Homocysteine serum levels correlate with the number of failed IVF cycles even when within normal range

Repeated implantation failure is a common challenge in daily practice. Homocysteine and vitamin B12 have been associated with reproductive processes among patients undergoing in vitro fertilization; however, their involvement in repeated implantation failure has not been assessed. We explored possible associations of serum homocysteine and vitamin B12 with repeated implantation failure. A retrospective analysis of 127 women who underwent ≥ 3 unsuccessful embryo transfers during 2005–2016, at the Fertility and In Vitro Fertilization Unit at Carmel Medical Center. After at least 3 IVF failures serum levels of homocysteine and vitamin B12 were measured. The mean patient age was 33.5 ± 5.2 years. The mean number of embryo transfers was 4.6 ± 1.5. The mean total cumulative number of embryos transferred was 10.4 ± 5.2. Mean serum levels of homocysteine were 8.6 ± 3.7 µM/L, and of vitamin B12 were 302.5 ± 155.3 pg/ml. Homocysteine levels were within the normal range (< 14 µM/L) in 95.8% of the patients. Yet, the levels of homocysteine correlated with both the number of failed embryo transfers (r = 0.34, p = 0.004) and the total cumulative number of transferred embryos (r = 0.36, p = 0.002). Our findings suggest an association between serum homocysteine levels and the occurrence of repeated implantation failure, even when homocystein levels were within the normal range. It should be studied whether nutritional supplementation to modulate serum homocysteine levels may improve treatment outcome.


