Journal club papers
Antenatal Dexamethasone for Early Preterm Birth in Low-Resource Countries. WHO Collaborators. (10/2020)
Steroids are key to reducing preterm infant morbidity and mortality. However, it is unclear whether this is also the case in low resource settings because a cluster randomised controlled trial published in 2015 found an unexpectedly increased risk of neonatal death, stillbirth and maternal infection. There has thus been continued controversy as to whether steroids should be offered in low resource settings – in most parts of the globe, does it save babies or harm them?!?
- This new multi-country randomised trial recruited 2852 women from 26 – 34 weeks gestation at risk of preterm birth across 6 countries; Bangladesh, India, Kenya, Nigeria, and Pakistan.
- Women were randomised 1:1 to 6 mg dexamethasone or placebo every 12 hours for a maximum of 4 doses or until discharge or birth.
- The trial was stopped early after interim analysis revealed a substantial benefit of dexamethasone treatment.
- Neonatal death occurred in 19.6% (278/1417) of infants in the dexamethasone group vs 23.5% (331/1406) in the placebo; reducing the risk of neonatal death by 16% (relative risk 0.84; 95% CI 0.72, 0.97).
- To prevent 1 neonatal death, 25 (95% CI 14, 110) cases need to be treated with dexamethasone.
- Dexamethasone reduced the risk of stillbirth or neonatal death by 12% (RR 0.88; 95% CI 0.78, 0.99; 25.7% vs 29.2%).
- Reassuringly, the incidence of possible maternal bacterial infection was also lower in the dexamethasone group (4.8% vs 6.3%; RR 0.75; 95% CI 0.56, 1.03) and there was no increase in neonatal hypoglycaemia.
This is excellent news. It confirms the benefit of corticosteroids in low resource settings for infants destined to be born preterm.
Why the divergent findings from 2015 trial? In the earlier trial, gestational age was assumed from LNMP recall OR clinical assessment of uterine size- both comparatively poor proxies for preterm gestation. It may have recruited more mature fetuses with growth restriction, a population arguably with ‘more to lose than gain’ with corticosteroids - at lower risk of respiratory distress but higher risk of infection/ other morbidities. In contrast, the current study recruited women who presented at 26+0- 33+6 with planned or imminent preterm birth, with ultrasound confirmation of gestational age. Further, a minimum level of neonatal care (access to oxygen, CPAP and oximetry) was ensured for all trial centres.
- In the Netherlands, restriction measures were first implemented on March 9, with the highest level of restrictions in place by March 23.
- This study used the national newborn screening program to investigate the impact of COVID-19 restriction measures (March 9, 15 and 23rd) on the incidence of preterm birth.
- From 2010 – 2020, 1,599,547 singleton births occurred in the Netherlands, with 56,720 after implantation of restrictions (March 9th).
- Compared to the previous years, the overall incidence of preterm birth was reduced by around 15-23% during for the 2 (ie 2 months before vs two months after), 3 and 4 months surrounding March 9: 2 months odds ratio (OR) 0.77 (95% CI 0.66, 0.91); 3 months OR 0.85 (95% CI0.73, 0.98); 4 months OR 0.84 (95% CI 0.73, 0.97).
- This reduction in preterm birth was driven by births >32 weeks.
- Note that the authors used a rather complex statistical approach (which the JCOTR team do not completely follow, but welcome the reader to take a look and enlighten us). It somehow uses birth data over 10 years to ‘virtually’ add far greater numbers to the 2,3,4 month comparisons, so there is sufficient power to find a possible difference.
Unprecedented reduction in births of very low birthweight (VLBW) and extremely low birthweight (ELBW) infants during the COVID-19 lockdown in Ireland: a ‘natural experiment’ allowing analysis of data from the prior two decades. Phillip RK et al. BMJ Global Health (08/2020)
- COVID-19 triggered a national lockdown in Ireland from March 12– May 18.
- In this study, the lockdown was used to explore whether socioenvironmental and maternal behaviour modification may impact on the incidence of very preterm birth in one province of Ireland
- Very low birthweight (VLBW <1500g) and extremely low birthweight (ELBW <1000g) were used as indicators of preterm birth.
- Trends from Jan – April 2001 – 2019 were compared to January – April 2020.
- The average VLBW rate per 1000 live births from 2001 – 2019 (Jan – April only) was 8.18 (95% Wald CI 7.21, 9.29), compared to 2.17 per 1000 live births (95% Wald CI 0.70, 6.74) during Jan-April 2020. This is a 73% reduction in the forecasted rate of VLBW infants.
- No ELBW infants were born during the 2020 period, compared to an average Jan – April rate of 3.0 per 1,000 live births in the previous two decades.
- The rate ratio comparing the risk of VLBW 2001 – 2019 to 2020 was 3.22 (95% Wald CI 1.21, 11.75).
- No increase in early pregnancy loss, miscarriage or stillbirth during the lockdown period was found, suggesting that the observed reduction in VLBW and ELBW infants was not due to a higher mortality rate.
- Denmark also experienced a national lockdown from March – April 2020.
- This study investigated preterm birth in a nationwide prevalence proportion study, where infants born during the lockdown period were compared to March – April of the previous 5 years (2015 – 2019).
- 31,180 live singleton infants were included in the study, with 1,566 (5.0%) born prematurely.
- Compared to births during March – April of the 5 years prior, the proportion of extremely (<28 weeks) and very premature (28 – 32 weeks) infants was significantly reduced during lockdown (p-value 0.003).
- This difference was largely attributed to a reduction in extreme prematurity (0.19 per 1000 births vs 2.19 per 1000 births).
- As the number of intrauterine deaths across the study period were not examined, we cannot exclude the possibility that the difference in extreme prematurity may be attributed to a larger than usual number of pregnancies resulting in intrauterine death.
This intriguing suite of studies have shown an association between a period of COVID lockdown and a reduced incidence of preterm birth.
While they are interesting observations, we do not think the message is that we should now proceed to lockdown all women while they are pregnant, for a number of important reasons.
Firstly, these are associations. We haven’t really garnered insight whether it was the lockdown itself, or some behaviour associated with lockdown.
Secondly, we do not have a complete picture on overall obstetric outcomes. For instance, lockdown (or recommending pregnant stay at home most of the time) may incur the price of reduced detection of fetal growth restriction and preeclampsia which could increase the stillbirth risk.
Thirdly, we do not have data on maternal health outcomes. No one enjoyed the lockdown and there may be an unacceptable cost to mental health, home finances, and potential strains to relationships.
Lastly, the ‘numbers needed to treat’ to prevent one case of preterm birth is likely to be very high. We would need to have many stay at home to prevent one case. In contrast, it is plausible that the numbers needed to ‘harm’ (such as adverse mental health impacts) are likely to be alarmingly low.
For these reasons we do not advocate clinicians start advising women remain locked down during their pregnancy.