Included studies

Five studies, reporting results from 3 different cohorts, were included in this review. Four observational studies (Brocklehurst 2011, Hollowell 2014, Hollowell 2015, Rowe 2018), reported results from the Birthplace in England cohort and UKMidSS cohort UK, and 1 retrospective cohort study (Stephenson-Famy 2018) reported results from the Washington State birth certificate cohort. Three studies (Brocklehurst 2011, Hollowell 2014 and Hollowell 2015) reported results from the same cohort (Birthplace in England cohort). The Birthplace publications differed in analysis due to different comparisons.

Four studies compared different BMI thresholds (Hollowell 2014, Hollowell 2015, Rowe 2018, Stephenson-Famy 2018). The results were stratified according to place of birth and parity. One study (Hollowell 2015) compared different places of birth and the results were stratified by parity. One study (Brocklehurst 2011) compared different places of birth, but only 1 of the outcomes was stratified by parity, other outcomes were not stratified by parity or BMI. The results from this study were not stratified by BMI or parity, so data from the 2 studies reporting further analysis from the same cohort were used (Hollowell 2014 and Hollowell 2015).

The studies were from England, Northern Ireland, Scotland, United States and Wales.

The included studies are summarised in Table 2.

See the literature search strategy in appendix B and study selection flow chart in appendix C.

Excluded studies

Studies not included in this review are listed, and reasons for their exclusion are provided in appendix J.

BMI <18.5 kg/m² versus BMI range 18.5 – 24.9 kg/m²

When a BMI <18.5 kg/m² was compared to the healthy weight range BMI, most of the evidence showed no evidence of an important difference for the outcomes of maternal admission to intensive care, modes of birth, maternal blood transfusion and transfer to an obstetric unit. The quality was mainly low for these outcomes due to imprecise findings, so should not be taken as definitive evidence of no difference between the groups. The evidence showed no important difference between groups for the combined outcome of obstetric interventions and adverse maternal outcomes (instrumental birth, intrapartum caesarean birth, augmentation, general anaesthesia, maternal blood transfusion, third/fourth degree tear, and maternal admission to higher level care) in nulliparous and mixed parity women whose planned place of birth was an obstetric unit, and no evidence of a difference in mixed parity women whose planned place of birth was a freestanding midwifery unit, alongside midwifery unit or home. There was an exception seen for the combined outcome of neonatal admissions/intrapartum stillbirth/early neonatal death in women of mixed parity who had planned birth in an alongside midwifery unit, where there was an important benefit for women with a BMI <18.5kg/m² when compared to the healthy weight range BMI. However, there was no evidence of a difference between the two BMI groups for this outcome in women of mixed parity who had planned birth at home or freestanding units, or women of any parity who had planned birth in the obstetric unit.

BMI range 25 – 29.9 kg/m² versus BMI range 18.5 – 24.9 kg/^m2

When BMI range 25 – 29.9 kg/m² was compared to BMI range 18.5 – 24.9 kg/m², most of the evidence showed no important difference between outcomes, and some of the evidence showed no evidence of an important difference. There were more intrapartum caesarean births in the obstetric unit in group BMI range 25 – 29.9 kg/m², in women of mixed parity.

There was no important differences between the BMI groups on transfer to an obstetric unit from home or in alongside unit for nulliparous women, and no evidence of a difference for transfers from home or an alongside unit for multiparous women. Two studies reported transfer to obstetric unit from a freestanding midwifery unit. One study from the UK, showed no important difference between groups in nulliparous women, and no evidence of a difference for multiparous women. However, 1 study from the US showed more transfers for nulliparous women with a BMI range 25 – 29.9 kg/m². Differences in the direction of effect could be attributed to the setting, as protocols for transfer may differ between the US and the UK. Most of the evidence was of high quality, with some evidence rated as moderate and low due to concerns around imprecision.

BMI range 30 – 35 kg/m² versus BMI range 18.5 – 24.9 kg/m²

When BMI range 30-35 kg/m² was compared to BMI range 18.5 – 24.9 kg/m², there was an important benefit for women of mixed parity in the higher BMI range, 30 – 35 kg/m², who had planned birth in a freestanding midwifery unit, with a reduction in the combined outcome of obstetric interventions and adverse maternal outcomes. The evidence was rated as moderate quality due to imprecision. High quality evidence showed no important difference for this outcome for nulliparous women who planned birth in the obstetric unit. High quality evidence also showed no important difference between groups for this outcome for planned birth in the obstetric unit, at home, or in an alongside midwifery unit, in women of mixed parity. High quality evidence also showed more intrapartum caesarean births in the group BMI range 30-35 kg/m², in the obstetric unit in women of mixed parity, and moderate quality evidence showed more transfers to the obstetric unit from home in multiparous women. The remaining outcomes showed no important difference or no evidence of an important difference between groups.

