Abstract
Purpose
The uterine microbiome of in vitro fertilization (IVF) patients was analyzed using next-generation sequencing (NGS) targeting 16S rRNA. Lactobacillus spp. were examined, with a special focus on Lactobacillus iners. The effects of antibiotic therapy on pregnancy outcomes were investigated.
Methods
A total of 257 IVF patients underwent endometrial microbiome testing. Patients were initially classified based on the percentage of Lactobacillus spp. into the Lactobacillus-dominant microbiome (LDM) group and the non-LDM group using a cutoff of 90%. Treatment was provided to non-LDM patients. Treated patients who improved on the second test were included in the Post-treatment group, and their pretreatment status was also examined.
Results
Lactobacillus was dominant in many IVF patients, but some patients showed Gardnerella or other bacteria associated with bacterial vaginosis. The treatment improvement rate for the non-LDM group was 81.4%, with an equivalent or better pregnancy success rate compared with the LDM group. The effect on pregnancy outcomes of Lactobacillus may differ by species, with L. crispatus and L. gasseri tending to act positively, whereas L. iners at ≥ 74.2% acts negatively.
Conclusions
This study shows that the recovery of an LDM in non-LDM IVF patients improves the composition of the endometrial microbiome, and pregnancy outcomes approach those of patients initially having an LDM. Furthermore, in LDM cases, L. iners species were also associated with lower pregnancy rates. These findings suggest that both the presence and type of Lactobacillus species are important for IVF success and that targeted microbiome treatment may improve reproductive outcomes.
Supplementary Information
The online version contains supplementary material available at 10.1007/s10815-025-03759-0.
Keywords: Assisted reproductive techniques, Endometrial microbiome, Lactobacillus iners, Lactobacillus-dominant microbiome (LDM), 16S ribosomal RNA
Introduction
Factors related to implantation success are broadly divided into those on the embryonic side and those on the maternal side. One could say that issues related to fetal factors are being resolved with selection of good embryos in association with morphological assessment and aneuploidy detection with preimplantation genetic testing for aneuploidy (PGT-A) [1]. On the maternal side, there is a wide range of factors that cannot be captured with the naked eye, including implantation failure associated with organic causes, the composition ratio of the endometrial microbiome, and abnormal immune responses via T-cells. Chronic endometritis may be one of the factors in implantation failure on the maternal side; it is a disease in which persistent inflammation of the endometrium caused by bacterial infection is seen. It is said to be present in 24–56% of women with recurrent miscarriages of two or more times [2–4]. Therefore, it has been pointed out that an association exists between chronic endometritis and implantation failure or recurrent pregnancy loss, and treatment for chronic endometritis is reported to lead to improved pregnancy outcomes [5–7]. Since universal guidelines for chronic endometritis are lacking, diagnostic criteria are uncertain. However, pathological examination of the endometrium and hysteroscopy are considered to be effective. When chronic endometritis is diagnosed, prescription of the wide-spectrum antibiotic doxycycline or other drugs is the standard protocol. At the same time, the opinion has also been expressed in recent years that drugs should be prescribed according to evidence with proper diagnosis and treatment to suppress the emergence of multidrug-resistant bacteria [8]. Therefore, accurate understanding of the intrauterine environment may be considered important to provide the best treatment and to improve pregnancy outcomes based on that treatment.
In the field of obstetrics and gynecology, the uterine cavity had long been thought to be sterile, but with advances in genetic analysis technology in recent years, the existence of a resident uterine microbiome has come to be recognized [9]. A 2015 Rutgers University study in the United States using next-generation sequencing (NGS) demonstrated bacteria in the uterus for the first time [10]. This discovery overturned the conventional wisdom that the “uterine cavity is sterile” and reminded us of the importance of accurately assessing the uterine environment. That paper demonstrated the co-existence of good bacteria and bad bacteria in the uterus, and it provided a new perspective that those bacteria affect pregnancy and reproduction. The following year, a Stanford University study showed that the proportion of Lactobacillus spp. in the uterine cavity directly affects the pregnancy rate. In a report by Moreno et al. involving infertility treatment patients who were undergoing in vitro fertilization (IVF), a comparison of the clinical results according to the percentage of Lactobacillus spp. present determined from the results of endometrial microbiome testing showed that the pregnancy rate in the normal group (Lactobacillus composition 90% or more) vs. the low Lactobacillus group was 70.6% vs. 33.3%, respectively, and the live birth rate was 58.8% vs. 6.7%, respectively [11]. In that study, women with Lactobacillus of at least 90% were confirmed to have a significantly higher pregnancy rate and live birth rate with IVF, demonstrating that the health of the endometrial microbiome contributes to successful pregnancy. From this report, the concept was born of a Lactobacillus-dominant microbiome (LDM) (> 90% Lactobacillus spp.) and a non-Lactobacillus-dominant microbiome (non-LDM) (< 90% Lactobacillus spp.). Lactobacillus spp. are the dominant bacteria in the female reproductive organs and play an important role in maintaining an acidic environment in the vaginal and uterine environments. The lactic acid generated by Lactobacilli maintains the pH in the uterus at 4 to 5 and suppresses the proliferation of bad bacteria, and it is thought to help maintain a uterine environment that is good for implantation and embryo growth. Successful embryo implantation may be hindered by an inflammatory response in the endometrium, so inflammatory mediators are strictly regulated when the blastocyst adheres to the epithelial endometrial wall, but non-LDM causes an inflammatory response in the endometrium [12, 13]. As described above, an interesting study regarding the communication between the endometrial microbiome and the endometrial epithelium has started. In addition, microbiome analysis of bacteria that are difficult to cultivate is also possible in endometrial microbiome testing using NGS. This provides a more detailed bacterial composition in the uterus and enables treatment plans to be decided according to their attributes. Measures for drug resistance to antibiotics are an important international issue for public health, and selection of appropriate antibiotics is important. In fact, 1.2 million people die annually from drug resistance [14], and in Japan, more than 5000 people are reported to die annually from methicillin-resistant Staphylococcus aureus (MRSA) and other types of bacteremia [15]. Given this background, wide-spectrum antibiotics should not be used lightly as a general protocol when antibiotics are administered, and it is thought that the effects of resistant bacteria can be minimized by providing treatment using appropriate dosages based on a proper understanding of microbial composition.
