The Periconceptional Healthy Dietary Pattern Mitigates Estradiol Response and Increases the Chance of Pregnancy after Conventional IVF/ICSI Treatment

Introduction

In Vitro Fertilization and Intracytoplasmic Sperm Injection (IVF/ICSI) are marginally successful techniques to achieve pregnancy in subfertile couples [1]. The propensity for IVF/ICSI treatment success is also determined by a sufficient ovarian response to Controlled Ovarian Hyperstimulation (COH), i.e. the number of growing ovarian follicles and the quantity of steroid hormone production [2]. If one can identify modifiable factors that affect the ovarian response to COH as well as the chance of ongoing pregnancy, new opportunities become available to improve and further personalize IVF/ICSI treatment. COH and supraphysiological estradiol (E2) levels observed during COH cycles detrimentally affect endometrial development [3, 4, 5]. Therefore, during traditional IVF/ICSI treatment, the embryo is possibly transferred to a relatively unreceptive endometrium [5]. There has been much debate about the relevance of the ovarian response on clinical outcomes, such as ongoing pregnancy [6, 7, 8, 9]. It could be helpful to not just focus on the influence of the ovarian response on IVF/ICSI treatment outcome through the effects of steroid hormone production on the endometrium, but to consider the influence of the ovarian response (of which E2 is a marker) on normal oocyte and follicular development [10, 11, 12].

Demographic, ultrasound and endocrine markers can predict the ovarian response to COH, and provide opportunities for a more personalized treatment [2]. In addition, there are various treatments to control the ovarian response to COH. Nevertheless, there is still residual variability in the ovarian response that cannot be explained by such factors, where possibly maternal environmental factors also play a role. Unfortunately, except for overweight, such factors comprise non- modifiable factors, with little variability between treatment cycles and are therefore not suitable to modify the ovarian response to COH [13, 14]. Modifiable factors, such as diet and other lifestyle factors could help control and modify the ovarian response to COH within the context of a woman’s demographic, ultrasound and endocrine constitution. In previous studies, we and others showed that periconceptional folic acid supplement use mitigates the E2 response to COH and improves embryo quality. Furthermore, the periconceptional diet and lifestyle associate with the subsequent chance of pregnancy after IVF/ICSI treatment, but also with the risk of subfertility and natural fecundity [15, 16, 17, 18, 19]. The underlying mechanisms are largely unknown. Within this body of evidence the current study aims to identify a habitual diet of women undergoing IVF/ICSI treatment using two different COH protocols that influences the E2 response to COH and clinical reproductive outcomes. In future, the results may contribute to improve the quality as well as the further personalization of subfertility treatment, in particular IVF/ICSI treatment.

Materials and Methods

The Food, Lifestyle and Fertility Outcome study was designed as a periconceptional cohort to investigate the relationship between nutrition, lifestyle and IVF/ICSI treatment outcomes in which a randomized controlled trial was embedded to study a mild and conventional COH protocol. This study has been described in detail before [17, 18, 19, 20]. Participation in the RCT was based on the patient’s willingness to undergo randomization to either a mild or conventional stimulation protocol or additionally, a BMI >29 kg/m2 precluded participation in the randomized controlled trial, as opposed to >37 kg/m2 in the observational study. After inclusion and before treatment commenced participants received a self-administered questionnaire addressing age, educational level, medical history, BMI, ethnicity, medication use, smoking, and use of folic acid and other vitamin supplements. In addition, all participants completed a Food Frequency Questionnaire (FFQ) to estimate food intake covering the previous 4 weeks. The FFQ was developed by the division of Human Nutrition, Wageningen University, and validated for intake of energy, B-vitamins, and mono- and polyunsaturated fatty acids [21]. The FFQ was provided on the day of oocyte retrieval or semen sample collection and returned on the day of embryo transfer. The Dutch Central Committee for Human Research and the medical ethical and institutional review board of the Erasmus MC, University Medical Centre in Rotterdam approved the study protocol. Participants provided written informed consent before participation and the obtained materials and questionnaires were processed anonymously.

Results

We identified three habitual dietary patterns (Table S1 for details) and further explored associations with E2 response using linear regression analysis. In contrast to the first two dietary patterns, the third dietary pattern significantly associates with the E2 response and was therefore further analyzed (r=-0.16; p=0.02). This diet is characterized by high intake of fruits, whole grains, margarine, and low intake of refined grains and snacks; it explained 8.9% of variance found in dietary intake and is further named the habitual Healthy dietary pattern. To compare baseline characteristics of women over the range of scores of adherence to the habitual Healthy dietary pattern, we divided the cohort into three equally sized groups of adherence to the habitual Healthy dietary pattern, i.e. low, medium and high. Table 1 depicts baseline characteristics of women undergoing IVF/ICSI treatment. In the high adherence group, a larger proportion of women reported the use of a folic acid containing (multi)vitamin supplement. However, the exclusion of women who did not use a vitamin supplement from the forthcoming regression analyses did not alter the results (see further). Other baseline characteristics show a comparable distribution over the tertiles of adherence (Table 1).

