Reply to the Letter to the Editor by Kember et al, Regarding Wilson et al., “A Position Modification Device for the Prevention of Supine Sleep During Pregnancy: A Randomised Crossover Trial” Published in the British Journal of Obstetrics & Gynaecology on 16 September 2024.Danielle L. Wilson, MSc., PhD1,2,3; Carley Whenn1; Maree Barnes, MBBS1,4; Susan P. Walker, MBBS, MD, FRANZCOG, DDU, CMFM2,5; Mark E Howard, MBBS, FRACP, GDEB, PhD1,41Institute for Breathing and Sleep, Austin Health, Heidelberg, Victoria, Australia;2Department of Obstetrics and Gynaecology, University of Melbourne, Parkville, Victoria, Australia.3School of Electrical Engineering and Computer Science, The University of Queensland, St Lucia, Queensland, Australia.4Department of Medicine, University of Melbourne, Parkville, Victoria, Australia.5Mercy Perinatal, Mercy Hospital for Women, Heidelberg, Victoria, Australia.Address correspondence to: Danielle L. Wilson, Institute for Breathing and Sleep, Level 5 Harold Stokes Building, Austin Health, Heidelberg, Victoria, Australia. Tel: 613 9496 3517; Fax: 613 9496 5124; e-mail: [email protected] Prof. Aris Papageorghiou,We thank Dr. Kember and his colleagues1 for their interest in our recently published article2. Firstly, we acknowledge the recent publication by Coleman et al3 regarding the benefits of the Prenabelt on foetal growth, as revealed through re-analysis of their data using Bayesian methods. Unfortunately, this paper3 was published after our initial submission. Interestingly, while the frequentist analysis approach just fell short of significance, a Bayesian approach shows promise for supine sleep position interventions, by throwing an interesting light on the likelihood of a benefit depending on prior clinical beliefs. It remains that the mechanisms of action need further exploration however, given that supplemental data from the original publication4 suggests that objectively measured supine sleep in a subset of women did not differ between those using Prenabelt v sham.Secondly, we recognise our study was constrained by the measurement device used. The Night Shift Sleep Positioner (Night Shift) with neck placement was initially chosen, however, during our study a chest belt was introduced by the manufacturer. We performed a sub-study with N=20 of the participants wearing the device at both neck and abdominal placements for one night. An epoch-by-epoch comparison revealed substantial agreement between the two devices for all sleep positions with 88% concordance overall (unweighted κ = .797 [SE .006], p<.001). Regrettably, we neglected to include these data within the original publication.Despite the limitations of our measurement device, we believe that distinguishing between neck and pelvic position would not have substantially impacted our conclusions, given the high concordance between neck and abdominal devices revealed above, and particularly given that the intervention showed a nonsignificant increase in the proportion of supine sleep overnight. As Dr. Kember’s team have demonstrated, sleep positions where the thorax and pelvis are discordant (e.g., supine thorax with left pelvis tilt) are infrequent, occurring in less than 6% of “real-world” sleeping positions during pregnancy.5We agree that the Night Shift lacked sufficient resolution to fully appreciate the complex interplay between maternal positioning and foetal haemodynamics, as mentioned in the limitations section of our paper. Indeed, measurement limitations persist with all position sensing methodologies including infra-red video monitoring, with Dr. Kember’s recent paper5 demonstrating that while the most commonly occurring sleeping positions (left-lateral, right-lateral and supine) were well-recognised by modelling, the twisted/hybrid positions had intermediate performance, with the modelling particularly challenged by left or right tilted positions. To address the limitations of the Night Shift, we have recently completed data collection on a subsequent study using tri-axial accelerometry at abdominal level to characterise maternal sleep position in degrees of roll around the axial plane, with an aim to capture the subtleties of maternal position on uteroplacental haemodynamics and foetal growth.As mentioned, this is an understudied area. Dr. Kember and his team share our drive to gather objective evidence of a causal pathway between supine sleep position and foetal wellbeing, which we believe is particularly important given current guidelines regarding safe sleeping position in late pregnancy are based on retrospective cohort studies of self-reported “going-to-sleep” position. Our team congratulates Dr. Kember and his colleagues for their excellent work in the sleep position during pregnancy space, and we look forward to the outcomes from their upcoming DOSAGE Study.Disclosure of Interests:The authors have no financial disclosures or conflicts of interest.Contributions of Authorship:Dr. Wilson, Prof. Walker and Prof. Howard drafted and revised the manuscript, Ms. Whenn and Dr. Barnes critically revised the manuscript.Details of Ethics Approval:Additional analyses comparing device location were approved under an amendment to the original approval by the Mercy Hospital for Women Human Research Ethics Committee project number 2020-015, on 3rd June 2021.References:https://doi.org/10.1111/1471-0528.179521. Kember A. Placeholder for Letter to the Editor for BJOG.2. Wilson Danielle L, Whenn C, Barnes M, Walker Susan P, Howard Mark E. A position modification device for the prevention of supine sleep during pregnancy: A randomised crossover trial. BJOG: An International Journal of Obstetrics & Gynaecology. 2024; .3. Coleman J, Grewal S, Warland J, Hobson S, Liu K, Kember A. Maternal positional therapy for fetal growth and customised birth weight centile benefit in a Bayesian reanalysis of a double-blind, sham-controlled, randomised clinical trial. BMJ Open. 2024; 14 (4): e078315.4. Coleman J, Okere M, Seffah J, et al. The Ghana PrenaBelt trial: a double-blind, sham-controlled, randomised clinical trial to evaluate the effect of maternal positional therapy during third-trimester sleep on birth weight. BMJ Open. 2019; 9 (4): e022981.5. Kember AJ, Zia H, Elangainesan P, et al. Transitioning sleeping position detection in late pregnancy using computer vision from controlled to real-world settings: an observational study. Sci Rep. 2024; 14 (1): 17380.