BMI ≥30 kg/m² versus BMI range 18.5 – 24.9 kg/^m2

When a BMI ≥30 kg/m² was compared to BMI range 18.5 – 24.9 kg/m², the evidence showed more transfers from a freestanding unit to the obstetric unit in nulliparous women with a BMI ≥30 kg/m² compared to nulliparous women with BMI range 18.5 – 24.9 kg/m². Transfer to an obstetric unit was the only outcome available for this comparison. The evidence was rated as moderate quality, with some concerns over risk of bias.

BMI >35 kg/m² versus BMI range 18.5 – 24.9 kg/^m2

A BMI >35 kg/m² was compared to BMI range 18.5 – 24.9 kg/m² in women planning birth in an obstetric unit. The evidence showed no important difference, or no evidence of an important difference for obstetric interventions and adverse maternal outcomes combined for nulliparous and multiparous women, and also for maternal admission to intensive care, or maternal blood transfusion in women of mixed parity. There were fewer instrumental births with a BMI >35 kg/m² in women of mixed parity than in the BMI range 18.5 – 24.9 kg/m², but more intrapartum caesarean births. There was an important harm for a BMI >35 kg/m² when compared to BMI range 18.5 – 24.9 kg/m² for neonatal admission or intrapartum stillbirth/early neonatal death for nulliparous and multiparous women. The quality of the evidence ranged from high to low, with concerns around imprecision.

BMI range >35 – 40 kg/m² versus BMI range 18.5 – 24.9 kg/^m2

When BMI >35 – 40 kg/m² was compared to BMI range 18.5 – 24.9 kg/m², high quality evidence showed no important difference between groups for obstetric interventions and adverse maternal outcomes combined in women of mixed parity in the obstetric unit. In addition, low quality evidence showed no evidence of an important difference in this outcome in women of mixed parity at home, in freestanding midwifery units or alongside midwifery units. The evidence was downgraded for imprecision and should therefore not be taken as definitive evidence of no difference between groups. There was an important harm of BMI range >35-40 kg/m² for neonatal admissions or intrapartum stillbirth/early neonatal death in the obstetric unit (high quality) and the freestanding midwifery unit (moderate quality). The evidence was downgraded for imprecision. Low quality evidence showed no evidence of a difference for this outcome at home or the alongside midwifery unit, with concerns around imprecision, so should not be taken as definitive evidence of no difference between groups. At a BMI range of >35-40 kg/m², is it national guidance that women plan their birth at an obstetric unit, therefore concerns around imprecision are due to small sample sizes of women planning birth in settings other than the obstetric unit.

BMI >35 kg/m² versus BMI ≤35kg/m²

A BMI >35 kg/m² was compared to a BMI ≤35kg/m² in women planning birth in an alongside midwifery unit. Most of the evidence showed no important difference or no evidence of an important difference between groups. The exceptions were a possible important harm for intrapartum caesarean births, for nulliparous women with a BMI range >35kg/m², but not multiparous women. There was a harm in category 1 and 2 caesarean births for nulliparous women with a BMI range >35 kg/m², but not multiparous women. There was also a harm in terms of postpartum haemorrhage in nulliparous women with BMI range >35 kg/m², but no evidence of an important difference in multiparous women. Most of the evidence was downgraded due to concerns around imprecision.

Planned places of birth at home or freestanding midwifery units or alongside midwifery units versus obstetric units

Some comparisons compared different planned places of birth to planned birth in an obstetric unit. The comparisons reported data on modes of birth, with stratifications by parity. The mean BMI range in both groups of women was 18.5 – 24.9 kg/m². Planned place of birth at home, in freestanding midwifery units, and alongside midwifery units all had a benefit over planned place of birth in obstetric units in terms of instrumental births, and caesarean births. There was an important benefit of planned place of birth at home, and in freestanding midwifery units over planned place of birth in obstetric units for spontaneous vaginal births in nulliparous women, but no important difference in multiparous women. There was no important difference between planned place of birth in alongside midwifery units and obstetric units in terms of spontaneous vaginal births for nulliparous or multiparous women. The evidence was rated as moderate to high quality, with some outcomes downgraded for imprecision.

There was no evidence identified for all the different places of birth compared to each other.