Table S1: Food group loadings for eligible habitual dietary patterns.

Table 1: Baseline characteristics of women undergoing controlled ovarian hyperstimulation, stratified for adherence to the habitual Healthy dietary pattern (n=203). Variables are depicted: Geometric Mean (Interquartile Range), unless stated otherwise. Table 2 shows biomarkers measured in serum prior to COH on CD2 and after stimulation treatment on the day of hCG administration. First, although not significant, baseline E2 and AMH concentrations are lower in the tertile of highest adherence (Table 2). In addition, serum concentrations of pyridoxine (PLP) and RBC folate concentrations are higher in the highest tertile (Table 2). After COH, the differences are comparable with PLP and RBC folate concentrations being higher in the highest tertile of adherence. Furthermore, AMH and E2 concentrations are lowest in the highest tertile (Table 2).

Albeit there was a statistically significant increase in E2 on the day of hCG, this increase was not statistically different between the three strata of adherence to the Healthy dietary pattern. However, when data were adjusted for total FSH dose, age, BMI, number of visualized follicles and baseline AMH concentrations, a non-linear (quadratic) association between adherence to the habitual Healthy dietary pattern and hCG-day E2 concentrations (r-model=0.73; βadherence=-0.11 [-0.18 - - 0.04]; p<0.01 and βadherence, squared =-0.054 [-0.09 - -0.017]; p<0.01) could be established. Thus, a one-point increase in adherence to the habitual Healthy dietary pattern lowers hCG-day E2 concentrations by 16.4%.

Table 2: Biochemical markers of women undergoing controlled ovarian hyperstimulation, stratified for adherence to the habitual Healthy dietary pattern (n=203). Variables are depicted: Geometric Mean (Interquartile Range) Table 3 shows the clinical reproductive outcome after IVF/ICSI treatment. There are no statistically significant differences between the groups. Although the fertilization rates appear different, this is likely an artifact of the fact that a binomial variable is created using a denominator and numerator with a skewed distribution. The number of suitable embryos is comparable. In the multivariable analyses, adjusted for the number of transferred embryo’s, age of the woman, BMI of the woman, treatment indication, fertilization rate and woman smoking status, we show a significant association between adherence to the habitual Healthy dietary pattern and the subsequent chance of ongoing pregnancy (ORadj: 1.61 [1.07 – 2.44]; p=0.02).