When a woman is diagnosed with chronic endometritis, prescription of the wide-spectrum antibiotic doxycycline or other drugs is a standard protocol. At the same time, as mentioned above, it has been recognized in recent years that antibiotics should be prescribed based on appropriate diagnosis and treatment in accordance with evidence to suppress the emergence of multidrug-resistant bacteria. In addition, Moreno et al. reported that the microbiomes in the vagina and uterus are similar but different [11]. The vaginal and endometrial microbiomes are also reported to differ in terms of diversity [16]. These findings highlight the importance of endometrial testing to improve pregnancy outcomes in patients with different vaginal and endometrial microbiomes, because the bacterial structure and composition of the vagina do not accurately reflect the bacteria that colonize the endometrium, where embryo implantation occurs, in all women.
As mentioned above, endometrial microbiome testing using NGS provides a more detailed bacterial composition in the uterus and enables treatment plans to be decided according to those attributes. In attempting to normalize the uterine microbiome, selection of various types of antibiotics and the administration of probiotics have been reported [17, 18], but they remain limited, and standard methods have not been established. It is also unclear how much they contribute to pregnancy. In the present study, the uterine microbiome was analyzed using NGS targeting 16S rRNA. In addition, the treatment effect and pregnancy outcomes with antibiotics were investigated in terms of these results, and Lactobacillus spp. up to the species level were examined, with a special focus on L. iners.
Materials and methods
Participants and study duration
This study was conducted from October 2021 to September 2023 in accordance with the guidelines of the Declaration of Helsinki. The participants were 257 women who underwent endometrial microbiome testing from among IVF cycle patients who gave written, informed consent after being advised of the significance and effectiveness of this testing. The study was approved by the ethics committee of the authors’ institution, and the trial was conducted as a prospective study (ethics committee approval number: osk001-R06). After testing, the outcomes were investigated in patients who underwent frozen-thawed embryo transfer. The inclusion criteria were Japanese women 20–45 years of age and IVF cycle patients who underwent oocyte retrieval and were planning a frozen embryo transfer. Cases of embryos brought in from other facilities and cases in which the embryos taken during the treatment of the older child were held were also included. The exclusion criteria were as follows: malignant disease or suspected malignant disease in the uterus; Asherman’s Syndrome; and a catheter could not be inserted into the uterus. There were no other specific detailed criteria, and the study included all patients undergoing IVF cycles who gave consent within the set period.
The primary endpoint was to investigate the status of the endometrial microbiome in IVF patients and pregnancy outcomes when the endometrial microbiome was non-LDM. The secondary endpoint was the type of treatment and its effects (changes in microbiome diversity and the effect of antibiotics) and differences in the effect depending on the species of Lactobacillus (especially L. iners) on pregnancy.
Treatment
A sensitivity protocol was performed first, followed by a metronidazole protocol. The metronidazole administration protocol was as follows: oral metronidazole 750 mg/day for 7 days and vaginal metronidazole 250 mg/day for 7 days. The sensitivity protocol was as follows: antibiotics to which the bacteria are known to be susceptible for 7 days.
Specimen collection, DNA extraction, and 16S rDNA sequence analysis with NGS
Specimens were collected from 10 days after the start of menstruation, with the timing selected so that collection would not be affected by menstrual blood. For the collection of uterine fluid, a MedGyn pipette (MedGyn Products, Westmont, IL, USA) was slowly inserted directly into the uterine cavity from the cervical opening, and fluid was aspirated by withdrawing the inner cannula. To prevent contamination, the vagina was thoroughly washed with benzalkonium chloride before inserting the catheter, and the first portion of the endometrial fluid obtained was discarded. The collected specimens were stored in OMNIgene-VAGINAL Tubes (DNA Genotek, Ottawa, Canada) placed in preservation solution and transported to Varinos Inc. at room temperature.
DNA extraction and sequence analysis of bacteria were conducted following the protocol of Varinos Inc. Specimens collected from patients were pretreated with MagNA Pure Bacteria Lysis Buffer (Roche, Basel, Switzerland), Proteinase K (Roche), and Lysozyme (Merck SA, Darmstadt, Germany) to solubilize tissue and bacteria. The Pathogen 1000 hp 3.2 protocol of MagNA Pure 24 was used in the DNA extraction work. For the extracted genomic DNA, the polymerase chain reaction (PCR) targeting the V1-V2 region of 16S rRNA genes was done. A sequence needed to determine the sequence with the Illumina sequencer, called an overhang sequence, was added to the 5′ terminal of the primer used during PCR. After PCR, samples were purified with KAPA HyperPure Beads (Roche), and impurities other than PCR products were eliminated. PCR was then done a second time to add an index. The index sequence conformed to the Nextera XT Index Kit v2 (Illumina) sequence. The KAPA HiFi Polymerase (Roche) was used in the second PCR response. Purification was done again with AMPure XP beads and taken as the library provided for the sequence. The concentration was measured with Qubit (Thermo Fisher Scientific, Waltham, MA, USA). Libraries from each specimen were mixed in equimolarly, and pair end sequencing using MiSeq (Illumina, San Diego, CA, USA) of 251 bp was performed. ZymoBIOMICS Microbial Community Standard (Zymo Research, Irvine, CA, USA) was used as a positive control, and UltraPureDNase/RNase-Free Distilled Water (Thermo Fisher Scientific) was used as a negative control.
Using Trimmomatic-0.38 [19] for Paired-End Reads output from the sequencer, adapter sequence and short leads were eliminated, and sequences were joined with EA-Utils fastq-join [20]. Using prinseq-lite-0.20.4 [21] for the sequence after joining, the primer binding region and low-quality leads (Q score < 25, lead length < 250 bp or ≥ 400 bp) were removed, and an operational taxonomic unit (OTU) was prepared (de novo OTU picking using uclust, sequence similarity threshold = 99.5%) with pick_otus.py of QIIME 1.9.1 [22]. A homology search was conducted with BLAST for representative sequences of each OTU, and identification was done at the species level (database used: prepared based on SILVA138 and STIRRUPS) [23, 24]. From the results of the BLAST search, sequences with homology of < 99% with the database and sequences with alignment regions of < 95% overall were excluded from the analysis due to the possibility of chimeras or other issues. Bacterial species thought to be derived from the reagents or work environment were also excluded from the analysis.
