Specific lifestyle factors and in vitro fertilization outcomes in Romanian women: a pilot study

Study sample, questionnaire, and clinical protocol

We enrolled 35 women who completed a first IVF cycle within the Assisted Reproduction Department at the 1^st Obstetrics and Gynecology Clinic in Cluj-Napoca, Romania, between April 8^th and October 1^st, 2019. All study participants provided written informed consent prior to participation in our study, and the research protocol was approved by the “Iuliu Hatieganu” University of Medicine and Pharmacy Ethics Committee (Institutional Review Board (IRB) approval number 51, March 11^th, 2019). We approached 37 women in total and had two participation refusals (94.6% participation rate). All study participants completed a physician-administered questionnaire. The questionnaire included 47 questions organized in six sections: demographic data (e.g., name, address, and date of birth); socioeconomic data (e.g., last graduated school); home and employment data (e.g., questions on potential exposure to xenobiotics at home and in the workplace); health status (e.g., medical and reproductive histories); lifestyle factors (e.g., smoking, physical activity, use of PCPs, and dietary habits) and questions about the male partner (e.g., age and medical history).

We abstracted clinical data from the participant medical file on antral follicle count (AFC) as an indicator of ovarian reserve, levels of hormones (follicle stimulating hormone (FSH), luteinizing hormone (LH), anti-Mullerian hormone (AMH), and estradiol), response to ovarian stimulation (peak estradiol, number of oocytes retrieved, and thickness of the endometrial mucosa), and IVF endpoints (oocytes fertilized, number and quality of embryos, chemical pregnancy, clinical pregnancy, and live births).

Urine, blood, ovarian follicular fluid, and endometrial flushing fluid specimens were collected at the time of oocyte retrieval, for future metal and genetic analysis. All biospecimens were immediately processed after collection (centrifugation for blood serum separation) and stored at either −20 °C for future analysis of metals, or −80 °C for future genetic analyses.

Patients underwent controlled ovarian stimulation following a baseline infertility evaluation. They were assigned to a clinical protocol according to individual phenotype and infertility characteristics, reproductive history, and patient preference. Either a long protocol (ovarian down regulation with agonist gonadotropin-releasing hormone (GnRH)) or antagonist GnRH protocol was implemented. The starting gonadotropin doses (Menopur (Menotropin) or Follitropin-α (Gonal-F)) were tailored for patient age, AMH level, number of small antral follicles, body mass index (BMI), and reproductive history. Ovarian stimulation was monitored by transvaginal ultrasound and serum estradiol levels, beginning 4/5 days after initiation and the gonadotropin doses were adjusted according to the ovarian response. For final oocyte maturation, hCG (choriogonadotropin-α (Ovidrel)) was administrated when more than three ovarian follicles exceeding 17 mm in diameter were detected by transvaginal ultrasound. Oocytes were retrieved by transvaginal follicle puncture 34–38 h after hCG administration. Collected oocytes in metaphase II arrest were fertilized by intracytoplasmic sperm injection (ICSI), or conventional insemination using fresh sperm from the male partner, 4–6 h after retrieval. Couples with male pathology (<2 million motile spermatozoa) or (≤4% normal morphology) received ICSI. Oocytes were evaluated 14–16 h after injection, and were graded as either normally fertilized (two pronuclei) or abnormally fertilized (single or poly pronuclear). Resulting embryo quality was assessed using the Istanbul consensus scoring system: I (best), II (good), III–IV (poor), and V (worst) (Balaban et al., 2011; American Society for Reproductive Medicine (ASRM) Practice Committee & Society for Assisted Reproductive Technology Practice Committee, 2013). Three or five days after the oocyte retrieval, one to two embryos were transferred to the patient’s uterus, considering clinical factors and patient preference. The remaining embryos were frozen for future use. Lutinus 100 mg was administrated three times a day intravaginally, as support for the luteal phase. A pregnancy was biochemically confirmed by serum beta hCG measurement (beta hCG > 20 mIU/ml) 14 days after the oocyte retrieval. An ultrasound examination confirmed a “clinical pregnancy” 2 weeks later, by visualization of one or more gestational sacs (Zegers-Hochschild et al., 2009). Patients’ obstetricians monitored the pregnancy and reported either spontaneous abortion or live birth.