Introduction
A substantial proportion of infertile couples experience repeated implantation failure (RIF). Though a universal definition of RIF has not been established, a commonly used definition is the failure to achieve pregnancy after the transfer of a total of 4 or more embryos in at least three embryo transfer events [1]. Embryo implantation may be regarded as an optimal result of embryo-endometrial cross talk. RIF 1 3 may be associated with: embryonic aneuploidy, increasing with patient's age [2]; anatomical abnormalities, specifically of the endometrium; hydrosalpinx; thickened zona pellucida; a suboptimal culture medium; and thrombophilia [3]. Solutions for RIF include pregestational genetic testing to select euploid embryos for transfer, operative hysteroscopy in cases of abnormal uterine cavity, assisted zona hatching, blastocyst transfer, pre-transfer endometrial injury, prophylactic anticoagulants at embryo transfer, elective salpingectomy in women with hydrosalpinx, improved culture medium [3] and immunotherapeutic agents [4]. The most significant factor affecting embryo implantation is patient's age [5]. Indeed, women who experienced RIF were older than their counterparts [6]. Nevertheless, the applicability of preimplantation genetic screening to improve rates of implantation is limited to age < 41 years [7], in addition to its prohibitive cost.
Euploid embryos display different developmental kinetic behavior on time-lapse compared to aneuploid embryos. Accordingly, an algorithm was proposed that improved the embryo selection process also in RIF [8].
Several investigations have examined the role of endometrial receptivity in RIF. Shi et al. observed a specific endometrial pattern on mRNA microarray that was associated with RIF [9]. Ruiz-Alonso et al. used endometrial receptivity array, as a therapeutic strategy for improving implantation success in personalized embryo transfers in patients with RIF [10]. However, even after thorough investigation and management of RIF, some couples still undergo unsuccessful treatment cycles.
Several studies have investigated associations of vitamin B12 and homocysteine with female reproduction. Vitamin B12's (cobalamin) cytoplasmic isoform promotes the conversion of homocysteine to methionine by the enzyme methyl transferase. This is essential for the demethylation of tetrahydrofolate, an important intermediate in the formation of nucleic acids. Serum levels of vitamin B12 below 200 pg/ ml are considered low [11]. The daily recommended nutritional dose is 2.4 µg and 2.6-2.8 during pregnancy and lactation [12]. B12's deficiency may be associated with diverse clinical manifestations [11].
Homocysteine is an amino acid formed during the conversion of methionine to cysteine. This process requires folic acid, vitamin B12 and B6. Therefore, any deficiency in these vitamins may result in an elevated homocysteine level [12]. Homocysteine serum levels above 14 µM/L are considered elevated, while levels are lower among children and women [11].
Elevated serum levels of homocysteine may result from increased production of methionine (transmethylation), decreased synthesis of cysteine (transsulfuration), decreased conversion of homocysteine to methionine (remethylation), decreased metabolism or decreased renal excretion of homocysteine, renal insufficiency or malabsorption [11]; and following the use of certain medications [13]. Hyperhomocysteinemia may also result from genetic mutations related to the C677T and A1298C MTHFR (methyltetrahydrofolate reductase) genes [14].
D'Elia et al. suggested that the number and maturity of oocytes in IVF may be related to the MTHFR polymorphism in these two mutations [15]. Furthermore, Enciso et al. found that MTFHR genotypes affect the production of aneuploid embryos, suggesting that MTFHR may modulate rates of chromosomal abnormalities [16]. Clement et al. in a study of male partners of infertile couples with failed ART cycles, found that highest prevalence of MTHFR gene single nucleotide polymorphism (SNPs) in this group were the heterozygous C677T, followed by the combined heterozygous C677T/A1298C, and then A1298C; these three variants represent 65% of the population. These variants correlated with elevated circulating homocysteine level of > 15 µMolar compared to the carriers of the wild type SNP [17]. Similarly, Tara et al. in a study of male partners of couples suffering from repeated (≥ 3) miscarriages, found that significant difference was detected in the frequency of MTHFR SNPs in male partners of the two groups (p = 0.019). Combined heterozygosity of MTHFR polymorphisms was a common phenomenon in the males with repeated miscarriages, compared with the control group (23.1% vs 14% respectively) [18].
Another etiology of hyperhomocysteinemia is inborn errors of metabolism; most frequently, a mutation in the cystathionine β synthase gene. The clinical manifestation of hyperhomocysteinemia is diverse, and includes an increased risk for congenital malformations (neural tube disease) and pregnancy complications such as placental abruption and miscarriages [11].
The intake of folic acid, vitamin B12 and vitamin B6 supplements may lower serum levels of homocysteine in persons with mild hyperhomocysteinemia [11]. Hall and Davidson described infertile patients with pernicious anemia who conceived following supplementation of folic acid [19]; El-Nemr described a patient with RIF and low serum levels of vitamin B12, who conceived spontaneously after supplementation of the vitamin [20]. Both folate and vitamin B12 serum levels were low in 269 women who underwent IVF: < 400 pg/ml and 474 pg/ml, respectively [21]. Higher levels of folate were associated with a better chance of achieving a clinical pregnancy in women who underwent IVF treatment [22].
Homozygotes for the 1298 CC variant of MTHFR were less likely to achieve a live birth after IVF or to have had a previous pregnancy, compared with women with the wild type variant. The authors concluded that MTHFR genotype is linked to a woman's potential to produce healthy embryos, possibly through interaction with genes related to DNA methylation [23].
An inverse correlation was found between homocysteine follicular levels and follicular diameter after ovarian stimulation, suggesting that high levels of homocysteine may adversely affect follicular development [24].
Associations of serum homocysteine and vitamin B12 levels with RIF have not been studied systematically. The present study looked for possible associations of serum homocysteine and vitamin B12 with the extent of RIF in women undergoing IVF.

Materials and methods
The study is a retrospective analysis of women treated by IVF who underwent 3 or more embryo transfer cycles (both fresh and cryopreserved) that did not result in a pregnancy. All the women were treated during 2005-2016 at the Fertility and IVF Unit, Carmel Medical Center. Homocysteine and vitamin B12 serum levels were tested and recorded on the general electronic file as part of the screening for RIF after at least 3 IVF failures. Demographic, clinical and laboratory parameters were also recorded and analyzed. Women with a medical history that could affect serum homocysteine or vitamin B12 levels, or women who used vitamin B12 supplementations, were excluded from the analysis.
The study was approved by the Institutional Review Board. Since this was a retrospective anonymous correlative analysis of electronic medical records, informed consent was not required.
Statistical analysis was performed using IBM statistics (SPSS) version 24. The continuous variables were presented as means ± SD or medians and ranges. Categorical variables were presented as percentages. Correlations of homocysteine and vitamin B12 serum levels with clinical and laboratory features were calculated using Pearson or Spearman correlations, in accordance with data distribution.
Differences in clinical and demographical characteristics between 2 categories of vitamin B12 level (normal, abnormal) and between 2 categories of homocysteine level (below median, above median) were analyzed using the Chi square test for the categorical variables, and the independent t-test or the Mann-Whitney test, as appropriate, for the continuous variables. p < 0.05 was considered as threshold for significance.

