With the improvement of assisted reproduction technology, and greater public awareness, a growing number of children are born to women with a history of infertility (Sutcliffe and Ludwig, 2007). Despite its high success rate, there are concerns about the impact of assisted conception on maternal and neonatal wellbeing. Compared with women who spontaneously conceive, women who undertake assisted conception experience greater levels of depression (Fisher et al, 2012), more negative pregnancy outcomes, unfavourable birth experiences and more negative neonatal outcomes (Giallo et al, 2011; Cooklin et al, 2012; Giallo et al, 2014). Several studies have suggested an association between assisted conception and an increased rate of gestational diabetes, pre-eclampsia and caesarean delivery. Neonatal outcomes have also been reported to be poor: babies born as a result of assisted conception are more likely to be admitted to the neonatal intensive care unit, and have an increased risk of congenital abnormalities, low birth weight and preterm delivery (Hammarberg et al, 2008a; 2008b; McGrath et al, 2010). However, controversial results have been reported by Shevell et al (2005), who indicated no association between assisted conception and fetal growth restriction, or fetal anomalies.
Childbirth has a reputation for being one of the most painful events in a woman's life (Lowe, 2002). A variety of non-pharmacological complementary approaches have been proposed to assist with the progress of physiological labour through effective pain management. These include, but are not limited to: massage therapy (Hosseini et al, 2013; Bolbol-Haghighi et al, 2016), transcutaneous electrical nerve stimulation (Sluka and Walsh, 2003), acupuncture (Tempfer et al, 1998), acupressure (Lee et al, 2004), yoga and relaxation (Field, 2011), exercise, visualisation, controlled breathing (Levett et al, 2016), reflexology (Smith et al, 2012), music therapy (Taghinejad et al, 2010), water immersion (Benfield et al, 2010), heat therapy and the use of a birth ball (Taavoni et al, 2016). A meta-analysis of 57 studies published from January 1990–December 2012 (Chaillet et al, 2014) showed that non-pharmacological methods to relieve labour pain could offer benefits to women and their newborns without additional harm.
Some women choose to use epidural analgesia to help decrease their stress and pain (Thomas et al, 1982). Use of opiate drugs for epidural analgesia has been reported to prolong second-stage labour (Agrawal et al, 2014) and to be associated with an increased risk of caesarean section (Herrera-Gómez et al, 2017). In contrast, Cochrane reviews (Anim-Somuah et al, 2011; Sng et al, 2014) have shown that early or late initiation of epidural analgesia for labour does not affect labour and birth outcomes.
Despite the increasing rate of assisted conceptions, research comparing labour pain relief in women who had and had not undertaken assisted conception is scarce, and the effect of epidural analgesia on birth outcomes in women who have undertaken assisted conception has been under-studied. Data derived from assisted reproduction technology registries only provide results for overall pregnancy outcomes, not on women's mode of labour pain relief. To address this existing gap in the literature, this study aimed to evaluate the labour and birth outcomes in women who had and had not undertaken assisted conception and how the use of labour pain relief affects their outcomes.
Methods
Design and ethics approval
This was a retrospective cohort study. Ethics approval was granted from the Human Research Ethics Committee at the Western Sydney Local Health District before the commencement of the study.
Setting and participants
Data for evaluation and analysis were collected from women who gave birth at Westmead Hospital between 1 July 2010 and 30 June 2017. Inclusion criteria were that the woman must have had a singleton pregnancy; given birth to a live, term baby between complete 37 and 43 weeks; and been primiparous. Data from multiparous women, twins or triplets and preterm births were excluded from the study.
Procedure
Women's records were selected from the ObstetriX database, an obstetric information system based on the clinical dataset developed in Australia by the New South Wales OBSTET consortium. It is a unique clinical data management tool and is used in most New South Wales Local Health Districts to collect obstetric data. The recorded data were collectively reported as an archetype report and were saved in a Microsoft Excel spreadsheet. The variables extracted from the ObstetriX database were demographic data; obstetric and gynaecological history; medical history; pregnancy, labour and birth outcomes; and the newborns' characteristics and outcomes.
