The impact of assisted reproductive technology and ovulation induction on breech presentation: A whole of population‐based cohort study

Multiple studies have suggested a likely association between breech presentation and assisted reproductive technology (ART) for conception. The aims were to determine whether conception via in vitro fertilisation (IVF) and ovulation induction (OI) is associated with fetal malpresentation at birth and to ascertain what mediating factors most significantly contribute to fetal malpresentation.


INTRODUCTION
Increasing demand for assisted reproductive technology (ART) has swept through most modern societies since the first in vitro fertilisation (IVF) birth in 1978. There had been a fivefold increase in treatment cycles across Australia and New Zealand from 1991 to 2019, 1,2 and similar growth in Europe and the USA over this time. 3 Numerous ART options are now available, including IVF with or without intracytoplasmic sperm injection. In ART procedures, hormone induction through ovulation induction (OI) is an initial step used to increase the opportunity for egg retrieval and subsequent pregnancy. 4 However, OI can also be used independently to support conception.
The safety of ART and OI continues to be scrutinised. These techniques have an increased risk of adverse pregnancy outcomes, such as low birthweight, preterm birth, hypertensive disorders of pregnancy and caesarean section, 5-7 when compared with spontaneously conceived pregnancies.
Breech presentation at parturition is associated with a higher risk of labour-related complications and increased costs of providing care due to resource-intensive vaginal, planned or emergency caesarean section birth. 8 A growing body of evidence suggests that breech and other malpresentations occur more frequently in pregnancies conceived via ART compared with spontaneous pregnancies, 5,[9][10][11][12] though several potentially confounding factors obfuscate the degree to which this additional risk can be attributed to ART itself. Previous cohort studies have varied in their conclusions, with some finding breech to be significantly associated with ART, [13][14][15] whereas others have found the association to be attenuated after the adjustment for potential confounders. 16,17 Little data on an association between breech and OI specifically exist, with previous studies suggesting a weaker association than that seen between breech and ART. 5,13 As demand for ART and OI grows, it is critical to ensure patients and clinicians are informed of pregnancy risks associated with these techniques. Shifts to frozen cycles and single-embryo transfer have been designed to minimise risks associated with multiple pregnancies. 18 Nonetheless, it is vital to understand the impact of ART on birth presentation. Interrogating the pathogenesis of breech may facilitate further refinement of ART procedures through improved understanding of physiological concepts behind the differing outcomes in ART and OI compared to spontaneous pregnancies.
Australia has one of the highest rates of ART births in the world, 2 making it an ideal population to conduct ART-focused, population-based cohort studies. The aim of this study was to determine the level of association between breech or other malpresentations and conception via ART and OI. Additionally, we aimed to identify whether ART or OI is an independent predictor of malpresentation and the mediating factors contributing to breech presentation.

MATERIALS AND METHODS
A retrospective whole-population-based cohort study, restricted to singleton pregnancies, was undertaken to investigate whether malpresentations at birth occurred more frequently in pregnancies conceived via ART or OI compared to spontaneously conceived pregnancies during the period of 1 July 2012 to 30 June 2018.

Data source
This study used Maternity1000, a whole-of-population-linked administrative data set. 19,20 Maternity1000 comprises all women

Outcome definition
The Perinatal Data Collection identifies mode of conception for each pregnancy. We have specifically chosen IVF ± ICSI (intracytoplasmic sperm injection) and OI. Births following artificial insemination by husband or donor (AIH/AID) and OI were included in the OI group; births following AIH/AID were absorbed only into the spontaneous group. Other modes of conception available in the data were egg donation and gamete intra-fallopian transfer.
These were excluded from analysis due to low rates for these modalities (n = 135 and 65, respectively).
The Perinatal Data Collection also details the recorded presentation of each baby at birth. Birth presentations included in this study were cephalic, breech, transverse/shoulder and face/brow. Face and brow presentations were combined due to their clinical similarity and low counts when separately classified.

Statistical analysis
Women and children's demographical and clinical characteristics were described in terms of means and standard deviations, and frequencies and percentages (%). Logistic regression was used to calculate the odds ratio (OR, and 95% confidence interval (CI)) of different birth presentations relative to the different modes of conception. Adjusted ORs were also calculated, with a priori inclusion of potentially confounding variables known to impact modes of conception and presentation at birth in logistic regression models. These were birthweight, gestational age at birth, maternal age, parity, history of previous caesarean section, smoking status during pregnancy, presence of any congenital abnormality and gender. Birthweight and gestational age at birth underwent z-score transformation to standardise weight and age into common distributions for analyses. Backward stepwise regression was performed to identify which confounding variable(s) contributed to breech presentation at birth. This was conducted by first removing non-significant variables from the regression model, followed by variables with the highest coefficients (β-values). Data analysis was undertaken using SAS 9.4 (SAS Institute Inc., North Carolina, USA).