Results
In total, 127 women who underwent 3 or more unsuccessful embryo transfer cycles were included in the study. The mean age was 33.5 ± 5.2 years (range: . The mean number of prior embryo transfers was 4.6 ± 1.5. The mean cumulative total number of transferred embryos was 10.4 ± 5.2. The mean serum levels of homocysteine were 8.6 ± 3.7 µM/L (range 4.1-29.3). The mean levels of vitamin B12 were 302.5 ± 155.3 pg/ml (range: 99-1331), (Table 1). No association was found between homocysteine levels and demographic, clinical and laboratory parameters, including the women's age. Homocysteine levels correlated with both the numbers of failed embryo transfers (r = 0.34, p = 0.004) (Fig. 1a) and the cumulative total number of transferred embryos (r = 0.36, p = 0.002) (Fig. 1b). These correlations were not influenced by demographic confounders (age, body mass index, smoking status, infertility type). No correlation was found between vitamin B12 levels and the number of failed embryo transfers, or with the cumulative total number of transferred embryos. The number of embryo transfers and the cumulative total number of embryos were correlated (p < 0.0001 r = 0.87) as expected. Ninety-six percent of the patients had normal homocysteine levels (< 14 µM/L) (Fig. 2). Low (< 200 pg/ml) serum levels of vitamin B12 were inversely associated with women's age: women with normal vitamin B12 levels were older than those with reduced vitamin B12 levels (34.3 ± 5.0 vs. 30.4 ± 5.4 years, p = 0.008). Homocysteine serum levels correlated inversely with vitamin B12 levels (r = − 0.35, p = 0.004).