Statistical analysis
All analyses were performed with R software version 4.3.2 (R Foundation for Statistical Computing, Vienna, Austria). To perform a beta diversity analysis for the Good group and the Pre-improved and Improved groups, principal coordinate analysis was performed using the weighted Unifrac distance from the results of the BLAST homology search for each OTU. A PERMANOVA test was conducted to determine whether there were differences in the above three bacterial microbiome pattern groups. To perform an alpha diversity analysis for the above three groups, the Shannon Index, Chao1 Richness, and Phylogenetic diversity were calculated from the BLAST homology search results for each OTU, and after performing a rarefaction analysis, the diversities of the three groups were compared. When limited to the Improved group in patients under 37 years of age, a receiver-operating characteristic (ROC) curve analysis was done, and the area under the curve (AUC) was calculated to investigate the diagnostic accuracy of pregnancy success according to the percentage of L. iners. Student’s t-test was performed when normality and homoscedasticity were obtained for quantitative data, and the Wilcoxon rank-sum test was performed when normality was not obtained. Fisher’s exact test was performed for count data. The Benjamini & Hochberg method was used as a multiple comparison correction. Significance was taken to be p < 0.05.
Results
A flow chart of the cases included in the analysis in this study is shown in Fig. 1.
Fig. 1.
Flow chart of the patients analyzed in this study. (1) The state of the endometrial microbiome in in vitro fertilization patients is analyzed using the results of the first test. (2)−1 Treatment improvement effect (from the perspective of bacterial diversity). Good group (n = 157) vs. Pre-treatment (n = 52) vs. Post-treatment (n = 52) of Improved Group I. (2)−2 More effective treatment protocols (antibiotics). Improved group I (n = 52) vs. Not-improved group (n = 24): a group that remained non-LDM and did not improve in the second test. (3) Regarding pregnancy outcomes, the Pre-ET group that was diagnosed as LDM in the initial test and the Pre-ET group that was diagnosed as non-LDM in the initial test but diagnosed as LDM in the second test after treatment were investigated. Pre-ET Good group (n = 45) vs. Pre-ET Improved group (n = 21). (4) Regarding the effect of Lactobacillus iners, a group with ≥ 90% L. iners (n = 4) and a group with ≥ 90% Lactobacillus other than L. iners (n = 15) were investigated. (5) ROC analysis was performed to determine the minimum abundance of L. iners that adversely affects pregnancy
The results are organized and analyzed based on the definitions of LDM and non-LDM [11]. This investigation is described with a focus on the following five points: status of the endometrial microbiome in IVF patients; treatment and its effects when the endometrial microbiome was non-LDM (changes in diversity and the effect of antibiotics); pregnancy outcomes; differences in effect depending on the species of Lactobacillus; and assessment of the effect of L. iners on pregnancy. The last three were assessed after embryo transfer.
Status of the endometrial microbiome in in vitro fertilization patients (n = 257)
The results of the first test in IVF patients were analyzed (Fig. 2 and Table 1). Although the endometrial microbiome at the time of the first test was predominantly Lactobacillus in many cases, there were a fair number of cases in which Gardnerella or other bacteria associated with bacterial vaginosis were present. When the cutoff value for the Lactobacillus percentage was taken to be 90%, the LDM percentage was 61.1%, and the non-LDM percentage was 38.9%.
Fig. 2.
Results of endometrial microbiome analysis in the initial test during the study period in patients who visited the clinic. The proportion of Lactobacillus is shown in blue
Table 1.
Average percentage of bacteria in the endometrial microbiome: patients’ data (as shown in Fig. 2)
| Name of bacterium (genus name) | Mean percentage (%) |
|---|---|
| Lactobacillus | 69.13 |
| Gardnerella | 9.57 |
| Atopobium | 5.41 |
| Streptococcus | 4.24 |
| Bifidobacterium | 2.66 |
| Prevotella | 1.97 |
| Escherichia | 1.58 |
| Megasphaera | 0.70 |
| Klebsiella | 0.66 |
| Aerococcus | 0.63 |
| Enterobacteriaceae | 0.50 |
| Dialister | 0.42 |
| Alloscardovia | 0.33 |
| Ureaplasma | 0.26 |
| Others | 1.93 |
Treatment and its effects when the endometrial microbiome was non-LDM
Treatment was provided to patients in the non-LDM group. They were the group of patients (n = 233) among the population who were non-LDM on the first test, and they excluded those who did not want and so did not take a second test afterward (n = 24). The background characteristics of participants in whom the treatment improvement effect was analyzed are shown in Table 2. In the first test, the group determined to be LDM was taken as the Good group (n = 157), and of the 100 patients first determined to be non-LDM, those determined to be LDM in the second test as a result of treatment were taken to be the Improved group I (Post-treatment group) (n = 52). In addition, the pre-treatment status of this Improved group I (Going to LDM = Pre-treatment group) was also taken as a participant group, and the three groups were compared.
Table 2.
Background characteristics of patients who were analyzed to examine the treatment improvement effect (n = 233) in Fig. 3
| Mean ± SD Number detected (detection rate) |
||||
|---|---|---|---|---|
| Item | Good Group (n = 157) |
Improved Group I (n = 52) |
Not Improved Group (n = 24) |
p value |
| Age (y) | 37.5 ± 4.0 | 36.6 ± 3.8 | 37.1 ± 3.8 | 0.390 |
| Height (cm) | 159.0 ± 5.18 | 157.6 ± 5.2 | 157.5 ± 6.1 | 0.159 |
| Weight (kg) | 52.6 ± 7.8 | 54.0 ± 10.4 | 51.0 ± 10.1 | 0.282 |
| Aspirin administration (yes/no) | 48/157 (30.6%) | 12/51 (23.5%) | 6/24 (25.0%) | 0.601 |
| Results of CD138 examination | 20/65 (30.8%) | 3/13 (23.1%) | 6/13 (46.2%) | 0.492 |
| Pregnancy history | 83/157 (52.9%) | 30/48 (62.5%) | 12/24 (50.0%) | 0.455 |
| Birth history | 36/157 (22.9%) | 20/48 (41.7%) | 10/24 (41.7%) | 0.015 |
| Miscarriage history | 55/157 (35.0%) | 13/25 (52.0%) | 9/21 (42.9%) | 0.215 |
To analyze the treatment improvement effect (from the aspect of bacterial diversity), an assessment was done by comparing each of the three groups defined above: Good group (n = 157) vs. Pre-treatment group (n = 52) vs. Improved Group I (Post-treatment group) (n = 52) (Fig. 3 and Table 3).