There was no evidence identified for the following outcomes: maternal death, or women’s experience of labour and birth.

See appendix F for full GRADE tables.

Included studies

See the literature search strategy in appendix B and economic study selection flow chart in appendix G.

A systematic review of the economic literature was conducted but no economic studies were identified which were applicable to this review question.

Excluded studies

Economic studies not included in this review are listed, and reasons for their exclusion are provided in appendix J.

The outcomes that matter most

The committee chose maternal death or severe maternal morbidity, mode of birth and postpartum haemorrhage as the critical outcomes for this review. The committee agreed that to understand the safety of planned place of birth, maternal death or severe maternal morbidity, and postpartum haemorrhage would be the best indicators of the most severe negative outcomes for women. The committee also agreed that it was essential to find out about mode of birth, and whether different planned places of birth for women with different BMIs or parities, led to differences in mode of birth.

The committee also agreed on the important outcomes for this review. They agreed that it was important to look at the neonatal outcomes and whether there were any risks associated with planned place of birth for the baby. They agreed that shoulder dystocia and admission to neonatal unit would reflect this. The committee discussed that admission to neonatal unit may not cover all aspects of morbidity, and in some instances, admissions would be due to precaution, however they agreed that this outcome would reflect the risks of separation between the mother and their baby. The committee agreed that shoulder dystocia was one of their key concerns in terms of neonatal morbidity and so used this as another indicator of morbidity of the neonate. The committee also discussed the importance of looking at breastfeeding rates, as this would be an indicator of the support available to the woman, and would also benefit the baby. The committee chose transfer to the obstetric unit as another important outcome as this would indicate if there had been any requirements or complications that could not be dealt with at the planned place of birth. The committee agreed that it was also important to find out about maternal satisfaction with labour and birth. The committee recognised the great importance of maternal satisfaction for place of birth, but they were aware that data on this outcome was likely to be sparse and unlikely to inform decision-making in a meaningful way, so they prioritised other outcomes as critical.

The quality of the evidence

The quality of the evidence for outcomes was assessed with GRADE and was rated from high to very low. Most of the evidence was downgraded due to imprecision around the effect estimate. Most women have a BMI in the range 18.5-25.9 kg/m², therefore sample sizes for women with other BMI ranges were smaller and affected imprecision. Some outcomes were also downgraded for inconsistency, as the difference in the direction of effect could not be explained by further subgroup analysis. Some studies were also downgraded for risk of bias concerns, mainly around selective reporting of subgroup analyses. All studies adjusted for confounders and the adjusted effect estimates were used, therefore the absolute effects were based on the relative effect applied to the observed effect rate.

Benefits and harms

The committee discussed the evidence for women with a healthy BMI range of 18.5 to 24.9 kg/m². This evidence showed that, for both nulliparous and multiparous women, planning birth in freestanding midwifery-led units, alongside midwifery-led units or at home, had benefits in terms of reduced obstetric interventions when compared to planning birth in an obstetric unit. The committee agreed that the evidence supports the current recommendation in the guideline that the rate of interventions is lower if birth is planned in midwifery-led units or at home for low-risk nulliparous and multiparous women. Therefore, the committee agreed to keep this recommendation in the guideline.

The committee were aware of data from the National Maternity and Perinatal Audit 2021 (Relph 2021) that shows that as BMI increases, the risk of intrapartum interventions, postpartum haemorrhage and adverse neonatal outcomes also increase. The National Maternity and Perinatal Audit 2021 was not formally included in the review as it did not meet the comparator criteria listed in the review protocol. They then reviewed the evidence to see if there were any identifiable risks with raised BMIs.

The committee discussed that the evidence in this review provided information about the risks of events occurring at different BMIs or BMI ranges compared to other BMIs or BMI ranges, but this was within defined planned places of birth. They discussed the limitations of the data presented, including the fact that for some of the outcomes they could not ascertain whether the benefits and risks presented were due to the planned place of birth or the BMI range. This is because some outcomes were not available for all planned places of birth and it was not possible to tease out whether a worse outcome would be better with a different planned place of birth. The committee had hoped the evidence would provide information between the risks and benefits of planning birth in a particular setting, for a given BMI or BMI range. Nonetheless, the committee agreed the evidence would be useful for advising women about their potential individual risks during labour and what support and care they might require, and therefore support them when planning their birth to identify where the best place of birth might possibly be for them. They discussed that overall the evidence used to make recommendations was of moderate to high quality and largely applicable to the UK context.