Disscussion
Our analysis found a correlation between serum homocysteine levels and the extent of RIF, as expressed by the number of failed embryo transfers and the total number of transferred embryos. Our results concur with previous studies that demonstrated an inverse association of homocysteine serum level with female fertility [23,24]. Our findings corroborate those of a prospective study of 263 women that investigated an association between periconceptional maternal biomarkers and embryonic growth. Embryonic growth rates were positively associated with vitamin B12 levels (p < 0.05) and inversely associated with homocysteine serum levels (p < 0.05) [25]. We suggest that defective embryonic implantation in the presence of higher homocysteine levels may underlie even earlier defective growth, presenting as RIF.
Diverse mechanisms can be suggested to explain the correlation between homocysteine serum levels and the extent of RIF. For one, homocysteine and folate may affect folliculogenesis, fertilization and early embryogenesis. Low follicular homocysteine was associated with better quality of oocytes and with a higher degree of maturity [26,27]. Although in the current study we measured homocysteine levels in venous blood, these levels reflect intrafollicular levels [28]. In addition, Ebisch et al. (2006) and others demonstrated an inverse association of higher levels of follicular homocysteine with embryo quality at day 3, as expressed by fragmentation and cleavage stage and less favorable embryo development. [27,29,30].
A possible explanation for a detrimental effect of elevated homocysteine levels on early embryo development relates to the inhibitory effect of homocysteine on methylation [31]. Follicular homocysteine levels are higher after ovarian stimulation [32]. In mammals, methylation patterns in the oocyte change after fertilization. Sato et al. discovered a deficiency in the methylation pattern of oocytes retrieved following ovarian stimulation [33]. This suggests that high homocysteine levels may adversely affect embryo cleavage and development by re-modulation of early embryonic methylation patterns.
Another possibility involves the metabolic pathways for homocysteine recycling to methionine and cysteine. Whereas in many cells, homocysteine is recycled by two pathways, via folate or cysteine beta synthase, in human oocytes, and in the early preimplantation embryo, the latter is poorly expressed or absent, and homocysteine is recycled almost exclusively by the folate pathway [31,34,35].
A third possible explanation relates to the chemical properties of homocysteine, which is a thiol-containing amino acid known to induce the release of reactive oxygen species (ROS) [36]. Elevated levels of ROS in follicular were inversely correlated with oocyte and embryo quality [37,38].
In a previous study, we found that oxidative parameters in follicular fluid, measured by the thermochemiluminescence assay, correlated with parameters of ovarian responsiveness and IVF outcome [39]. In addition, oxidative parameters in embryo culture media were associated with chances of implantation [40]. Oxidative stress produced by elevated intrafollicular homocysteine levels may serve as an additional mechanism that explains our results.
Finally in folate-deficient mice, Li et al. demonstrated that the lack of folate restrained decidual angiogenesis, which resulted in significant abnormalities in vascular density Fig. 2 The distribution of homocysteine levels in the study population. Homocysteine serum levels of 95.8% of the study population were < 14 µM/L (upper normal level) and in the enlargement and elongation of the vascular sinus [41]. Moreover, expression of vascular endothelial growth factor and placental growth factor (VEGFA, VEGFR2 and PLGF, respectively), which are key molecules regulating decidual angiogenesis and maternal spiral artery remodeling, were reduced. Thus, folate deficiency can damage decidual angiogenesis and may be related to the vasculotoxic properties of homocysteine and the lack of balance of the reproductive hormone. In contrast, Gao examined in mice, two essential genes associated with uterine receptivity and embryo implantation: Pgr and Cdh1 [42]. The methylation and expression of both genes were not affected by the lack of folate. In other words, folate deficiency does not impair embryo implantation. However, homocysteine levels were not measured in that study.
Another finding of our study is an inverse correlation of serum levels of homocysteine and vitamin B12 (p < 0.004). These findings are in concordance with the report that vitamin B12 deficiency causes elevated serum homocysteine, by inhibiting its conversion to methionine [43].
An additional interesting finding in our study was that low (< 200 pg/ml) serum levels of vitamin B12 were inversely associated with patients' age (p = 0.008). We do not have a good explanation for this, which should be further studied. An interesting basic research, by Qi et al. performed in endometrial stromal cells in vitro, found that Estradiol-17β stimulated cystathionine-β synthase (that catalyzes the first step of the transsulfuration pathway, from homocysteine to cystathionine) expression and hydrogen sulfide (a potent proangiogenic factor) production in proliferative premenopausal. These results may perhaps shed some light on putative mechanism of homocysteine relation to fetal implantation [44].
The most prominent finding in our study was that even normal serum levels of homocysteine correlated with the extent of RIF. Notably, 95.8% of the patients in our study had a normal homocysteine levels. Our findings suggest that the optimal range of serum homocysteine for successful embryo implantation may be lower than that of other contexts. Interestingly, Rubini et al. found in a periconception cohort, using three dimensional fetal ultrasound and virtual reality techniques, that high maternal homocysteine within the reference range is negatively associated with first trimester embryonic growth and birth weight, and the effects of homocysteine are dependent on conception mode (IVF/ ICSI vs natural conception) [45].
In other previous studies, mild to moderate levels of hyperhomocysteinemia were associated with increased risk for severe cardiovascular outcome [46,47]. In the context of homocysteine and RIF, a more sensitive and higher resolution scale of serum levels may be needed. Since homocysteine levels vary by age, sex, and ethnicity [48] the optimal range needed for successful implantation should be determined from a large-scale, prospective study that would include a diverse population of women of different age groups and ethnic backgrounds.
Our results may be limited by the retrospective nature of our study, the lack of a more fertile control group (women who conceived within the first 3 transfer cycles) and the relatively small sample size. In addition, the measurement of homocysteine serum levels is not routinely performed and costly.
Future studies should determine whether nutritional supplementation or other therapy to modulate serum homocysteine levels may affect IVF outcome. Reducing homocysteine levels by dietary supplements or other means may become an auxiliary tool in the management of RIF. This hypothesis cannot be tested other than by a future prospective randomized double-blind controlled trial. Whereas applicability of preimplantation genetic screening was limited to younger women, alleviating homocysteine levels by dietary supplements may be inexpensive and more accessible [7].

Conclusion
Our study found an association between serum homocysteine levels and severity of repeated implantation failure, even when homocysteine levels were within the normal range. Future studies may show whether nutritional supplementation or other therapy to modulate serum homocysteine levels may improve treatment outcome.
Author contributions NF-M: Acquisition of data, Statistical analysis. HG: Critical revision of the manuscript for important intellectual content. IF: Analysis and Interpretation of data. SL-B: Analysis and Interpretation of data. IB-S: Supervision, Critical revision of the manuscript for important intellectual content. ZW-M: Conception and design of the manuscript, Critical revision of the manuscript for important intellectual content.

Funding
The study was not founded.
Data availability All data and study materials are available if needed.

Conflict of interest
The authors declare that they have no conflict of interest regarding the publication of this article.

Ethical approval
The study was approved by the institutional ethical review board (IRB).