Fig. 3.
Improvement in the treatment effect. Endometrial microbiome data in the Good group (n = 157) vs. the Pre-treatment group (n = 52) vs. the Post-treatment group (n = 52) (Improved Group I). The proportion of Lactobacillus is shown in blue. Pie chart of the Good group: Results of the first test in the group that was LDM on the first test. Pie chart of the Pre-treatment group: Results of the first test in the group that was non-LDM on the first test and became LDM in the second test. Pie chart of the Post-treatment group (Improved Group I): Results of the second test in the group that was non-LDM in the first test and became LDM in the second test
Table 3.
Details of bacterial abundance in the Good group vs. the Pre-treatment group vs. the Post-treatment group in Fig. 3
| Good group | Pre-treatment group | Post-treatment group | |
|---|---|---|---|
| Lactobacillus | 98.75 | 22.33 | 99.40 |
| Gardnerella | 0.22 | 23.51 | 0.05 |
| Atopobium | 0.01 | 11.95 | 0.05 |
| Streptococcus | 0.11 | 7.48 | 0.00 |
| Escherichia | 0.02 | 7.26 | 0.00 |
| Bifidobacterium | 0.06 | 6.62 | 0.06 |
| Prevotella | 0.16 | 5.94 | 0.03 |
| Megasphaera | 0.00 | 2.70 | 0.00 |
| Enterobacteriaceae | 0.04 | 1.92 | 0.00 |
| Klebsiella | 0.01 | 1.74 | 0.02 |
| Dialister | 0.04 | 1.34 | 0.01 |
| Aerococcus | 0.01 | 1.18 | 0.04 |
| Enterococcus | 0.00 | 0.78 | 0.00 |
| Sneathia | 0.01 | 0.68 | 0.00 |
| Others | 0.55 | 4.58 | 0.33 |
In the Pre-treatment group, the Lactobacillus spp. percentage was low, and there was much Gardnerella or other bacteria associated with bacterial vaginosis, but in Improved Group I (Post-treatment group), the endometrial microbiome very much resembled that of the Good group that was originally determined to be LDM, with Lactobacillus dominant. The similarity (diversity) of the two was investigated with α diversity analysis and β diversity analysis (Figs. 4 and 5).
Fig. 4.
Treatment improvement effect: α diversity analysis. Good group (n = 157) vs. Pre-treatment group (n = 52) vs. Post-treatment group (n = 52) (Improved Group I). Species uniformity is seen with the Shannon Index, species abundance with Chao1 Richness, and phylogenetic diversity with PD Whole Tree. ***: p < 0.001, **: p < 0.01, *: p < 0.05, n.s.: p > 0.05
Fig. 5.
Treatment improvement effect: β diversity analysis. Good group (blue squares) vs. Pre-treatment group (red circles) vs. Post-treatment group (yellow triangles) (Improved Group I). PERMANOVA test: p = 0.001
In the Post-treatment group, diversity was shown to be less than in the Pre-treatment group and to approach that in the Good group after treatment. The uniformity of species is shown with the Shannon Index, the abundance of species with Chao1, and the phylogenetic diversity with PD Whole Tree. The results of β diversity analysis showed the Post-treatment group to have an endometrial microbiome close to that of the Good group, as noted above in Fig. 3 and Table 3. This could be said to demonstrate that diversity converges as a result of treatment intervention (PERMANOVA test: p = 0.001). Thus, the results of a comparison of diversity in the groups showed that, in the Post-treatment group, not only did the Lactobacillus percentage improve, but diversity converged and approached that of the Good group, indicating improvement in the state of the endometrial microbiome (Fig. 5).
In the following, the question of which treatment methods (antibiotics and probiotics) were effective is examined. The following two groups were investigated for the purpose of analyzing the improvement rate according to treatment policy: Improved group I (Post-treatment) (n = 52), which improved on the second test, vs. the Not-improved group (n = 24), which did not improve on the second test and remained non-LDM. The improvement rate with each treatment policy is described below. The group with treatment centered on metronidazole (metronidazole protocol) [17, 18] had a better improvement rate than the group who received treatment considering bacterial sensitivity (sensitivity protocol) (Table 4). The improvement rates (percentage of cases that were non-LDM on the first test, but became LDM on the second test with treatment) were compared using Fisher’s exact test with differences in treatment method (metronidazole protocol vs. sensitivity protocol), and the results showed a significantly higher rate with the metronidazole protocol than with the sensitivity protocol (LDM cutoff was taken to be a Lactobacillus percentage of 90%) (Table 4). For patients whose microbiome did not recover with the first treatment, further treatment with a combination of the metronidazole and sensitivity protocols was given, and the final improvement rate, including second treatment, was 81.4%.
Table 4.
Comparison of improvement rates by treatment protocol: Metronidazole protocol vs. Sensitivity protocol (n = 76)
| Improved group I | Not-improved group | Improvement rate | |
|---|---|---|---|
| Metronidazole protocol | 45 | 14 | 76.3% |
| Sensitivity protocol | 7 | 10 | 41.2% |
Fisher’s exact test, p = 0.015
Pregnancy outcomes
An association analysis was performed for pregnancy outcome in the Pre-ET Good group vs. the Pre-ET Improved group (Fig. 1). It was previously reported by another research group that the pregnancy rate was higher in an LDM group than in a non-LDM group [11]. Therefore, the present study was performed to investigate whether the pregnancy rate in the Improved group that became LDM with treatment recovered to the same level as the Good group, in which the endometrial microbiome was originally determined to be LDM and was in a good state. When the backgrounds of the patients who were included in the analysis were compared between the successful pregnancy group and the unsuccessful pregnancy group, significant differences were seen in age, whether aspirin was administered, and embryo grade. To adjust the patient backgrounds of the two groups, patients below the age of 37 years were analyzed again (Table 5, < 37 years old). After narrowing down the analysis subjects to those who were under 37 years of age, there was no longer a significant difference between the two groups. A comparison of the pregnancy success rate for the Pre-ET Good group vs. the Pre-ET Improved group after adjustment is shown in Table 7.
Table 5.
