Postnatal persistence of fetal cardiovascular remodelling associated with assisted reproductive technologies: a cohort study

To assess the postnatal persistence of fetal cardiovascular remodelling associated with assisted reproductive technologies (ART) in children at 3 years of age.


Introduction
It is now estimated that more than 5.4 million babies have been conceived worldwide since the first in vitro fertilization (IVF) baby was born in 1978. 1 Thus, the potential health risks associated with these treatments are of great importance to public health. Although most of children conceived by assisted reproductive technologies (ART) are born healthy, concerns regarding long-term cardiometabolic health in ART children are increasing due to the accumulative evidence suggesting that ART may have lasting negative repercussions for the health of individuals conceived through these techniques. 2,3 Various studies have shown that children born after ART have poorer metabolic and vascular profiles than naturally conceived children, with elevated systolic and diastolic blood pressures, higher fasting glucose levels, increased central adiposity and vascular dysfunction with pulmonary hypertension. [4][5][6][7][8] According to those authors, the changes mentioned seems to be independent of prematurity, birthweight and parental characteristics, suggesting that features related to ART could directly contribute to these adverse health effects. 9 Recently, cardiac remodelling in utero that persists into early postnatal life was demonstrated in ART offspring. 10 In that study, fetal echocardiography showed signs of cardiac remodelling and dysfunction in ART fetuses mainly in the form of larger atria, shorter ventricles, thicker myocardial walls, reduced longitudinal motion and impaired relaxation. A postnatal follow up of the same cohort at 6 months of age showed not only postnatal persistence of cardiac remodelling but also increased blood pressure and vascular wall thickness. 10 However, cardiovascular outcomes in later stages of life are still controversial, with studies reporting changes in left ventricular function 11,12 and others suggesting right ventricular dysfunction that is only evident under stressful conditions. 13 The aim of the present study was to evaluate the persistence, into childhood, of cardiac findings observed in ART fetuses. For this purpose, we conducted a cardiovascular follow-up study from our prenatal ART cohort into early childhood.

Study population and protocol
A prospective cohort study including 80 children conceived by ART and 80 spontaneously conceived (controls) recruited from fetal life and followed up to 3 years of age (see Supporting Information for details of the prenatal 10 and follow-up cohorts). Preimplantation genetic diagnosis, oocyte donation, multiple pregnancies, any maternal medical disease, fetal malformations, chromosomal anomalies or any pregnancy complications leading to delivery before 34 weeks of gestation were considered exclusion criteria. The study protocol was approved by the Institutional Review Board at Hospital Clinic, and written parental consent was obtained from all study participants.
The child's follow-up evaluation was scheduled at 3 years of corrected age including anthropometric data, echocardiography and vascular assessment. Examiners were blinded to conception type. Anthropometric data included the child's height, weight and body mass index measured at the time of the examination.