Objective: to assess the effectiveness and acceptability of a pillow-like position modification device to reduce supine sleep during late pregnancy, and to determine the impacts on the severity of sleep-disordered breathing (SDB) and fetal wellbeing. Design: Randomised cross-over study Setting and population: Individuals in the third trimester of pregnancy receiving antenatal care at a tertiary maternity hospital in Australia. Methods: Participants used their own pillow for a control week and an intervention pillow for a week overnight, in randomized order. Sleep position and total sleep time for each night of both weeks was objectively monitored, with a sleep study and fetal heart rate monitoring performed on the last night of each week. Main Outcome Measures: Percentage of sleep time in the supine position, apnoea-hypopnoea index, fetal heart rate decelerations and birthweight. Results: Forty-one individuals were randomized with data collected on 35 participants over 469 nights. There was no difference in percentage of total sleep time in the supine position overnight between the control or intervention pillow week (13.0% [6.1, 25.5] v 16.0% [5.6, 27.2], p = .81), and no difference in the severity of SDB or fetal heart rate decelerations across weeks. However, increased supine sleep was significantly related to a higher apnea-hypopnea index (r s = .37, p = .003), lower birthweight (r s = -.45, p = .007) and lower customised birthweight centile (r s = -.45, p = .006). The proportion of supine sleep each night of the week varied widely both within and across participants, despite awareness of side-sleeping recommendations. Conclusions: The adoption of a pillow designed to discourage supine sleep was not effective in late pregnancy, with women spending an average of one hour per night supine. Supine sleep was associated with SDB and lower birthweight. Alternative devices should be investigated, incorporating lessons learnt from this study to inform trials of supine sleep minimisation in pregnancy. Clinical Trial Registration Number – ACTRN12620000371998 (Australia New Zealand Clinical Trials Registry)