Comparison of background characteristics between the pregnancy success group and the pregnancy failure group in patients who underwent analysis of pregnancy success and of the microbiome
| All embryo transfer patients (n = 170) | Under 37 years of age (n = 66) | |||||||
|---|---|---|---|---|---|---|---|---|
| Successful pregnancy (GS = 1) |
Unsuccessful pregnancy (GS = 0) | P value | Test | Successful pregnancy (GS = 1) |
Unsuccessful pregnancy (GS = 0) | P value | Test | |
| No | 82 | 88 | - | 39 | 27 | |||
| Age (y) | 36.2 ± 3.8 | 38.1 ± 3.7 | 0.002* | Student’s t-test | 32.9 ± 2.2 | 33.7 ± 1.8 | 0.172 | Wilcoxon rank-sum test |
| Height (cm) | 158.9 ± 5.3 | 158.8 ± 5.3 | 0.907 | Student’s t-test | 158.0 ± 5.3 | 158.3 ± 4.3 | 0.768 | Student’s t-test |
| Weight (kg) | 53.4 ± 9.5 | 52.9 ± | 0.978 | Wilcoxon rank-sum test | 53.6 ± 11.1 | 52.5 ± 7.7 | 0.945 | Wilcoxon rank-sum test |
| Aspirin administration (yes/no) | 17/82 (20.7%) | 31/88 (35.2%) | 0.042* | Fisher’s exact test | 5/39 (12.8%) |
7/27 (25.9%) |
0.206 | Fisher’s exact test |
| CD138 examination result |
8/19 (42.1%) |
11/40 (27.5%) | 0.372 | Fisher’s exact test |
0/4 (0.0%) |
3/10 (30.0%) |
0.506 | Fisher’s exact test |
| Embryo grade | 75/82 (91.5%) | 65/88 (73.9%) | 0.004* | Fisher’s exact test | 39/39 (100.0%) | 25/27 (92.6%) | 0.164 | Fisher’s exact test |
| Pregnancy history | 37/79 (46.8%) | 48/87 (55.2%) | 0.351 | Fisher’s exact test | 14/37 (37.8%) |
9/27 (33.3%) |
0.795 | Fisher’s exact test |
| Birth history | 19/79 (24.1%) | 26/87 (29.9%) | 0.485 | Fisher’s exact test | 6/37 (16.2%) |
3/27 (11.1%) |
0.722 | Fisher’s exact test |
| Miscarriage history | 25/69 (36.2%) | 27/77 (35.1%) | 1.000 | Fisher’s exact test | 9/30 (30.0%) |
5/25 (20.0%) |
0.537 | Fisher’s exact test |
Table 7.
Pregnancy success rate in women under 37 years of age (n=66) (3) Pre-ET Good Group vs. Pre-ET Improved Group (Fisher’s exact test)
| A. Comparison of pregnancy rate (p=0.065) | |||
| Successful pregnancy | Unsuccessful pregnancy | Success rate | |
| Pre-ET Good Group | 23 | 22 | 51.1% |
| Pre-ET Improved Group | 16 | 5 | 76.2% |
| B. Comparison of ongoing pregnancy rate (p = 0.430) | |||
| Successful pregnancy | Unsuccessful pregnancy | Success rate | |
| Pre-ET Good Group | 20 | 25 | 44.4% |
| Pre-ET Improved Group | 12 | 9 | 57.1% |
There were no significant differences in the background characteristics that should be considered between these two groups (Table 6; Fisher’s exact test). A similar trend was also seen for the ongoing pregnancy rate. The pregnancy success rate and the ongoing pregnancy rate in all transfer cases were equivalent in the Pre-ET Good group vs. the Improved group (Supplemental Table 1) (Table 7).
Table 6.
Comparison of background characteristics between the Pre-ET Good Group and the Pre-ET Improved Group in patients aged 37 years or younger (n=66)
| Under 37 years of age | ||||
|---|---|---|---|---|
| Pre-ET Good Group |
Pre-ET Improved Group |
P value | Test | |
| No. of data | 45 | 21 | ||
| Age (y) | 33.1±2.2 | 33.4±1.8 | 0.780 | Wilcoxon rank-sum test |
| Height (cm) | 159.4±4.6 | 155.4±4.4 | 0.002 | Student’s t-test |
| Weight (kg) | 52.2±8.5 | 55.1±12.0 | 0.657 | Wilcoxon rank-sum test |
| Aspirin administration (yes/no) |
8/45 (17.8%) |
4/21 (19.0%) |
1.000 | Fisher’s exact test |
| CD138 examination result |
3/11 (27.3%) |
0/3 (0.0%) |
1.000 | Fisher’s exact test |
| Embryo grade |
43/45 (95.6%) |
21/21 (100.0%) |
1.000 | Fisher’s exact test |
| Pregnancy history |
14/45 (31.1%) |
9/19 (47.4%) |
0.260 | Fisher’s exact test |
| Birth history |
4/45 (8.9%) |
5/19 (26.3%) |
0.111 | Fisher’s exact test |
| Miscarriage history |
10/45 (22.2%) |
4/10 (40.0%) |
0.255 | Fisher’s exact test |
Differences in effect depending on species of Lactobacillus
In recent years, it has been reported that not only the proportion of Lactobacillus, but also the level of contribution to pregnancy differs depending on the species. In particular, L. crispatus is known to have a positive effect on pregnancy, whereas L. iners has a negative effect on pregnancy. Since this investigation was conducted with a focus on L. iners, the LDM Pre-ET Improved Group < 37 years (n = 21) was divided into two groups as follows: a group in which L. iners comprised ≥ 90%, almost all L. iners (n = 4), vs. a group in which bacteria other than L. iners comprised ≥ 90%, almost no L. iners (n = 15). A group with a mixture of various Lactobacillus species (mixed group, n = 2) was excluded. When the pregnancy success rate and ongoing pregnancy rate were compared, the group that included almost no L. iners was found to have a significantly higher pregnancy success rate (p = 0.037) (Table 8). Thus, the pregnancy success rate in the Improved group, in which a disrupted endometrial microbiome was thought to be an infertility factor, was even higher with the exclusion of cases that included much L. iners. A comparison of pregnancy outcomes for each Lactobacillus species in the Pre-ET Improved group showed trends for higher pregnancy success rates and ongoing pregnancy rates when the only Lactobacillus species was L. crispatus or L. gasseri than when it was only L. iners (Supplemental Table 2).
Table 8.