Statistical analysis
The data were downloaded from the Excel spreadsheet to SPSS Advanced Statistics version 24.0. Descriptive statistics were used to summarise the demographic data. Associations between contributing factors and outcomes were examined using the c2-test (or Fisher exact test if applicable) and Student t tests, where appropriate. Univariate and multivariate regression modelling were used to estimate crude and adjusted effects of factors of interest. Gaussian, logistic or multinomial logistic link functions were used for continuous, dichotomous or multicategorical outcome variables, respectively. Odds ratios (ORs) were calculated to report the effect size for logistic models. Differences were considered significant when P<0.05. Data preparation and cleaning were performed in Excel and validated through visual review, value range checks, field type checks, and logical checks by researchers. Statistical analyses were performed in R, version 3.1.2.
Results
During the period of data review, there were 39 994 births in the hospital. Data were collected from 11 135 records for the analysis (575 women who had undertaken assisted conception and 10 560 women who had conceived spontaneously) (Figure 1).
The baseline characteristics of the women are shown in Table 1. Compared to women who had spontaneously conceived, women who had undertaken assisted conception were more likely to be older, non-smokers, under obstetrician or GP-led care during their pregnancy and have a larger body mass index (BMI) (P<0.001). Although the difference in rate of gestational diabetes was statistically significant between the two groups (P=0.043), it was not clinically significant (0.5% vs 0.1%). There were no differences between the two groups when comparing other baseline characteristics (P>0.05) (Table 1), nor were there differences between the two groups for other medical conditions including renal, endocrine, neurological and autoimmune diseases, or for other antenatal complications, such as abnormal placental site, antepartum haemorrhage after 20th week of pregnancy, fetal growth restriction, fetal anomaly and threatened premature labour (P<0.05) (results not shown in the Tables).
| Variable | Type of conception | P | |
|---|---|---|---|
| Assisted (n=575) | Spontaneous (n=10 560) | ||
| Mean maternal age, years ± SD | 31.2 ± 4.9 | 27.5 ± 4.5 | <0.001 |
| Australian-born, n (%) | 196 (34.1%) | 3371 (31.9%) | 0.299 |
| Body mass index (BMI) at first antenatal visit, n (%) | |||
| <18.5 (underweight) | 24 (4.2%) | 870 (8.2%) | |
| 18.5–24.9 (normal) | 327 (56.9%) | 6634 (62.8%) | <0.001 |
| 25–29.9 (overweight) | 134 (23.3%) | 2135 (20.2%) | |
| ≥30 (obese) | 90 (15.7%) | 921 (8.7%) | |
| Smoking during this pregnancy, n (%) | 8 (1.4%) | 558 (5.3%) | <0.001 |
| Drug use during this pregnancy, n (%) | 0 (0.0%) | 98 (0.9%) | 0.036 |
| Alcohol use during this pregnancy, n (%) | 3 (0.5%) | 126 (1.2%) | 0.205 |
| Domestic violence, n (%) | |||
| Physical abuse | 2 (0.3%) | 136 (1.3%) | 0.073 |
| Frightened by partner | 4 (0.7%) | 114 (1.1%) | 0.505 |
| Gestational hypertension, n (%) | 22 (3.8%) | 12 (0.1%) | 0.787 |
| Gestational diabetes, n (%) | 3 (0.5 %) | 12 (0.1%) | 0.043 |
| Model of care during pregnancy, n (%) | |||
| Midwife-led | 227 (39.5%) | 5944 (56.3%) | <0.001 |
| Obstetrician or GP-led | 348 (60.5%) | 4616 (43.7%) | |
SD: standard deviation
The percentage of women receiving epidural analgesia was higher in women who had undertaken assisted conception (54.6%) than women who had spontaneously conceived (46.6%), with significant differences between the two groups (P<0.001). However, after adjusting for country of birth, BMI, maternal age, model of care during pregnancy and gestational hypertension, this difference was not statistically significant (OR=0.86; 95% CI=0.72–1.03) (Table 2).
| Type of analgesia | Type of conception | Crude OR (95% CI) | P | Adjusted OR* (95% CI) | P | |
|---|---|---|---|---|---|---|
| Assisted (n=575) | Spontaneous (n=10 560) | |||||
| Non-pharmacological† or no first stage | 214 (37.2%) | 4520 (42.8%) | Reference | – | Reference | |
| Intramuscular or intravenous narcotics | 47 (8.2%) | 1117 (10.6%) | 1.13 (0.82–1.55) | 0.473 | 1.04 (0.75–1.44) | 0.816 |
| Epidural | 314 (54.6%) | 4923 (46.6%) | 0.74 (0.62–0.89) | 0.001 | 0.86 (0.72–1.03) | 0.110 |
Ref. = reference value. OR = odds ratio.
