DATA MINING AND MODELLING THE IMPACT OF IN VITRO FERTILISATION ON OFFSPRING HEALTH OUTCOMES

Abstract

In vitro fertilisation (IVF) conceived offspring are at increased risk of adverse birth outcomes such as low birth weight (LBW), small for gestational age (SGA), and preterm birth (PTB) when compared to spontaneously conceived offspring. However, few causal links between specific IVF procedures and birth outcomes have been robustly established. According to the Developmental Origins of Health and Disease (DOHaD) hypothesis, environmental influences during periconception have the potential to programme fetal development in a way that is determinant of long-term health and disease risk. IVF is an exogenous intervention in human reproduction during this critical developmental window of life. Considering that IVF conceived children now account for >2% of UK births annually, the potential impact of IVF treatment on future adult disease risk at the population level is considerable. Investigating key links between IVF treatment parameters and differences in birth outcomes may aid in identifying high-risk factors which could be altered to reduce said risk. We hypothesised that variation in IVF treatment procedures are associated with differences in offspring birth weight (BW) outcomes, and tested this by: 1) investigating how changes in IVF treatment practices and laboratory conditions relate to BW trends over time, and 2) developing bespoke means of augmenting existing data sources, including those held by the Human Fertilisation and Embryology Authority (HFEA), to investigate associations between IVF treatment parameters and differences in BW. We explored temporal BW changes in a single-centre dataset comprising almost 25 years (1991-2015) of routinely recorded data submitted to the HFEA, and found that gestation-adjusted singleton BW increased by 300g in total or 12.6g/year on average (95% CI: 4.5g - 20.7g). Frozen/thawed embryo transfer (FET) conceived singletons were born 52g heavier on average vs. their fresh ET conceived counterparts (95% CI: 1.9g-102.0g). We then explored associations between IVF parameters and birth outcomes using the results of an investigation linking a national survey of embryo culture systems to HFEA data, and found culture medium, among other treatment and patient prognostic factors, to be associated with odds of live birth (p=0.045). Incubator type was associated with gestational age (GA) (p=0.037). Blastocyst (vs. early cleavage) stage culture was associated with increased odds of live birth (OR=1.39, 95% CI: 1.32-1.45), lower GA (-0.15 weeks, 95% CI: -0.23 to -0.06), increased risk of preterm birth (OR=1.43, 95% CI: 1.24-1.66), and higher GA-adjusted BW (36.5g, 95% CI: 16.2g-56.9g). The impact of exposure to the embryo culture system on IVF success and subsequent birth outcomes is robustly shown. However, confounding between embryo culture system changes and clinic may have masked other treatment effects. Findings from a multicentre cross-sectional study covering seven years (2007-2014), confirmed our single-site finding of increased BW among FETs (92g, 95% CI: 57g-127.6g) after accounting for IVF treatment, embryo culture system, and patient characteristic factors. This difference increased to 180g (95% CI: 60.1g-298.8g) when limited to a sibling analysis in which patient factors were held constant. This series of studies shows that BW outcomes differ with respect to variation in IVF treatment practices and embryo culture system exposures.

Date of Award	1 Aug 2018
Original language	English
Awarding Institution	The University of Manchester
Supervisor	Stephen Roberts (Supervisor), Edward Johnstone (Supervisor) & Daniel Brison (Supervisor)