Comparison of pregnancy success rates with and without Lactobacillus iners (L. iners) in the Improved group under 37 years of age
| A. Comparison of pregnancy rate (p=0.037) | |||
| Successful pregnancy | Unsuccessful pregnancy | Success rate | |
| Mostly L. iners Group in which L. iners was ≥90％ | 1 | 3 | 25.0% |
| Few L. iners Group in which Lactobacillus species other than L. iners were ≥90％ | 13 | 2 | 86.7% |
| B. Comparison of ongoing pregnancy rate (p=0.303) | |||
| Successful pregnancy | Unsuccessful pregnancy | Success rate | |
| Mostly L. iners Group in which L. iners was ≥90％ | 1 | 3 | 25.0% |
| Few L. iners Group in which Lactobacillus species other than L. iners were ≥90％ | 9 | 6 | 60.0% |
Assessment of the effect of L. iners on pregnancy
An additional ROC analysis was conducted to determine how large the proportion of L. iners must be to act negatively on pregnancy (Fig. 6 and Table 9).
Fig. 6.
Investigation of the cutoff value to predict pregnancy success with the L. iners percentage. When limited to the Improved group in patients under 37 years of age, an ROC analysis was performed to see how much the L. iners percentage explains pregnancy success, and the AUC was 0.713. The Youden index was calculated from the ROC analysis results, and the optimal cutoff value for the L. iners percentage to predict pregnancy success was calculated to be 74.2%
Table 9.
Cutoff value of L. iners abundance for predicting pregnancy success (See legend of Figure 6)
| Cutoff value ofL. iners abundance | NLiDM | LiDM | Sensitivity | Specificity | Accuracy | P value | Youden Index | ||
|---|---|---|---|---|---|---|---|---|---|
| PregnancySuccess | PregnancyFailure | PregnancySuccess | PregnancyFailure | ||||||
| 74.2 | 15 | 2 | 1 | 3 | 0.938 | 0.600 | 0.857 | 0.028 | 0.538 |
| 48.2 | 14 | 2 | 2 | 3 | 0.875 | 0.600 | 0.810 | 0.063 | 0.475 |
| 21.6 | 13 | 2 | 3 | 3 | 0.813 | 0.600 | 0.762 | 0.115 | 0.413 |
| 0.1 | 12 | 2 | 4 | 3 | 0.750 | 0.600 | 0.714 | 0.280 | 0.350 |
| 91.5 | 15 | 3 | 1 | 2 | 0.938 | 0.400 | 0.810 | 0.128 | 0.338 |
| 94.5 | 15 | 4 | 1 | 1 | 0.938 | 0.200 | 0.762 | 0.429 | 0.138 |
| 97.6 | 15 | 5 | 1 | 0 | 0.938 | 0.000 | 0.714 | 1.000 | −0.063 |
The result showed an AUC of 0.713. The Youden index was also calculated from the results of the ROC analysis, and the optimal cutoff value for the L. iners percentage to predict pregnancy success was calculated to be 74.2%. From this result, the endometrial microbiome in which L. iners accounts for 74.2% or more is highly likely to have a negative effect on pregnancy. In LDM cases, the percentage of L. iners was 74.2% or higher in 21% of cases. Thus, there is a possibility that, even in LDM cases, there are patients who will require treatment.
Discussion
The following findings were demonstrated as a result of an investigation at our clinic of the endometrial microbiome of IVF patients. Although Lactobacillus was dominant, it showed an equivalent or better pregnancy success rate. The level of contribution to pregnancy of Lactobacillus may differ depending on the species, with L. crispatus and L. gasseri tending to act positively, whereas L. iners acts negatively. The percentage of L. iners for it to have a negative effect on pregnancy was ≥ 74.2%. The fact that L. iners acts negatively on pregnancy has previously been reported, but this is the first report to show a base level for it. L. iners accounted for ≥ 74.2% of the endometrial microbiome in 21% of LDM patients. Thus, it is possible that, among LDM patients, there are some who will need treatment. If a treatment protocol for L. iners is established in the future, further improvement may be achieved.
In the present study, the endometrial microbiome of IVF patients was investigated, and though there was agreement with other reports [18], the condition of the endometrial microbiome of IVF patients has become clearer.
Diversity was shown to be less in the Post-treatment group than in the Pre-treatment group and to approach that in the Good group after treatment. It could be said that this demonstrates that diversity converges as a result of treatment intervention. Thus, the results of a comparison of diversity in the groups showed that, in the Post-treatment group, not only did the Lactobacillus percentage improve, but diversity converged and approached that of the Good group, indicating improvement in the state of the endometrial microbiome. Diversity converged when the uterine environment was improved following treatment and Lactobacillus became dominant, but with respect to this point, there are major differences with the intestines where diversity is considered to be important [25]. The results of the present investigation showed that a strategy using metronidazole as the first treatment was more effective in improving the endometrial microbiome. Metronidazole is an antibiotic thought to be very effective against anaerobic bacteria below the diaphragm, and it has also been shown to be effective in the treatment of the endometrial microbiome that has been replaced with anaerobic bacteria. This is also supported by the fact that metronidazole is superior in terms of tissue transference into the pelvis. In strategies against detected bacteria, the following two points are important. The abundance and type of Lactobacillus are more important the first time, and metronidazole is the basic strategy. However, if metronidazole is ineffective, for the second and subsequent times, an antibiotic to which the detected bacterium is sensitive needs to be selected, since it has been demonstrated that the species is one that cannot be treated with metronidazole. Although Lactobacillus was dominant in the endometrial microbiome of many IVF patients, there were also a fair number of patients with Gardnerella or other bacteria associated with bacterial vaginosis. It could be said that bacterial vaginosis is a state in which a healthy endometrial microbiome has been replaced with anaerobic bacteria, and metronidazole would naturally be the gold standard treatment. It is thought that a personalized approach should be based on this. The best management at present is to assess the abundance of Lactobacillus in the initial sequence and treat it with metronidazole. For cases that do not respond to this, using an antibiotic that the detected bacteria are sensitive to would be better. The method of using an antibiotic that the detected bacteria are sensitive to in addition to metronidazole in the initial management (for example, methods such as adding clindamycin to initial metronidazole for Streptococcus) should also obviously be considered, but from the perspective of overuse of antibiotics, such an approach should be adopted cautiously.