Women who had undertaken assisted conception were more likely than women who had spontaneously conceived to undergo cervical ripening before labour (8.3% vs 4.5%, respectively), receive oxytocin for induction or augmentation of labour (67.7% vs 57.5%, respectively), have an instrumental birth (35.7% vs 29.4%, respectively) and have postpartum blood loss more than 500 ml (16.8% vs 13.3%, respectively) (P<0.001) (Table 3). No significant differences were found between the two groups in terms of other labour, birth and neonatal outcomes (Tables 3 and 4).
| Outcome | Type of conception | P | |
|---|---|---|---|
| Assisted (n=575) | Spontaneous (n=10 560) | ||
| Cervical ripening before labour, n (%) | |||
| Yes (prostaglandin gel, Foley catheter) | 48 (8.3%) | 479 (4.5%) | <0.001 |
| Use of oxytocin for induction of labour, n (%) | 389 (67.7%) | 6077 (57.5%) | <0.001 |
| Antenatal/intrapartum antibiotics, n (%) | 4 (0.7%) | 95 (0.9%) | 0.780 |
| Mean duration of first stage of labour ± SD | 5.5 ± 3.7 | 5.7 ± 3.8 | 0.215 |
| Mean duration of second stage of labour ± SD | 1.5 ± 1.2 | 1.4 ± 1.4 | 0.156 |
| Mode of birth, n (%) | |||
| Normal vaginal birth | 344 (59.8%) | 7008 (66.4%) | |
| Instrumental birth | 205 (35.7%) | 3109 (29.4%) | 0.004 |
| Caesarean birth | 26 (4.5%) | 443 (4.2%) | |
| Episiotomy/tears, n (%) | |||
| No tear | 237 (41.2%) | 4012 (38.0%) | |
| Second degree tear/episiotomy | 168 (29.2%) | 3215 (30.4%) | |
| Third–fourth degree tear | 41 (7.1%) | 600 (5.7%) | 0.099 |
| Other tears (first degree, clitoral, labial, parauretral, etc) | 129 (22.4%) | 2733 (25.9%) | |
| Postpartum blood loss (ml), n (%) | |||
| <500 | 478 (83.1%) | 9156 (86.7%) | |
| 500–1500 | 25 (4.3%) | 271 (2.6%) | 0.011 |
| >1500 | 72 (12.5%) | 1133 (10.7%) | |
SD: standard deviation; differences statistically significant at P<0.05
| Outcome | Type of conception | P | |
|---|---|---|---|
| Assisted (n=575) | Spontaneous (n=10 560) | ||
| Sex | |||
| Female | 281 (48.9%) | 479 (4.5%) | <0.001 |
| Male | 294 (51.5%) | 6077 (57.5%) | <0.001 |
| Birth weight (g) | |||
| <2500 | 19 (3.3%) | 282 (2.7%) | |
| 2500–4000 | 529 (92%) | 9697 (91.8%) | 0.481 |
| >4000 | 27 (4.7%) | 581 (5.5%) | |
| Apgar score <7 at 5 minutes | 9 (1.6%) | 154 (1.5%) | 0.976 |
| Neonatal complication at birth | 99 (17.2%) | 1575 (14.9%) | 0.148 |
Differences statistically significant at P<0.05
To investigate the role of labour pain relief in maternal and neonatal outcomes, a regression analysis was performed, adjusted only for labour pain relief. Primary results showed that ORs of stage 1 labour duration (OR=0.01; 95% CI=0.00–0.03; P=0.023), instrumental birth (OR=0.81; 95% CI=0.67–0.98; P=0.027) and postpartum blood loss of 500–1500 ml (OR=0.6; 95% CI=0.39–0.91; P=0.016) were less in women who did not receive labour pain relief (results not shown in Tables). However, after adjusting for maternal age, BMI, model of care during pregnancy, cervical ripening before labour and labour pain relief, no statistically significant differences were found between the two groups in terms of use of oxytocin for induction or augmentation of labour, mode of birth, episiotomy or tear, postpartum blood loss, Apgar score at 5 minutes and neonatal complications at birth (Table 5).