Statistical analysis

We calculated the fertilization proportion per woman (%) as the ratio between the number of fertilized oocytes and the number of retrieved oocytes, and the proportion of high-quality embryos per woman (%), as the ratio between the number of embryos (grades I and II) and the number of retrieved oocytes.

We first examined associations between 17 individual lifestyle factors and IVF outcomes among oocytes (n = 194) and embryos (n = 79). We operationalized responses to a physician-administered questionnaire as ordinal variables (e.g., 1 for never to 5 for every day) to describe the daily to monthly frequencies of 17 lifestyle habits and the intensity of behaviors, including exposure to tobacco smoke and psychosocial stress, physical activity, use of PCPs, and dietary exposures. We then used principal component analysis (PCA) to describe the variability in patterns of women’s lifestyle habits and behaviors. Based on their covariances, PCA summarizes a larger number of variables as a smaller number of independent “summary variables”, factors that represent latent constructs (Yong & Pearce, 2013). We employed a varimax rotated PCA, based on a polychoric correlation matrix, to summarize 17 individual variables describing various lifestyle habits and behaviors (Grace-Martin, 2019). We selected two PCA factors summarizing women’s use of PCPs, dietary habits and physical activity based on a scree plot, eigenvalues >2.0, and >10% variance explained by each factor (DiStefano, Zhu & Mîndrilǎ, 2009; Yong & Pearce, 2013). We only retained lifestyle habits and behaviors factor loadings >|0.5| in each PCA factor to ensure moderate-strong correlations to the summary PCA factors. We then multiplied each participant’s value by the retained factor loading to calculate weighted sum scores for each summary PCA factor.

We used the two retained PCA factors as simultaneous predictors of IVF outcomes in multivariable regression models to estimate associations of lifestyle habits and behavior patterns with IVF outcomes, adjusted for age in years (Polanska et al., 2014), BMI in kg/m² (Sarais et al., 2016), current smoking status (Domar et al., 2012), and education as a measure of socioeconomic status (Swift & Liu, 2014), which were selected a priori as confounders based on the literature. We used Poisson regression with robust error variance employing generalized estimating equations (GEE) to estimate the relative risks (RR) and 95% confidence intervals (CI) for dichotomized IVF outcomes including oocyte fertilization, embryo quality, clinical pregnancy, and live birth as outcomes in relation to women’s lifestyle habits and behavior patterns. The GEE accounted for “clustered” outcomes within each woman (Zou, 2004). Using negative binomial regression, we also calculated effect estimates for associations between baseline AFC and women’s lifestyle habits and behavior patterns. We used linear regression to estimate the association between continuous baseline AMH levels (ng/mL), endometrial thickness (mm), and peak estradiol levels (pg/mL) as intermediate IVF outcomes with women’s lifestyle habits and behavior patterns. We checked for and excluded influential observations defined as Dfbeta >|1.96|. We set statistical significance as α = 0.05 for a 2-tailed test. Statistical analysis was performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

IVF pregnancy and live birth rates in Romania and other European countries

In 2016, The European IVF Monitoring Consortium (EIM) reported that the average IVF clinical pregnancy rate was 28.0% per oocyte retrieval, varying between 13.2–57.1% among countries, while the average ICSI clinical pregnancy rate was 25.0% per oocyte retrieval, varying between 18.7–42.6% among different countries (European IVF-monitoring Consortium (EIM) for the European Society of Human Reproduction and Embryology (ESHRE) et al., 2020). Romania initiated 3,479 IVF and ICSI cycles in 2016, with clinical pregnancy rates of 31.6% and 28.3% per oocyte retrieval, respectively, similar to those reported for other EU countries (e.g., IVF pregnancy – France 21.1%, Italy 21.6, Denmark 20.9%, Germany 28.8%, Belgium 27.8%, Ireland 40.4%, Hungary 31.4%, the Netherlands 31.0%, and the UK 32.8%; ICSI pregnancy – Czech Republic 22.5%, Bulgaria 26.3%, Portugal 23.0%, Sweden 27.4%, Ireland 38.8%, Iceland 30.2%, Poland 25.8%, the Netherlands 32.9%, and the UK 33.1%). In 2016, Romania had 23.4% live deliveries per oocyte retrieval, similar to or greater than in, Poland (23.1%), Iceland (23.1%), the Netherlands (22.2%), Belgium (20.3%), Germany (19.4%), and France (18.3%), but lower than in Norway (24.2%), Sweden (24.3%), and the UK (28.5%) (European IVF-monitoring Consortium (EIM) for the European Society of Human Reproduction and Embryology (ESHRE) et al., 2020).