In a comparison of pregnancy outcomes in the initial LDM group (Good group) and the group that was initially non-LDM but became LDM with improvement following treatment (Improved group I), Improved group I was found to have better outcomes than the Good group. One of the factors related to infertility in Improved group I was disruption of the endometrial microbiome, and the outcomes in this group could be said to have been improved by improving the endometrial microbiome. In contrast, in the initial LDM group (Good group), in which disruption of the endometrial microbiome was not a cause of infertility, and there was no room for improvement, there were not thought to be any factors that could contribute to improved outcomes.
Lactobacillus spp. is known to have very different biological activities depending on the species. The main species confirmed to be present in the uterus include L. crispatus, L. gasseri, L. jensenii, and L. iners [9]. Of them, the lactic acid production capacity of L. iners is known to differ from that of other Lactobacillus species. In a paper by Witkin et al., it was reported that the lactic acid production capacity of L. crispatus was superior to that of L. iners [26], showing the utility of L. crispatus in terms of its antimicrobial effect. In addition, L. iners is known to produce small amounts of H2O2 and to generate inerolysin, a cytotoxic toxin, so that differences in this function may also affect the endometrium [27]. A comparison of the pregnancy success rate and ongoing pregnancy rate in two groups, an LDM group with almost all L. iners (≥ 90%) (n = 4) vs. an LDM group with almost no L. iners, in which species other than L. iners account for ≥ 90% (n = 15), showed that the group that included almost no L. iners had a significantly higher pregnancy success rate (p = 0.037) in the present study. From this, it would seem that it is necessary to compare the clinical results between patients who are L. iners-dominant and all other patients. Whether L. iners itself is a pathogenic bacterium, whether L. iners creates a poor environment, or whether this is a state prior to a transition to non-LDM, should be clarified in the future.
Furthermore, although L. iners has been reported to act negatively on pregnancy, the specific threshold for such action is unknown. In the present study, that value was ≥ 74.2%, which shows the base level for the first time. In LDM patients, those in whom L. iners was ≥ 74.2% accounted for 21%, and it may be that, even among LDM cases, there are some cases that will require treatment. In treating L. iners-dominant individuals, there may be cases that can be handled without the use of antibiotics and with nutritional intervention alone to improve the environment, such as prebiotics and probiotics. Establishment of a treatment protocol for L. iners in the future may provide hope for further improvement.
Collection of uterine fluid can sometimes be technically difficult and painful for the patient, so some have considered using the vaginal microbiome instead. With regard to differences in the vaginal microbiome and endometrial microbiome, though they are close and affect each other, caution is needed because they may differ depending on the patient [11, 17]. These reports have a concordance rate of around 80% at best, meaning that cases that should be treated may be missed. Thus at present, the most reliable method of selecting an antibiotic is to directly assess the uterus.
Limitations
The number of cases in the present study was small, and the results were limited, especially regarding pregnancy. Furthermore, although 80% of non-LDM patients improved and became LDM, the number of cases was limited because some non-LDM patients did not agree to a second examination.
Moreover, factors on the embryo side and those on the uterine side that affect pregnancy should each be considered, and though an ideal investigation would be to conduct PGT-A in all cases and eliminate factors on the embryo side, this cannot be done in Japan due to restrictions imposed by the insurance system, and a method was adopted that filters the effects of the embryo by age and morphological assessment of the embryo.
Conclusion
The results of the present investigation of the endometrial microbiome in IVF patients at our clinic elucidated the endometrial microbiome of IVF patients in detail. The rate of improvement of the endometrial microbiome with the treatment protocol at our institution was 81.4%. The Improved group recovered an endometrial microbiome pattern that was the same as the group that was LDM from the start (Good group), and it showed an equivalent or better pregnancy success rate. In addition, the possibility that L. iners at a percentage ≥ 74.2% acts negatively on pregnancy was found in this study. It is our hope that this information will be beneficial for the improvement of the implantation environment.
Supplementary Information
Below is the link to the electronic supplementary material.
(DOCX 22.7 KB)
(DOCX 23.4 KB)
Author contribution
All authors have reviewed the final version to be published and agreed to be accountable for all aspects of the work. Concept and design: MH. Critical review of the manuscript for important intellectual content: MH, K H, MY. Acquisition and analysis of data: MH. Writing of the manuscript: MH.
Funding
Not applicable.
Data availability
The data that support the findings of this study are available from the corresponding author, MH, upon reasonable request.
Declarations
Ethical approval and consent to participate
All procedures followed were in accordance with the ethical standards of the responsible committee on human experimentation (institutional and national) and with the Helsinki Declaration of 1964 and its later amendments. Informed consent was obtained from all patients for being included in the study. This article does not contain any studies with human or animal subjects performed by any of the authors. The Committee of Hanaoka IVF Clinic Shinagawa approved the collection and use of biological materials for this study.
Conflict of interest
The authors declare no competing interests.