| Factor | Crude OR (95% CI) | P | Adjusted OR* (95% CI) | P |
|---|---|---|---|---|
| Duration of stage one labour | 0.18** (–0.11 to 0.48) | 0.22 | 0.2∞ (–0.09 to 0.48) | 0.177 |
| Use of oxytocin in labour | ||||
| No | Reference | Reference | ||
| Yes | 0.65 (0.54–0.77) | <0.001 | 0.83 (0.67–1.02) | 0.077 |
| Mode of birth | ||||
| Normal vaginal birth | Reference | Reference | ||
| Instrumental birth | 0.74 (0.62–0.89) | 0.001 | 0.09 (0.07–0.12) | 0.527 |
| Caesarean birth | 0.84 (0.56–1.26) | 0.393 | 1.25 (0.82–1.92) | 0.301 |
| Episiotomy/tears | ||||
| No tears or other tears | Reference | Reference | ||
| Second degree tear/episiotomy | 1.04 (0.86–1.25) | 0.694 | 0.97 (0.8–1.17) | 0.722 |
| 3rd–4th degree tears | 0.79 (0.57–1.11) | 0.175 | 0.78 (0.55–1.1) | 0.154 |
| Postpartum blood loss (mL) | ||||
| <500 | Reference | Reference | ||
| 500–1500 | 0.57 (0.37–0.86) | 0.008 | 0.7 (0.46–1.09) | 0.112 |
| >1500 | 0.82 (0.64–1.06) | 0.131 | 0.94 (0.72–1.22) | 0.627 |
| Apgar at 5 min | ||||
| ≥ 7 | Reference | Reference | ||
| <7 | 0.93 (0.5–1.97) | 0.835 | 1.13 (0.6–2.44) | 0.722 |
| Neonatal complication at birth | ||||
| No | Reference | Reference | ||
| Yes | 0.84 (0.68–1.06) | 0.133 | 1 (0.79–1.26) | 0.970 |
Discussion
This study investigated labour and birth outcomes in women who had given birth after spontaneous conception and women who had given birth after undertaking assisted conception. The authors evaluated how the use of labour pain relief affected maternal and neonatal outcomes. Results of the study showed that a greater number of women who had undertaken assisted conception were older, overweight or obese and non-smokers. These findings were not surprising, because previous research has shown an association between infertility, obesity (Gesink Law et al, 2007) and older age (Tendais and Figueiredo, 2016).
The findings showed that women who had undertaken assisted conception were more likely than women who had spontaneously conceived to be in the care of an obstetrician or GP. It could be assumed that women with a history of infertility might have other concurrent medical conditions (Palomba et al, 2016) that require them to be in the care of an obstetrician or GP due to a high-risk pregnancy; however, these findings did not support this assumption, as there were no significant differences in the women's medical conditions and antenatal complications when the two groups were compared. Despite this, it may not be possible to rule out this assumption because women who gave birth before the 37th week of pregnancy were excluded from data collection, and preterm birth might have been due to their antenatal complications.
The other assumption that could be made is that nulliparous women who undertake assisted conception are likely to be more concerned about their birth outcomes than women who spontaneously conceive, and may be more likely to request a caesarean section, which is also more likely to occur when a woman is in the care of a private obstetrician (Roberts et al, 2000; Gillet et al, 2011). Research has shown that women who undertake assisted conception have a 2.95 times higher risk of planned caesarean birth (Chien et al, 2015); however, the results of this study did not support this association and showed no difference between the two groups in their mode of birth and the rate of their caesarean section. Because this was a retrospective study using patient records, it was not possible to investigate the reason why women who had undertaken assisted conception were more likely to be in the care of an obstetrician or GP. Further prospective studies are needed to clarify this.