Healthy dietary habits, physical activity, and IVF outcomes

Food products in Romania are purchased mostly from supermarkets that mainly sell products imported from other European countries (e.g., Spain, Italy, Greece, Poland, and Turkey). As regards the physical activity, a review of data collected from 38 European countries showed similar mean weekly durations of moderate to vigorous physical activity in Romania (599 min/week), Poland (599 min/week), and Hungary (593 min/week), which was less than in Bulgaria (675 min/week), Croatia (775 min/week), Germany (637 min/week), Greece (667 min/week), and the Netherlands (960 min/week), but greater than in Italy (212 min/week), France (259 min/week), Spain (357 min/week), the UK (259 min/week), Denmark (468 min/week), and Austria (499 min/week) (Loyen et al., 2016).

Previous studies indicated a potential relationship between diet and infertility treatment outcomes, although the results have been mixed (Sanderman, Willis & Wise, 2022). A diet rich in unsaturated fats, vegetables, fish, and whole grains was associated with a positive impact, whereas a diet rich in saturated fats and sugar had a negative impact on fertility outcomes in both women and men (Gaskins & Chavarro, 2018). Fish intake, likely due to its omega 3 polyunsaturated fatty acids content, was associated with a greater likelihood of blastocyst formation (Braga et al., 2015), and also, with a greater likelihood of live birth in women undergoing IVF (Nassan et al., 2018). Vitamin D3 in fatty fish may also play an important part in female reproduction (Anagnostis, Karras & Goulis, 2013) by influencing follicle development (Irani & Merhi, 2014). However, the beneficial effects of fish consumption may be overshadowed by exposure to reproductive toxicants found in seafood, including mercury (Al-Saleh et al., 2008; Butts et al., 2021) and polychlorinated biphenyls (PCBs) (Meeker et al., 2011; Bloom et al., 2017). In this pilot study, we found that monthly fish consumption was associated with higher likelihoods of good quality embryos and live birth, albeit non-significant. Greater vegetable and fruit intake was associated with better embryo quality in a study of 269 Brazilian IVF patients (Braga et al., 2015). However, increased exposure to pesticide residues on vegetables and fruit has also been associated with a lower probability to achieve a clinical pregnancy and live birth from IVF in a U.S. study (Chiu et al., 2018). Here, we found that greater consumption of fruits and vegetables was associated with higher probabilities of pregnancy and live birth (with statistical significance for fruit consumption).

Regular exercise helps to prevent energy excess and heightens sensitivity to insulin, which is beneficial for reproductive function (Redman, 2006). Also, regular exercise programs and changes in lifestyle leading to weight loss in overweight and obese women have a positive impact on menstrual function, improving ovulation and subsequent fertility (Silvestris et al., 2018). Women’s physical activity before initiating IVF was associated with higher pregnancy and live birth rates in a meta-analysis of eight studies, and also with a small increase (not statistically significant) in the implantation rate (Rao, Zeng & Tang, 2018). Yet, results from a population health survey comprising 3,887 women under the age of 45 years indicated an association between robust exercise (daily exercise or exercise until extreme tiredness) and subfertility, although no associations were reported with reduced intensity exercise (Gudmundsdottir, Flanders & Augestad, 2009). Also, in a prospective cohort study including 2,232 women, at least 4 h per week of higher intensity exercise for a year or more preceding the IVF cycle was related to greater (2.8-fold) odds of cycle cancellation, a 2-fold increase in implantation failure, and a 40% decrease in live births after the first IVF cycle compared to women who did not exercise regularly (Morris et al., 2006). Our analysis of individual lifestyle habits and behaviors and clinical IVF outcomes showed that routine weekly exercise was associated with a non-significant lower likelihood of live birth. However, in the PCA analysis including physical activity together with healthy dietary habits as one principal component, we found that combined healthy diet and greater physical activity was associated with a non-significant higher likelihood of pregnancy, and significantly greater AMH and peak estradiol levels. Still, given our small pilot sample, a larger investigation is needed to more definitively estimate the impact of dietary habits and physical activity on clinical and intermediate IVF outcomes among Romanian women.