The committee discussed the best way to present the evidence so it could be used by women and clinicians when discussing their planned place of birth. They noted that they could not establish a BMI range cut-off above which planning birth in a specific setting was no longer recommended. However, the evidence showed that some outcomes increase with a higher BMI and the committee agreed to make a recommendation to summarise the key messages from the data and to present these risks in tables in an appendix to the guideline so it could be used to inform women when discussing and planning their place of birth.

The committee noted that one of the main areas of concern for women when planning place of birth are outcomes that lead to the separation between them and their baby. The committee noted that this should not be a concern for women when planning place of birth and that, in line with existing guideline recommendations, if a woman is transferred to an obstetric unit after the birth, healthcare professionals should ensure that her baby goes with her. However, the committee agreed that women may be separated from their baby if the baby requires admission to the neonatal unit, particularly if a higher level of neonatal care required transfer of the baby to another hospital, and this would be a concern for all women.

The committee first discussed the evidence for women with a booking BMI of <18.5kg/m² (that is, with a BMI lower than the ‘healthy’ range) compared to women with a booking BMI of 18.5 – 24.9 kg/m². Across all settings (obstetric units, alongside midwifery units, free-standing midwifery units and home) there was no evidence of an important difference for any of the available outcomes, with the exception of the combined neonatal outcome (neonatal admission, stillbirth, neonatal death), where there was a benefit for women with the lower BMI when planning birth in an alongside midwifery unit.

The committee next discussed the evidence for the comparison of women with a booking BMI in the range 25-29.9 kg/m² compared women with a booking BMI in the healthy weight range 18.5 to 24.9 kg/m². Across all settings (obstetric units, alongside midwifery units, freestanding midwifery units and home) there was no evidence of an important difference for most of the available outcomes, but there were 2 outcomes which showed an increased risk for women with a higher BMI:

1. Women planning birth in an obstetric unit (mixed parity) are more likely to have a caesarean birth. The committee noted that there was no data available from other planned places of births on caesarean births.
2. Nulliparous women planning birth in a freestanding midwifery unit were more likely to be transferred to an obstetric unit. However, for this difference, the committee noted that this evidence came from a US setting, and although the facilities at the freestanding midwifery units resemble those in the UK, there would be other differences such as reasons for transfer between the two settings that could explain the contradictory evidence. The committee also had concerns over the low quality of the evidence from the US setting. Therefore the committee based their decisions on other evidence from the UK setting that showed no differences between nulliparous women in the two BMI groups. As a result, the committee agreed that they would not highlight specific risks related to transfer to the obstetric unit from a freestanding midwifery unit, for nulliparous women with a booking BMI in the range 25-29.9 kg/m².

Overall, the committee therefore agreed to only highlight the risks relating to the first difference for this set of results.

The committee then discussed the evidence for women with a booking BMI in the range 30 – 35 kg/m², compared to women with a healthy booking BMI of 18.5 – 24.9 kg/m². As with the evidence for women with a booking BMI in the range 25 – 29.9 kg/m², there was no evidence of an important difference in the outcomes for the majority of comparisons, but for 3 outcomes there was a difference. For 2 of these differences an increased risk was shown for women with an increased BMI:

1. Women planning birth in an obstetric unit (mixed parity) are more likely to have a caesarean birth. The committee noted that there was no data available from other planned places of births on caesarean births.
2. Multiparous women planning birth at home, are more likely to be transferred to the obstetric unit.

The committee noted that these were the same risks that had been identified in the evidence for the previous BMI range 25 – 29.9 kg/m² and this reinforced their recommendations to highlight these risks to women with a raised BMI.

The third difference identified in this comparison was an increased risk for women with a healthy BMI, which was not what the committee expected to see. The increased risk was for the combined outcome of obstetric interventions and adverse maternal outcomes in a freestanding midwifery unit. However, for planned place of birth in all other settings (obstetric unit, alongside midwifery unit and home) there was no important difference, or no evidence of an important difference for this outcome. The committee discussed that this combined outcome included many different outcomes: instrumental births, intrapartum caesarean births, augmentation, general anaesthesia, maternal blood transfusion, third/fourth degree tears, and maternal admission to higher level care, and the evidence did not explain which specific component of the combined outcome contributed to the increased rate seen in women in the lower BMI range. They agreed that this lack of clarity regarding this outcome, the fact that the difference had only been seen in one setting, and was contrary to their expectations about risks increasing with increased BMI, meant that it should not be included in their recommendation.