Footnotes
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.Liang Z, Wen Q, Li J, Zeng D, Huang P. A systematic review and meta-analysis: clinical outcomes of recurrent pregnancy failure resulting from preimplantation genetic testing for aneuploidy. Front Endocrinol (Lausanne). 2023;14:1178294. 10.3389/fendo.2023.1178294. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.McQueen DB, Perfetto CO, Hazard FK, Lathi RB. Pregnancy outcomes in women with chronic endometritis and recurrent pregnancy loss. Fertil Steril. 2015;104:927–31. 10.1016/j.fertnstert.2015.06.044. [DOI] [PubMed] [Google Scholar]
- 3.Zolghadri J, Momtahan M, Aminian K, Ghaffarpasand F, Tavana Z. The value of hysteroscopy in diagnosis of chronic endometritis in patients with unexplained recurrent spontaneous abortion. Eur J Obstet Gynecol Reprod Biol. 2011;155:217–20. 10.1016/j.ejogrb.2010.12.010. [DOI] [PubMed] [Google Scholar]
- 4.Bouet PE, El Hachem H, Monceau E, Gariépy G, Kadoch IJ, Sylvestre C. Chronic endometritis in women with recurrent pregnancy loss and recurrent implantation failure: prevalence and role of office hysteroscopy and immunohistochemistry in diagnosis. Fertil Steril. 2016;105:106–10. 10.1016/j.fertnstert.2015.09.025. [DOI] [PubMed] [Google Scholar]
- 5.Kitaya K, Matsubayashi H, Takaya Y, Nishiyama R, Yamaguchi K, Takeuchi T, et al. Live birth rate following oral antibiotic treatment for chronic endometritis in infertile women with repeated implantation failure. Am J Reprod Immunol. 2017. 10.1111/aji.12719. [DOI] [PubMed] [Google Scholar]
- 6.Cicinelli E, Matteo M, Trojano G, Mitola PC, Tinelli R, Vitagliano A, et al. Chronic endometritis in patients with unexplained infertility: prevalence and effects of antibiotic treatment on spontaneous conception. Am J Reprod Immunol. 2018. 10.1111/aji.12782. [DOI] [PubMed] [Google Scholar]
- 7.Johnston-MacAnanny EB, Hartnett J, Engmann LL, Nulsen JC, Sanders MM, Benadiva CA. Chronic endometritis is a frequent finding in women with recurrent implantation failure after in vitro fertilization. Fertil Steril. 2010;93:437–41. 10.1016/j.fertnstert.2008.12.131. [DOI] [PubMed] [Google Scholar]
- 8.Yasuo T, Kitaya K. Challenges in clinical diagnosis and management of chronic endometritis. Diagnostics. 2022;12:2711. 10.3390/diagnostics12112711. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Baker JM, Chase DM, Herbst-Kralovetz MM. Uterine microbiota: residents, tourists, or invaders? Front Immunol. 2018;9:208. 10.3389/fimmu.2018.00208. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Franasiak JM, Scott RT Jr. Reproductive tract microbiome in assisted reproductive technologies. Fertil Steril. 2015;104:1364–71. 10.1016/j.fertnstert.2015.10.012. [DOI] [PubMed]
- 11.Moreno I, Codoñer FM, Vilella F, Valbuena D, Martinez-Blanch JF, Jimenez-Almazán J, et al. Evidence that the endometrial microbiota has an effect on implantation success or failure. Am J Obstet Gynecol. 2016;215:684–703. 10.1016/j.ajog.2016.09.075. [DOI] [PubMed] [Google Scholar]
- 12.Dominguez F, Gadea B, Mercader A, Esteban FJ, Pellicer A, Simón C. Embryologic outcome and secretome profile of implanted blastocysts obtained after coculture in human endometrial epithelial cells versus the sequential system. Fertil Steril. 2010;93:774-782.e1. 10.1016/j.fertnstert.2008.10.019. [DOI] [PubMed] [Google Scholar]
- 13.Cha J, Vilella F, Dey SK, Simón C. Ten critical topics in reproductive medicine. Molecular interplay in successful implantation. Washington: Sanders; 2013. pp. 44–8.
- 14.Antimicrobial Resistance Collaborators. Global burden of bacterial antimicrobial resistance in 2019: a systematic analysis. Lancet. 2022;399:629–55. 10.1016/S0140-6736(21)02724-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Tsuzuki S, Matsunaga N, Yahara K, Gu Y, Hayakawa K, Hirabayashi A, et al. National trend of blood-stream infection attributable deaths caused by Staphylococcus aureus and Escherichia coli in Japan. J Infect Chemother. 2020;26:367–71. 10.1016/j.jiac.2019.10.017. [DOI] [PubMed] [Google Scholar]
- 16.Ichiyama T, Kuroda K, Nagai Y, Urushiyama D, Ohno M, Yamaguchi T, et al. Analysis of vaginal and endometrial microbiota communities in infertile women with a history of repeated implantation failure. Reprod Med Biol. 2021;20:334–44. 10.1002/rmb2.12389. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Kadogami D, Nakaoka Y, Morimoto Y. Use of a vaginal probiotic suppository and antibiotics to influence the composition of the endometrial microbiota. Reprod Biol. 2020;20:307–14. 10.1016/j.repbio.2020.07.001. [DOI] [PubMed] [Google Scholar]
- 18.Kadogami D, Kimura F, Hanada T, Tsuji S, Nakaoka Y, Murakami T, et al. Impact of Lactobacillus in the uterine microbiota on in vitro fertilization outcomes. J Reprod Immunol. 2023;160:104138. 10.1016/j.jri.2023.104138. [DOI] [PubMed] [Google Scholar]
- 19.Bolger AM, Lohse M, Usadel B. Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics. 2014;30:2114–20. 10.1093/bioinformatics/btu170. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Erik A. Comparison of sequencing utility programs. Open Bioinforma J. 2013;7:1–8. PMCID: PMC8081347
- 21.Schmieder R, Edwards R. Quality control and preprocessing of metagenomic datasets. Bioinformatics. 2011;27:863–4. 10.1093/bioinformatics/btr026. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, et al. QIIME allows analysis of high-throughput community sequencing data. Nat Methods. 2010;7:335–6. 10.1038/nmeth.f.303. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Quast C, Pruesse E, Yilmaz P, Gerken J, Schweer T, Yarza P, et al. The SILVA ribosomal RNA gene database project: improved data processing and web-based tools. Nucleic Acids Res. 2013;41:D590–6. 10.1093/nar/gks1219. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Fettweis JM, Serrano MG, Sheth NU, Mayer CM, Glascock AL, Brooks JP, et al. Species-level classification of the vaginal microbiome. BMC Genomics. 2012;13(Suppl 8):S17. 10.1186/1471-2164-13-S8-S17. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 25.Witkin SS, Mendes-Soares H, Linhares IM, Jayaram A, Ledger WJ, Forney LJ. Influence of vaginal bacteria and D- and L-lactic acid isomers on vaginal extracellular matrix metalloproteinase inducer: implications for protection against upper genital tract infections. MBio. 2013;4:e00460-13. 10.1128/mBio.00460-13. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Zheng N, Guo R, Wang J, Zhou W, Ling Z. Contribution of Lactobacillus iners to vaginal health and diseases: a systematic review. Front Cell Infect Microbiol. 2021;11:792787. 10.3389/fcimb.2021.792787. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Johansen J, Atarashi K, Arai Y, Hirose N, Sørensen SJ, Vatanen T, et al. Centenarians have a diverse gut virome with the potential to modulate metabolism and promote healthy lifespan. Nat Microbiol. 2023;8:1064–78. 10.1038/s41564-023-01370-6. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
(DOCX 22.7 KB)
(DOCX 23.4 KB)
Data Availability Statement
The data that support the findings of this study are available from the corresponding author, MH, upon reasonable request.