Introduction: Stillbirth is a significant global public health issue, with approximately 98% occurring in low- and middle-income countries. The Solomon Islands is a Pacific nation with poor perinatal outcomes and very little previous research investigating stillbirth. Methods: We conducted a retrospective cohort study investigating all stillbirths at the National Referral Hospital in Honiara, Solomon Islands, between January 2017 and December 2018. Causes of stillbirth and risk factors were classified on review of available case files. Results: Over two years, there were 341 stillbirths and 11,056 total births at the National Referral Hospital (30.8 stillbirths per 1000 births). Cause of death was documented for 198 and 142 full case files were available. Most stillbirths occurred antenatally (n=170/198) and 62% were at preterm gestations (<37 weeks). Low birthweight (<2500g) was present in 59% (n=84/142) and preventable maternal conditions, including hypertensive disorders and syphilis, were present in 42% (n=59/142) of cases. Acute events caused 46% of intrapartum deaths and 92% of these had inadequate intrapartum monitoring. Conclusion: Our study is the first to investigate causes of stillbirth in the Solomon Islands. We found a large proportion of preventable stillbirths and significant gaps in documentation. This highlights the importance and feasibility of a national registry. There is an urgent need for targeted training in data collection, improved quality of antenatal and intrapartum care and community awareness to reduce preventable stillbirths in the Asia-Pacific.

Objective: Many growth charts provide single centile cutoffs for each week of gestation, yet fetuses gain weight throughout the week. We aimed to assess whether using a single centile per week distorts the proportion of infants classified as small and their risk of stillbirth across the week. Design: Retrospective cohort study. Setting: Victoria, Australia. Population: Singleton, non-anomalous infants born from 2005-2015 (529,261). Methods: We applied growth charts to identify small-for-gestational-age (SGA) fetuses on week-based charts (single centile per gestational week) and day-based charts (centile per gestational day). Main outcome measures: Proportions <10th centile by each chart, and stillbirth risk amongst SGA infants. Results: Using week-based charts, 12.1% of infants born on the first day of a gestational week were SGA, but only 7.8% on the final day; ie. an infant born at the end of the week was 44% less likely to be classed as SGA (p<0.0001). The relative risk of stillbirth amongst SGA infants born on the final day of the week compared with the first was 1.47 (95%CI 1.09-2.00, p=0.01). Using day charts, SGA proportions were similar and stillbirth risk equal between the beginning and end of the week (9.5% vs 9.9%). Conclusions: Growth standards using a single cutoff for a gestational week overestimate the proportion of infants that are small at the beginning of the week and underestimate the proportion at the end. This distorts the risk of stillbirth amongst SGA infants based on when in the week an infant is born. Day-based charts should be used

Objective: To compare emergency department (ED) presentations and hospital admissions for urgent early pregnancy conditions in Victoria before and after the onset of COVID-19 lockdown on 31 March 2020. Design: Population-based retrospective cohort study Setting: Australian state of Victoria Population: Pregnant women presenting to emergency departments or admitted to hospital Methods: We obtained state-wide hospital separation data from the Victorian Emergency Minimum Dataset and the Victorian Admitted Episodes Dataset from January 1, 2018, to October 31, 2020. A linear prediction model based on the pre-COVID period was used to identify the impact of COVID restrictions. Main outcome measures: Monthly ED presentations for miscarriage and ectopic pregnancy, hospital admissions for termination of pregnancy, with subgroup analysis by region, socioeconomic status, disease acuity, hospital type. Results: There was an overall decline in monthly ED presentations and hospital admissions for early pregnancy conditions in metropolitan areas where lockdown restrictions were most stringent. Monthly ED presentations for miscarriage during the COVID period were consistently below predicted, with the nadir in April 2020 (790 observed vs 985 predicted, 95% CI 835-1135). Monthly admissions for termination of pregnancy were also below predicted throughout lockdown, with the nadir in August 2020 (893 observed vs 1116 predicted, 95% CI 905-1326). There was no increase in ED presentations for complications following abortion, ectopic or molar pregnancy during the COVID period. Conclusions: Fewer women in metropolitan Victoria utilized hospital-based care for early pregnancy conditions during the first seven months of the pandemic, without any observable increase in maternal morbidity.