The next BMI range the committee discussed was for women with a booking BMI >35 kg/m² together with the further analysis of women with a booking BMI in the range >35-40 kg/m² (both of which were compared to women with a booking BMI in the healthy range, 18.5 – 24.9 kg/m²). The committee noted that this evidence showed that women with a booking BMI >35 kg/m² or 35 – 40 kg/m² planning birth in an obstetric unit (nulliparous or multiparous) or a FMU (mixed parity) were more likely to experience the combined outcome of neonatal admission, intrapartum stillbirth or early neonatal death, and agreed that this increased risk should be included in the planned risks table. The committee discussed that combining neonatal outcomes in this way did not provide enough information on neonatal admissions that were low dependency, or did not result in serious outcomes. However, the committee discussed that this outcome was still informative for making decisions about planned place of birth, as neonatal units are located alongside obstetric units. Therefore they agreed that it was important to highlight this as a risk for the groups of women with a BMI >35 kg/m² or 3540 kg/m², so they could use this information to make decisions on their planned place of birth.

The committee also discussed that women with a booking BMI of >35 kg/m² planning birth in the obstetric unit were more likely than those of a BMI in the range 18.5-24.9 kg/m² planning birth in the obstetric unit, to have an intrapartum caesarean birth. There was no evidence for this outcome for other planned birth settings so the committee were unable to comment on the association of planned place of birth with intrapartum caesarean birth, but evidence for the combined outcome measure of obstetric interventions and adverse maternal outcomes found no difference between the higher (>35 kg/m² or 35-40 kg/m²) and healthy BMI ranges in any setting. Therefore the committee agreed that this risk should be highlighted for all women with a booking BMI >35 kg/m², based on the available evidence.

The committee discussed the evidence that compared women with a booking BMI ≥35 kg/m² to women with a booking BMI ≤35 kg/m², who planned their birth in an alongside midwifery unit. The committee noted that over 90% of the women with a booking BMI ≥35 kg/m² had a BMI between 35.1 to 40 kg/m², so they acknowledged that the evidence may not be applicable to women with a BMI >40 kg/m². The committee noted that the evidence showed an increased risk for nulliparous women with a BMI ≥35 kg/m² of intrapartum or emergency caesarean births, and postpartum haemorrhage. They noted that for intrapartum caesarean births, the increased risk was only seen when using a more liberal confidence interval of 90%, rather than 95%, but agreed that the absolute risks presented in the table of risks would highlight this. They acknowledged that these data came from women who planned their birth in alongside midwifery units, but without data from other planned birth settings the committee could not comment on the risks associated with this particular setting. However, they agreed that the care and support required in the event of these outcomes could only be provided in the obstetric unit and so the need for transfer from the alongside midwifery unit to the obstetric unit should be something that nulliparous women should consider in their decision making when planning their place of birth. The committee noted that there was no evidence of an important difference for multiparous women with regard to these outcomes.

Cost effectiveness and resource use

The committee discussed the fact that the current recommendations suggest that all women with a booking BMI of 30-35 kg/m² should have an individual assessment when planning place of birth. The committee suggested that the removal of this hard cut-off and instead the inclusion of a risk table may mean that more women with lower BMIs choose birth at home, or at an alongside midwifery unit or freestanding midwifery unit, and that this may reduce resource use whilst respecting and promoting individual choice with respect to place of birth.

Other factors the committee took into account

The committee were aware of data from the National Maternity and Perinatal Audit (NMPA) (Relph 2021) which showed that the likelihood of a woman experiencing an intrapartum intervention or adverse maternal outcome, or her baby experiencing very serious complications following birth, increases as BMI increases. The committee were therefore expecting to see increased risks in the higher BMI groups compared to women with a healthy BMI, and agreed that the data from the evidence review and the NMPA complemented each other, and supported their decision to alert women to these increased risks.

The committee discussed the fact that BMI ranges representing a healthy weight, overweight, or obesity may differ in women from different ethnic groups, and that this should be taken into consideration when assessing the risks for women at different BMIs. However, the committee noted that the evidence they had reviewed included a proportion of women from black, Asian and minority ethnic groups ranging from 12 to 16% (which is representative of the UK population) but that there was no separate evidence from the review on women from specific ethnic groups so the committee were unable to make separate recommendations. To provide additional information for users of the guideline the committee cross-referenced to the NICE guideline on the classification of overweight and obesity which provides guidance on how to adjust ranges for different ethnic groups. In addition, the committee noted that women from certain ethnic and socioeconomic groups may be likely to be overweight or obese and so the recommendations may apply to a higher proportion of women in these groups than in other groups.