Echocardiography
Echocardiography was performed following a standardised protocol 14 using a Vivid q (General Electric Healthcare, Horten, Norway) with 2-10 MHz phased-array transducer. Children were studied when resting quietly. A complete two-dimensional M-mode and Doppler echocardiographic examination was performed initially to assess structural heart integrity and morphometry. Left and right atrial planimetric areas were measured on a 2D image from an apical four-chamber view at end-systole (greatest dimension, just before mitral or tricuspid valve opening). Ventricular base-to-apex length and transverse diameter were measured on a 2D image from an apical four-chamber view at end-diastole. Left and right ventricular sphericity indexes were calculated as base-to-apex length/mid-transverse diameter, Ventricular end-diastolic septal and lateral free wall thicknesses were measured by M-mode from a parasternal long-axis view. 14,15 Systolic function of both ventricles was evaluated using shortening fraction, cardiac output, tricuspid and mitral annular plane systolic excursion (TAPSE and MAPSE) and annular systolic peak velocities (S'). 16 Left shortening fraction was calculated from internal ventricular diameters obtained from a parasternal long-axis view by M-mode, using the equation: (end-diastolic dimensionend-systolic dimension)/end-diastolic dimension. Left and right stroke volumes were calculated as p/4 9 (aortic or pulmonary valve diameter) 2 9 (aortic or pulmonary artery systolic flow velocity-time integral). Left and right cardiac outputs were calculated as stroke volume 9 heart rate. Diameters of the aortic and pulmonary valves were measured in frozen real-time images during early to mid-systole by the leading-edge-to-edge method; aortic diameter was obtained from the parasternal long-axis view, and the pulmonary artery diameter was obtained in a parasternal short-axis view. 16 Ascending aorta flow velocity integral was measured with pulsed Doppler from an apical five-chamber view, and the pulmonary artery flow velocity integral was recorded from a standard parasternal short-axis view with the sample volume placed immediately distal to the pulmonary valve. Velocity-time integrals were calculated by manual trace of the spectral Doppler area. TAPSE and MAPSE were measured real time in an apical four-chamber view by placing the M-mode cursor at the atrioventricular junction, marked by the tricuspid valve rings at the right free wall. Maximum amplitude of motion was taken as the extent of displacement between end-systole and end-diastole, and measured in millimetres. Tissue Doppler was applied at tricuspid and mitral lateral annuli from an apical fourchamber view, to record S' in centimetres/second. 16 Diastolic function of both ventricles was evaluated by atrioventricular peak velocities at early diastole and atrial contraction (E/A ratios), E deceleration time, diastolic annular peak velocity (E') and left isovolumic relaxation time (IRT). Atrioventricular flow were obtained from an apical four-chamber view, placing the pulsed Doppler sample volume just below the valve leaflets. E deceleration time was measured as the time from the maximum mitral/tricuspid velocity to the baseline. Tissue Doppler was applied at tricuspid and mitral lateral annuli from an apical fourchamber view to obtain E'. Left IRT was obtained from the pulsed Doppler waveform of the aortic blood flow, from the end of the aortic wave to the beginning of the mitral early filling wave.

Vascular assessment
Vascular assessment included blood pressure and carotid wall thickness by ultrasound.
Systolic and diastolic blood pressures were obtained at the beginning of the medical evaluation from the brachial artery using a validated ambulatory automated Omron 5 Series device, while the child was resting. Mean blood pressure was calculated as ((2 x diastolic)+systolic)/3.
Carotid ultrasound assessment was performed by a skilled sonographer using a Vivid q (General Electric Healthcare). Longitudinal clips of the far wall of both carotid arteries were obtained approximately 1 cm proximal to the bifurcation using a 3.33-10.0 MHz linear-array transducer. Carotid intima-media thickness (IMT) was measured offline according to a standardised protocol based on a trace method with the assistance of a computerised programme ECHOPAC SOFTWARE ONLY. To obtain IMT, three end-diastolic still frames were selected across a length of 10 mm and analyzed for mean and maximum IMT, and the average reading from these three frames then calculated. 17

Baseline and perinatal characteristics
Baseline and perinatal characteristics of the study are shown in Table 1. Parental baseline characteristics were similar among the study groups, with the exception of older parental age and higher rate of nulliparity in the ART group as compared with spontaneously conceived ones. As expected, the ART group showed a worse perinatal outcome with earlier gestational age at delivery, lower birthweight and birthweight centile as compared with controls. There was also a nonsignificant trend for higher prevalence of pre-eclampsia, gestational diabetes, prenatal corticoid exposure, caesarean section and admission to neonatal intensive care unit in the ART group as compared with controls.

Study protocol in early childhood
Anthropometric and cardiovascular results at 3 years of age are shown in Table 2. Both groups showed similar age and anthropometric characteristics at evaluation. ART children showed a larger right atrium together with more spherical ventricles and similar myocardial thickness as compared with spontaneously conceived children. While cardiac output and tissue Doppler values were similar among groups, ART children showed decreased shortening fraction and mitral/tricuspid ring displacement together with prolonged IRT. In addition, blood pressure and carotid IMT were significantly higher in the ART group than in the group of spontaneously conceived children ( Figure 1). Most cardiovascular changes remained statistically significant even after adjustment for potential confounding factors such as parental age, gestational age at delivery and birthweight centile.