Objective: Vaginal birth after caesarean (VBAC) has been suggested to be associated with an increased risk of obstetric anal sphincter injury (compared with primiparous women who birth vaginally). However, prior studies have been small, or used outdated methodology. We set out to validate whether the risk of obstetric anal sphincter injury among women having their first VBAC is greater than that among primiparous women having a vaginal birth. Design: State-wide retrospective cohort study. Setting: Victoria, Australia. Population: All births (455,000) between 2009-2014. Methods: The risk of severe perineal injury between first vaginal birth and first vaginal birth after previous caesarean section was compared, after adjustment for potential confounding variables. Covariates were examined using logistic regression for categorical data and Wilcoxon rank-sum test for continuous data. Missing data were handled using multiple imputation; the analysis was performed using regression adjustment and Stata v16 multiple imputation and teffects suites. Results: Women having a VBAC (n=5,429) were significantly more likely than primiparous women (n=123,353) to sustain a 3rd or 4th degree tear during vaginal birth (7.1 vs 5.7%, p<0.001). After adjustment for mode of birth, body mass index, maternal age, infant birthweight, episiotomy and epidural, there was a 21% increased risk of severe perineal injury (relative risk 1.21 (95%CI 1.07 – 1.38)). Conclusions: Women having their first vaginal birth after caesarean section have a significant increased risk of sustaining a 3rd or 4th degree tear, compared with primiparous women having a vaginal birth. Patient counselling and professional guidelines should reflect this increased risk.

Objective: Identify the proportion of infants reclassified if sex-specific birthweight charts were used, and if this reclassification has an impact on the correlation between birthweight centile and adverse perinatal outcome. Design: Retrospective cohort study Setting: Victoria, Australia. Population: All infants born from 2005-2015 (529,261) Methods: We applied GROW centiles, either adjusted or unadjusted for fetal sex. We compared proportions of small for gestational age (SGA, <10th centile) infants, then the populations of males considered small only by sex-specific charts and females considered small only by unadjusted charts. Main Outcome Measures: Stillbirth, combined perinatal mortality, NICU admissions, Apgars <7 at 5 minutes, emergency caesarean sections. Results: Of those <10th centile by unadjusted charts, 39.6% were male, and 60.5% female. Using sex-specific charts, 50.3% <10th centile were male and 49.7% female. 9,449 (19.2%) females that were SGA according to unadjusted charts were appropriate for gestational age (AGA,>10th-<90th centile) using sex-specific charts. These reclassified newborn females were not at increased risk of adverse outcomes compared with an AGA infant, but were at increased risk of being iatrogenically delivered for suspected growth restriction (RR 4.90, 95%CI 4.39–5.48). 8,048 male infants were reclassified as SGA by sex-specific charts (25% SGA increase). Compared with AGA infants, these reclassified male newborns were at greater risk of stillbirth (RR 1.94, 95%CI 1.30-2.90) and all other adverse perinatal outcomes. Conclusions: Sex-specific growth standards classify a new high-risk cohort of male infants as SGA, and exclude a cohort of females, whose risk is no greater than appropriately grown infants.

Objective The My Baby’s Movements (MBM) trial aimed to evaluate the impact on stillbirth rates of a multifaceted awareness package (MBM intervention). Design Stepped-wedge cluster-randomised controlled trial. Setting Twenty-seven maternity hospitals in Australia and New Zealand. Population Women with a singleton pregnancy without major fetal anomaly ≥28 weeks’ gestation from August 2016-May 2019. Methods The MBM intervention was implemented at randomly assigned time points with sequential introduction into 8 clusters of 3-5 hospitals at four-monthly intervals. The stillbirth rate was compared in the control and intervention periods. Generalised linear mixed models controlled for calendar time, clustering, and hospital effects. Outcome Measures Stillbirth at ≥28 weeks’ gestation. Results There were 304,853 births with 290,219 meeting inclusion criteria: 150,079 in control and 140,140 in intervention periods. The stillbirth rate during the intervention was lower than the control period (2.2/1000 births versus 2.4, odds ratio [OR] 0.91, 95% Confidence Intervals [CI] 0.78-1.06, p=0.22). The decrease was larger across calendar time with 2.7/1000 in the first 18 months versus 2.0/1000 in the last 18 months (OR 0.74; 95% CI 0.63-0.86; p≤0.01). Following adjustment, stillbirth rates between the control and intervention periods were not significantly different: (aOR 1.18, 95% CI 0.93-1.50; p=0.18). No increase in secondary outcomes, including obstetric intervention or adverse neonatal outcome, was evident. Conclusion The MBM intervention did not reduce stillbirths beyond the downward trend over time, suggesting hospitals may have implemented best practice in DFM management outside their randomisation schedule. The role of interventions for raising awareness of DFM remains unclear