This study showed that 54.6% of women who had undertaken assisted conception used epidural analgesia, compared with 46.6% of women who had spontaneously conceived. However, after adjusting for several confounding variables, including country of birth, BMI, maternal age, model of care during pregnancy and gestational hypertension, the odds of having an epidural did not differ between the two groups. This indicated that the type of pain relief used during labour was not solely affected by the type of conception and there were other factors that affected labour pain management techniques. For example, research has shown that giving birth without epidural analgesia was associated with unfavourable social and organisational factors, such as labour occurring too quickly, medical contraindication and unavailability of anaesthetists (Le Ray et al, 2008). Another study (Jeschke et al, 2012) reported that firsttime labour, the desire for birth without pain, and previous use of epidural analgesia were independent predictors of the use of epidural analgesia in labour. Because this was a retrospective study using patient records, the authors were unable to investigate the social situations of patients, their desire for labour pain relief and the availability of anaesthesiologists. Nevertheless, it is acknowledged that the choice of labour pain relief is affected by many factors, not only the type of conception.
The χ2-test showed that women who undertook assisted conception were more likely to receive oxytocin during labour, undergo instrumental birth and have postpartum haemorrhage. However, regression analysis adjusted for maternal age, BMI, model of care and epidural analgesia showed no significant difference between women who undertook assisted conception and women who spontaneously conceived, indicating that the dynamics of labour and birth were affected not only by the type of conception, but also by other confounding factors contributing to labour and birth outcomes. For example, research by Adams et al (2015) investigated the role of labour analgesia and showed that women's use of non-pharmacological pain management decreased the likelihood of their baby being admitted to a special care nursery, and the use of epidural analgesia increased the likelihood of instrumental birth and neonatal complications at birth necessitating admission to a special care nursery. In other research, Carlson et al (2017) evaluated the association between model of care, BMI, and maternal and neonatal outcome. Their study showed that, compared to women with high BMI who were in the care of obstetricians, women with high BMIs who were cared for by a nurse-midwife were less likely to have an operative vaginal birth and a third- or fourth-degree perineal tear, and were more likely to have labour anaesthesia and synthetic oxytocin augmentation. Nevertheless, they reported that the rates of unplanned caesarean birth, postpartum haemorrhage, maternal intrapartum fever and neonatal intensive care unit admission were similar between the two groups.
Implications for practice
The results of this study confirmed that the type of conception was not the sole indicator of the labour and birth outcomes, and that there were many other factors that contributed to the wellbeing of mothers and their newborns. These findings will help inform best practice to support labouring women who undertake assisted conception and provide them with the most appropriate information about the factors that can affect their perinatal outcomes, such as their country of birth, BMI, model of care, medical history and labour analgesia. The results can also help facilitate shared decision-making to help women who undertake assisted conception to receive appropriate care during labour and birth and achieve the most favourable outcomes. This will in turn help to promote the general health and quality of life of these women, followed by significant effects on the wellbeing of the community and society.
Limitations and strengths of our study
This study had limitations that are inherent in the use of administrative data. The retrospective design of the study necessitated the use of routinely collected data. The authors relied on information entered by clinical midwives into a database not purpose-built for this research, and excluded the records of women with missing data, although no difference was found in the baseline demographics (age and country of birth) between excluded and included women. Records of multiparous women, which might have yielded different outcomes, were also excluded from analysis. Despite these shortcomings, the large sample size of this study (11 135 women) adds further evidence of labour and birth outcomes in women who do and do not undertake assisted conception, and how labour pain relief modulates this association. Future research needs to investigate labour and birth outcomes in multiparous women with a history of infertility and assisted conceptions, and should evaluate the use of labour pain relief in these women and how it affects their outcomes.
Conclusion
Use of labour pain relief is associated not only with the type of conception, but also with other factors, such as country of birth, BMI, maternal age, model of care during pregnancy and gestational hypertension. In addition, having undertaken assisted conception is not the sole indicator of oxytocin use during labour, instrumental birth and postpartum haemorrhage, as factors such as maternal age, BMI, model of care and epidural analgesia facilitate this association. These findings provide a better understanding of the management of factors leading to poor outcomes in women who undertake assisted conception, and will help to develop better strategies to reduce obstetric complications in this population.