gynecology, Ha'Emek Medical Centre, Afula 18101, Israel. E-mail: ibs@clalit.org.il
Submitted on November 24, 2000; resubmitted on May 4, 2001

## References

Ben-Shlomo, I., Wiener-Megnagi, Z., Golan, J., Eyali, V. et al. (
1997
) Individual implantation rate: proposal for a new index for evaluation of assisted reproduction results.
Fertil. Steril.
,
68
,
816
–819.
Carey, J.C., Klebanoff, M.A., Hauth, J.C., Hillier, S.L. et al. (
2000
) Metronidazole to prevent preterm delivery in pregnant women with asymptomatic bacterial vaginosis.
N. Engl. J. Med.
,
342
,
534
–540.
Clarke, L.M. (1992) Isolation of Genital Mycoplasmas. In Isenberg, H.D. (ed) Clinical Microbiology Procedures Handbook, Vol. 2. ASM, Washington D.C., USA, pp. 824–832.
Czernobilsky, B. (
1978
) Endometritis and infertility.
Fertil. Steril.
,
30
,
119
–120.
Egbase, P.E., al-Sharhan, M., al-Othman, S., al-Mutawa, M. et al. (
1996
) Incidence of microbial growth from the tip of the embryo transfer catheter after embryo transfer in relation to clinical pregnancy rate following in-vitro fertilization and embryo transfer.
Hum. Reprod.
,
11
,
1687
–1689.
Fanchin, R., Harmas, A., Benaoudia, F., Lundkvist, U. et al. (
1998
) Microbial flora of the cervix assessed at the time of embryo transfer adversely affects in vitro fertilization outcome.
Fertil. Steril.
,
70
,
866
–870.
Ghinsberg, R. C. and Nitzan, Y. (
1994
) Chlamydia trachomatis direct isolation, antibody prevalence and clinical symptoms in women attending outpatient clinics.
New Microbiol.
,
17
,
231
–242.
Hay, P.E., Lamont, R.F., Taylor-Robinson, D., Morgan, D.J. et al. (
1994
) Abnormal bacterial colonisation of the genital tract and subsequent preterm delivery and late miscarriage.
Br. Med. J.
,
308
,
295
–298.
Isenberg, H.D. and D'Amato, R.F. (1995) Indigenous and pathogenic microorganisms of humans. In Murray, P.R. et al. (eds) Manual of Clinical Microbiology. ASM, Washington D.C., USA, p.12.
Kurki, T., Sivonen, A., Renkonen, O., Savia, E. et al. (
1992
) Bacterial vaginosis in early pregnancy and pregnancy outcome.
Obstet. Gynecol.
,
80
,
173
–177.
McCoy, M.C., Katz, V.L., Kuller, J.A. and Killam, A.P. (
1995
) Bacterial vaginosis in pregnancy: an approach for the 1990s.
Obstet. Gynecol. Surv.
,
50
,
482
–488.
McGregor, J.A., French, J.I., Parker, R., Draper, D. et al. (
1995
) Prevention of premature birth by screening and treatment for common genital tract infections: result of a prospective controlled evaluation.
Am. J. Obstet. Gynecol.
,
173
,
157
–167.
MS-A6 (1999) Published by National Committee for Clinical Laboratory Standards. Wayne, PA 19087, USA.
Paulson, R.J., Sauer, M.V. and Lobo, R.A. (
1990
) Factors affecting embryo implantation after human in-vitro fertilization: a hypothesis.
Am. J. Obstet. Gynecol.
,
163
,
2020
–2023.
Ralph, S.G, Rutherford, A.J. and Wilson, J.D. (
1999
) Influence of bacterial vaginosis on conception and miscarriage in the first trimester: cohort study.
Br. Med. J.
,
319
,
220
–223.
Shalev, E., Giladi, Y., Matilsky, M. and Ben-Ami, M. (
1995
) Decreased incidence of severe ovarian hyperstimulation syndrome in high risk IVF patients receiving intravenous albumin: a prospective randomized study.
Hum. Reprod.
,
6
,
1373
–1376.
1995
) The signals and molecular pathways involved in implantation, a symbiotic interaction between blastocyst and endometrium involving adhesion and tissue invasion.
Hum. Reprod.
,
10
,
1579
–1562.
Templeton, A. and Morris, J.K. (
1998
) Reducing the risk of multiple births by transfer of two embryos after in-vitro fertilization.
N. Engl. J. Med.
,
339
,
573
–577.
Templeton, A., Morris, J.K. and Parslow, W. (
1996
) Factors that affect outcome of in-vitro fertilization treatment.
Lancet
,
348
,
1402
–1406.