Main findings
This study demonstrates the persistence of cardiovascular changes associated to ART from fetal life to early childhood, supporting the concept of primary cardiovascular programming in ART offspring. Children conceived by ART showed signs of cardiac remodelling such as a larger right atrium and more spherical ventricles, as well as cardiac dysfunction demonstrated by decreased ring displacement and prolonged relaxation time. These changes are consistent with our previous report in the same cohort during fetal life, which also demonstrated larger atria, more globular ventricles, and signs of systolic and diastolic biventricular dysfunction. 10

Strength and limitations
Our major strength is the study design of a well-phenotyped cohort recruited from conception up to childhood including complete fetal, infant and child echocardiographic data. Original prenatal and follow-up populations showed similar baseline and perinatal characteristics (Tables S1 and S2). 10 As limitations, we acknowledge that the present study might be underpowered, as sample size was not estimated at this evaluation in childhood and there was a 20% loss in follow up. Detailed information regarding the population loss in follow up is shown in Appendix S1. We also acknowledge the presence of potential underlying confounding factors that are almost impossible to rule out due to the intrinsic baseline (such as parental infertility per se, socio-economic status, nutrition and lifestyle habits or maternal comorbidities) and perinatal factors (mainly coexistence of prematurity or placental disease) usually associated to ART pregnancies. For example, in our study we observed differences in parental age, parity, gestational age and weight at birth; statistical analysis was adjusted for these potential confounders, although this correction may be suboptimal for comparing cardiac characteristics. The mechanisms driving these changes in ART children remain to be elucidated; confounding factors such as advanced maternal age are well-known contributors to adverse pregnancy outcomes, 18 including low birthweight, a condition that has a high prevalence in the ART population and is also directly related with fetal cardiovascular programming. 19,20 Future studies including subfertile couples conceiving spontaneously could help better control for confounders and provide new insights into the mechanisms underlying cardiovascular changes observed in ART offspring.

Interpretation
Our results are also in line with other recent studies suggesting cardiovascular differences in childhood or adolescence of subjects conceived by ART. [11][12][13] Zhou et al. 11 reported signs of left ventricular hypertrophy and dysfunction in ART children at 2-6 years of age. Liu et al. 12 reported left ventricular reduced motion and diastolic dysfunction in ART children at 5 years of age, together with a nonsignificant trend for more spherical ventricles. Von Arx et al. 13 demonstrated right ventricular dysfunction under stressful conditions of high-altitude exposure in ART preadolescents who were born at term with a normal birthweight. Although most studies report significant cardiovascular changes associated to ART, the pattern of cardiac remodelling and dysfunction differs among populations. We report dilated atria and more spherical and less efficient ventricles (both left and right), Zhou et al. We hypothesise that cardiac changes observed in ART offspring are mainly secondary to increased vascular stiffness leading to increased cardiac pressure that needs to be compensated by adapting ventricular shape (more spherical and/or more hypertrophic) and dilating the atria. Differential cardiac patterns could be observed depending on the age at assessment: more evident and predominant right changes in fetal life evolved to less prominent changes as the myocardium becomes more mature and compliant and pulmonary pressures decrease or are shifted to the predominant left ventricle in postnatal life. Differences among studies could also be explained by heterogeneity in the factors potentially involved in the pathophysiological origins of these findings: (1) parental parameters and the cause and severity of infertility; (2) ovarian stimulation and its effects on the oocyte and endometrium; (3) ART laboratory procedures (manipulation of gametes and embryos, culture conditions, transfer at the blastocyst stage, vitrification techniques); (4) maternal environment and perinatal comorbidities (such as prematurity and fetal growth restriction). 21 Improvements and changes made to ART techniques over the time, could also be the possible cause of the observed differences in initial studies that have not been reproducible or that vary widely.
Regarding vasculature, we demonstrated increased blood pressure and carotid wall thickness in ART children compared with naturally conceived ones. These data is consistent with our vascular neonatal results in this cohort. 10 It is also in agreement with the previously reported systemic and pulmonary vascular dysfunction in late childhood. 8,22 However, results on systemic blood pressure in ART offspring are controversial. Dissimilarities among studies have been recently analysed in a meta-analysis including 872 IVF-ICSI offspring from 10 different studies. Those authors demonstrated an overall increase of 1.88 mmHg in systolic blood pressure among ART children as compared with those spontaneously conceived. Interestingly, blood pressure was statistically significantly higher in the ART children born in 1990-1999 than in those born in 2000-2009, suggesting an improvement over the years most likely due to changes and maturation of ART techniques. 23 The pathophysiology underlying the vascular impairment observed in ART remains unresolved, although experimental research suggests premature vascular ageing and arterial hypertension probably related to epigenetic alterations in vascular key factors such as the endothelial nitric oxide synthase gene. 24