Ethics approval
This study received ethical approval from the Townsville Hospital

Descriptive characteristics
There were 365 138 births in Queensland, Australia, between No significant differences were present between the three conception groups in terms of gestational age at birth, gender or maternal body mass index (Table 1). Women were older in the ART group compared to those in the spontaneous and OI groups (35.0 ± 4.8 years vs 29.7 ± 5.7 and 31.7 ± 4.4 years, respectively). Parity, previous birth by caesarean section, smoking status and congenital anomalies were higher in the spontaneously conceived group ( Table 1). The most significant of these differences was smoking status; women who had conceived spontaneously were 5 to 10 times more likely to smoke before and after 20 weeks when compared with women who conceived via OI and ART.
ART and OI were also both significantly associated with de-  (Table 2).
Backward stepwise regression revealed that birthweight had the greatest effect on the odds of breech in both ART and OI groups, second only to maternal age in the ART group, as shown in Supporting Information, Table S1, and Figure 1. As birthweight increased, the odds of breech decreased in both ART and OI groups.
Additionally, as maternal age increased, the odds of breech increased in the ART group. Multiparity and gestational age were found to be protective.

Main findings
In this whole-population-based cohort study, ART and OI were associated with a significantly increased odds of breech presentation at birth. The adjusted OR for breech was observed to be similar between ART and OI pregnancies. Low birthweight and increasing maternal age were the largest positive contributing factors to breech in these pregnancies. No significant associations were observed between mode of conception and other malpresentations.
Previous cohort studies have found breech or malpresentation to occur more frequently in ART pregnancies compared with those conceived spontaneously, 5,9,13-15 a finding further confirmed in our study. Only two previous studies, however, included OI pregnancies. One found the association between OI and breech to be insignificant after adjustment 13 ; the other found it to be weaker than that between ART and breech. 5 In contrast, our study demonstrated similar significantly increased odds of breech with conception via both ART and OI over a large whole-populationbased cohort with appropriate separation of conception methods and malpresentations. Our findings strengthen the importance and significance of previous reports through a three-way study design directly comparing spontaneous pregnancy to both ART and OI. In addition, we simultaneously examined multiple types of birth presentation, rather than limiting scope to breech only or grouping malpresentations together to analyse as a single variable. As such, we additionally identified that no significant associations were observed between mode of conception and the odds of malpresentations other than breech.
Previous cohort studies on the topic have been restricted to singleton babies delivered at term 13 and either not adjusted for potential confounders 5,9,15 or found risks of breech to be fully attenuated after adjustment. 16,17 Of these, the study by Romundstad et al. 16

Clinical implications
At this time, the importance of recognising ART and OI as significant risk factors for breech could inform improved antenatal counselling and closer scrutiny regarding fetal presentation for women who conceived via these methods, particularly around 36 weeks' gestation. The absolute risk increases in breech from spontaneous pregnancy in these high-risk cohorts were found to be 1.6 and 2.7% for OI and ART, respectively. This is an im- Given the diversity and size of this study cohort, and the significance of the data, we suggest no delay or additional research is required before this information is integrated into current clinical practice.

Research implications
Based on our findings, there are several areas of research warranting further consideration.
First, the similarly increased associations with breech observed in the ART and OI groups suggest a shared underlying mechanism likely with the OI procedures, which are common to both pathways. 4  Additionally, economic modelling based on the POP study found that screening would be expected to virtually eliminate undiagnosed breech presentation and reduce fetal mortality and could be cost effective. 28 Given the present study's findings of associations between breech and conception via ART and OI, and that low birthweight was found to be the largest positive contributing factor F I G U R E 1 Backward stepwise regression representing the odds of breech presentation in women undergoing ART (assisted reproductive technology) and OI. ICSI, intracytoplasmic sperm injection; IVF, in vitro fertilisation; OI, ovulation induction.
to breech among these pregnancies, a study assessing the efficacy of a third-trimester scan for the detection of both birthweight and fetal presentation in this population group could be considered.

Strengths and limitations
The strengths of this study include the whole-population-based design, large sample size and diversity and robustness of the data, which granted the ability to adjust for several risk factors and calculate precise odds ratios. The natural diversity of the Australian population in terms of ethnicity, with more than a quarter of the population born overseas 29 ; maternal age and parity; and relatively high rates of IVF ± ICSI and OI lend this study a high degree of generalisability. This is the first contribution on this topic from Australia and includes pregnancies reflective of current ART practices due to the recent time period studied. In addition, the threeway study design comparing spontaneous pregnancy to both ART and OI allowed us to expand on the current literature regarding the impact of hormone induction processes utilised in both procedures on the likelihood of breech and demonstrate the presence of a likely shared underlying aetiology for breech presentation.
Nonetheless, we lacked data on some factors related to both breech presentation and ART, including maternal uterine anomalies, drug and medication use other than smoking during pregnancy and complications such as placenta praevia. We were also unable to specify the subtypes of IVF ± ICSI used, namely fresh or frozen transfer. Finally, it is possible that our results are a conservative estimate of the true odds, as data on the number of babies that were breech late in pregnancy but underwent a successful external cephalic version were not available in the data set.

CONCLUSIONS
Pregnancies conceived via both ART and OI are significantly associated with an increased approximate risk of breech presentation mediated by low birthweight and advanced maternal age.
This suggests a shared underlying mechanism for the aetiology of breech, for which further research is needed. Clinical assessment of fetal presentation at 36-37 weeks' gestation, especially given modified antenatal care programs with the SARS-CoV2 pandemic, and counselling with respect to this increased risk are warranted for women considering conception via ART.

ACKNO WLE DGE MENTS
The

SUPPORTING INFORMATION
Additional supporting information may be found online in the supporting information section at the end of the article.