Discussion

This study shows that strong adherence to a periconceptional habitual Healthy dietary pattern characterized by high intake of fruit, whole grains and margarine and low intake of refined grains and snacks associates with lower E2 concentrations on the day of hCG and a higher chance of ongoing pregnancy after IVF/ICSI treatment. The results from this study should be considered within the context of its strengths and limitations. The FOLFO study is a prospective periconception cohort designed to address the influence of nutrition and lifestyle on IVF/ICSI treatment outcome. In this study we used all participants of both the observational and randomized controlled trial, besides overweight status and patient preference there where no differences in inclusion criteria in either study, additionally, all analyses where adjusted for BMI and stimulation protocol. The habitual dietary patterns obtained using PCA from FFQ data result in sensible food-item based outcomes that are easily translated into patient care. However, it is important to consider the tendency of the individual to over represent socially desirable food habits and generally underestimate their overall food intake. Also, food habits are culturally determined, sometimes requiring an adapted FFQ for different ethnicities [23], in this study however, excluding non-western ethnicities did not alter the result. Furthermore, recall bias when completing FFQ is not an issue of concern because of the prospective design in which the outcome of the study was not known at the moment of the FFQ data collection. Even though PCA tends to emphasize sample specific factors; the high accordance between the various studies in the generated dietary patterns that are considered healthy and their associations with treatment outcome is a testament to the external validity of this method [16, 24]. Controlling the E2 response to COH using a modifiable factor, such as nutrition, could help to further personalize IVF/ICSI treatment and possibly increase the chance of pregnancy. In our study, however, the ovarian response itself was not a significant predictor of the chance of ongoing pregnancy. The role of high hCG-day E2 concentrations on clinical IVF/ICSI treatment outcome remains much debated, with an uncertain influence [6, 7, 8, 9]. However, the effect of a high average number of transferred embryos (≥3) in these studies is likely to dilute or distort any effects of the E2 response on these clinical outcomes. In addition, in vitro studies clearly show that high E2 concentrations associate with endometrial gene expression profiles that are considered not beneficial for implantation [3, 25]. Possibly the influence of the ovarian response on IVF/ICSI treatment outcomes should not be solely due to the effects of E2 on the endometrium, but also due to the influence of COH and the ovarian response (of which E2 is a marker) on oocyte and follicular developmental competency, by influencing follicular gene-expression profiles [10, 11], but also because the cytochrome P450 enzyme reaction cycles (used for steroid synthesis) results in a significant oxidative load [12] and thereby possibly oxidative stress. In our study, the chance of ongoing pregnancy was highest in women with strong adherence to the Healthy dietary pattern, which is in accordance with a previous study in the same cohort [16]. Unfortunately, we were not able to study the influence on live birth, of ultimate relevance for couples undergoing IVF/ICSI treatment. The beneficial effects of reduced oxidative stress and folic acid supplement use on embryo and placental quality could lead to beneficial reproductive outcomes, possibly without affecting the ovarian response [15, 18, 26]. This is supported by the finding that nutrition also affects the chance of pregnancy in couples not undergoing IVF/ICSI treatment [24]. Therefore, improved pregnancy rates and a lower E2 response can pose two separate effects of nutrition. Further studies need to address whether the association between nutrition, the ovarian response and improved pregnancy rates have a common causal pathway and if so; whether there is a cumulative effect of nutrition on IVF/ICSI treatment outcomes by affecting both embryo quality and the E2 response. Several studies show beneficial effects of multivitamin supplement use including folic acid on IVF/ICSI treatment outcomes. In our study women with the highest adherence to the habitual Healthy dietary pattern were most likely to report multivitamin supplement use. This indeed raises the question whether a Healthy diet or multivitamin supplement use determines the outcomes? Nevertheless, including multivitamin supplement use in the analyses or excluding all nonusers did not influence the significant effects of nutrition on the ovarian response or chance of pregnancy. This suggests indeed that the observed associations are mediated through the habitual Healthy diet. Earlier studies allow for speculation on possible mechanisms through which nutrition affects the ovarian response. Previously we showed that folic acid supplement use lowered the ovarian response to COH comparable to a diet high in fruit and whole grains characterized by higher folate concentrations in the current study [17]. Additionally, Kanakkaparambil, et al. similarly showed mechanistic insight into an increased E2 response in ewes who received an opposite poor methyl- deficient diet, through up regulation of the Follicle Stimulating Hormone-Receptor (FSHR) as a consequence of excessive oxidative stress [27]. Physiological up regulation of FSHR may also induce an anti-oxidant response through up regulation of glutathione [28, 29]. Also, many polymorphisms in folate metabolism have been identified, which have been shown to influence ovarian metabolism [30]. However, since there is no known influence of these polymorphisms on nutritional habits, it is unlikely that these polymorphisms confound our findings (i.e. mendelian randomization). It is also interesting to note that at baseline women with high adherence to the habitual Healthy dietary pattern have lower AMH levels, a predictor for ovarian reserve and ovarian response. It would be of interest to confirm the hypothesis that a healthy dietary pattern would lead to a lower antral follicle count and thereby a lower ovarian response. Nevertheless, we suggest that the beneficial effects of the habitual Healthy dietary pattern could be mediated by the accompanied higher folate concentrations. The anti-oxidant capability of folate and the high anti-oxidant content of fruits propose a mechanism through which both can independently affect the E2 response through reducing oxidative stress, notwithstanding the fact that a healthy dietary pattern is a prerequisite of adequate folate levels.

Conclusion

We show that strong adherence to a periconceptional habitual Healthy dietary pattern decreases the E2 response to COH, while increasing the subsequent chance of ongoing pregnancy after IVF/ICSI treatment. These findings support the rationale and necessity for initiation of the use of a healthy diet in women planning pregnancy, but also legitimize preconception counseling with regard to the influence of nutrition on IVF/ICSI treatment outcomes. Furthermore, this finding provides room for a more personalized approach of COH treatment. It is unclear whether the decreased E2 response or improved embryo quality through better nutrition is causal for the higher pregnancy rates, possibly it is both. In vitro research of granulosa cells could therefore help pinpoint underlying mechanisms. The anti-oxidant capacities of high fruit consumption may explain the effects on both the E2 response and improved pregnancy rates and is in accordance with previous evidence.

Acknowledgements

The authors gratefully acknowledge the contribution of Dr. Jolanda Boxmeer and Professor Nick Macklon (Department of Obstetrics & Gynecology, University of Southampton) to the FOLFO Study. We also thank the IVF team of the Division of Reproductive Medicine for the sample collection and the laboratory technicians and research assistants of the Laboratories of Clinical Chemistry and Endocrinology of the Erasmus MC, University Medical Centre in Rotterdam, The Netherlands, for the laboratory determinations.

Contribution to Authorship

JMT analyzed the data and wrote the manuscript. MV analyzed the data and provided analysis tools. JHdV provided essential materials and tools. JL provided reagents and analysis tools. JSEL was responsible for the clinical care of the patients and contributed to the writing of the manuscript. RPMST conceived the study, supervised all aspects and contributed to all versions of the manuscript.