Conclusions
With children conceived through ART now forming a sizeable subgroup of the population, further large clinical follow-up studies are required to establish the clinical relevance of evaluating cardiovascular health in those born after ART. The long-term consequences of cardiovascular changes observed in fetal life and childhood are not fully understood. Observations in adolescence suggest that cardiovascular dysfunction appears only under stress conditionssuch as high altitudeallowing us to speculate that ART programs individuals with subclinical changes more susceptible to developing disease under certain stressful circumstances later in life. 8,22 The ability to determine which individuals may be more susceptible to developing subsequent cardiac and/or vascular disease would be of great importance for public health strategies, hence, the clinical relevance of continuing long-term studies in this population. The challenge here is that ART techniques keep changing rapidly over time, evolving relevant improvements in ovarian stimulation protocols, oocyte and embryo vitrification techniques 25 and culture conditions. 26,27 Discrepancies among studies could be explained by improvements made to these techniques over time in order to achieve pregnancy, transferring the 'best' embryo in the optimal conditions. The real impact of these factors on the embryo phenotype (morphology, developmental kinetics, physiology and metabolism) still needs to be understood in both animal and human models. Therefore, large collaborative studies over time are warranted to address the impact on offspring health complexity of different ART techniques.
As a final conclusion, this and previous studies support the concept of primary cardiovascular programming associated to ART. Further research is warranted to determine the exact mechanisms underlying these cardiovascular changes and the potential long-term consequences on the health of these children and future adults.
From a clinical perspective and regardless of the need to clarify the specific mechanisms, the existence of fetal programming in these infants presents important opportunities to improve cardiovascular health in a relevant proportion of the population. Our study demonstrates significant changes in cardiovascular structure and function in ART children as compared with those naturally conceived. These changes are subclinical, with most cardiovascular indexes lying within normal ranges, explaining why most children are asymptomatic and without clinical signs of disease. However, subclinical changes in cardiovascular structure and function in the early stages of life may represent an underlying mechanism for increased cardiovascular risk later in life. In fact, some of cardiovascular differences, such as those reflected by significant increases in blood pressure and IMT, are recognised as potential risk factors for subsequent cardiovascular disease. Future studies are warranted to evaluate the long-term persistence of these cardiovascular changes throughout life and better to assess the potential risks associated to these findings.

Disclosure of interests
None declared. Completed disclosure of interests form available to view online as supporting information.

Contribution to authorship
BVA and FC were responsible for the conception, design and writing of the study. AS, AS-M and LG-O collaborated in database collection, quality control and patient recruitment. MR-L and MC-L also assisted with statistical analyses. GC, BB, MS and JB contributed as Senior investigators in the discussion and conclusions of the document. EG also contributed as Senior Investigator and approval of final document.

Funding
This project has been funded with support of the Instituto de Salud Carlos III ((PI14/00226, INT16/00168 and PI15/ 00130) integrated in the National R&D Plan and cofinanced by the ISCIII general evaluation subdirectorate, by the European Regional Development Fund (FEDER acronym in spanish) and by Erasmus Programme of the European Union (Framework Agreement number: 2013-0040)). Research leading to these results received funding from "La Caixa" Foundation; Cerebra, the Foundation for the Brain Injured Infant and Young People and the support of the Secretaria d'Universitats i Recerca del Departament d'Economia i Coneixement de la Generalitat de Catalunya (grant SGR n_ 928, 2013_FI0667, 2016FI_B01184).

Acknowledgements
We would like to thank all the parents and children for their commitment and effort when participating in our study. BVA would like to thank CONACyT (Consejo Nacional de Ciencia y Tecnolog ıa, Ciudad de M exico) for its initial support.

Supporting Information
Additional Supporting Information may be found in the online version of this article: Table S1. Baseline characteristics of controls and cases included in the fetal and follow-up studies. Table S2. Fertility characteristic and perinatal outcomes of controls and cases included in the fetal and follow-up cohorts.