select search filters
briefings
roundups & rapid reactions
before the headlines
Fiona fox's blog
A population cohort analysis published in PLOS ONE looks at whether prescriptions for asthma medications, whether continued or discontinued during pregnancy, are associated with premature births, stillbirths and birth weights in babies.
Prof Seif Shaheen, Clinical Professor of Respiratory Epidemiology, Queen Mary University of London, said:
“In this observational study the authors have used routine data to investigate whether the use or discontinuation of prescribed asthma medication during pregnancy is associated with adverse perinatal outcomes, including premature birth and low birth weight. The use of routine ‘big’ data is becoming very popular in epidemiology, but an important limitation of this approach, if one is trying to determine whether particular risk factors might have an independent role in causing an outcome of interest, is that data on important confounders are often missing. In this study adjustment has been made for socioeconomic status using only a crude, area-based measure, rather than an individually based measure such as educational attainment; thus residual confounding by social background is likely. Another important potential confounder, maternal obesity, which is a risk factor for both asthma in the mothers and also adverse perinatal outcomes, has not been controlled for at all, because data on body mass index were missing for a high proportion of women. Whether or not asthma or asthma medications per se are likely to influence perinatal outcomes is therefore very uncertain, based on this study.”
Prof Andrew Shennan, Professor of Obstetrics, King’s College London, said:
“Nearly 1 in every 10 pregnant women have asthma requiring medication, and many are concerned about the safety of these medications when pregnant. This population-based study confirms that pregnancy outcomes are worse with asthma. However, early birth and less breast feeding were more likely in those who discontinued treatment, stressing the importance of continuing treatment in pregnancy. More than 1 in 4 women stopped treatments in this study, illustrating this is a common issue. The study confirmed that the commonly prescribed inhalers were safe. Women should continue medication in pregnancy, but seek medical advice before and early in pregnancy to ensure the best medication is being used and appropriately monitored.”
Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:
“This is a well-conducted study, in statistical terms. The researchers have done a good job. There are some difficulties in interpreting precisely what the findings mean, but that isn’t the fault of the researchers, who have been scrupulously open about what they have and haven’t been able to show. The difficulties arise from unavoidable problems in researching questions like this.
“It’s an observational study – that is, the researchers had no input into what the mothers in the study did. The mothers, and their doctors, made the choices of what medication to take, just as if they were not in this study – in fact the data were collated and analysed after all those choices had been made and the outcomes of the pregnancies were already known. The researchers simply used records of what had happened to make their analysis. That’s really the only possibility for studying associations between asthma, and particular asthma medications, and pregnancy outcomes. Mothers clearly can’t be assigned by the researchers to have asthma or not, and it wouldn’t be ethical for researchers to allocate specific asthma treatments to the mothers, because that might potentially harm the mothers (by making their asthma worse) or harm their children. So the study has to be observational. This means that there are inevitably many differences between women with and without asthma, and women with asthma who took different medications (or no medication at all), apart from the asthma and the medications taken. Any of these other differences could, in part or completely, be the real cause of any difference in pregnancy outcomes between people taking different medications, and not the medications (or the asthma) at all. So all that can be reported is that there are associations between some of the treatment characteristics and the pregnancy outcomes, and not that the treatments cause the outcomes. The researchers make this very clear.
“A particular issue with any study of this nature is what’s called “confounding by indication”. There will always be reasons why some people are prescribed one medication and others another medication, and these reasons may be part or all of what causes any differences in pregnancy outcomes between the mothers who took different asthma medications, rather than any direct effects of the medications themselves. The researchers investigated possible aspects of confounding by indication, by also including women who stopped taking asthma medication while they were pregnant, but they (rightly) report that one can never be sure that confounding by indication is no longer present. They write that “We acknowledge the difficulties of disentangling the effects of asthma from its management.” They also carried out various statistical adjustments to allow, as far as they could, for other differences between the women who took different medications, in known risk factors for premature birth, possible growth restrictions for the baby, and discontinuation of breast feeding. But it’s impossible to be sure that everything important has been allowed for, which is why this study can tell us about associations, and not about exactly what causes what. However, because they did find associations between asthma treatments (or no treatment at all) and prematurity, stillbirths and low birth weight, their advice to consider increasing monitoring and support for pregnant women with asthma is important.
“Because the researchers studied a very large proportion of all pregnancies in Wales between 2000 and 2010, their data can give a fair idea of the risks of the outcomes in that population – though women that were known for various reasons (not directly connected to asthma) to be at particular risk of a poor pregnancy outcome were not included in the data analysis, so the risks in the Welsh population as a whole would be a bit higher than the figures here. For women who did not have asthma at all, about 57 in every thousand pregnancies in this group gave birth at 37 weeks or earlier. The only really firm evidence for a difference in this rate, for people with asthma, is for those who took no medication because they discontinued it during pregnancy. For them, the study indicates that about 68 in every thousand might give birth at 37 weeks or earlier, but there’s quite a lot of statistical uncertainty about this figure – it could plausibly be anywhere between 60 per thousand and 78 per thousand, that is, between 3 and 21 per thousand more than the women without asthma. This indicates that having asthma but not taking medication during pregnancy is associated with an increased risk of a birth at 37 weeks or earlier – but the risk is in any case not all that high, and the increase is also not large. And this is an association – the study can’t definitely tell us that the unmedicated asthma causes the increased risk.
“For women who did not have asthma, about 8 in every thousand pregnancies resulted in a birth at 32 weeks or earlier, which is obviously much more premature than the 37 weeks figure, but also much less common. Again the clearest evidence of a difference in women with asthma is in those who discontinued their medication, where the number of births at 32 weeks or before could be between 1 and 9 greater than the number in women without asthma. However, in this case, there is also some evidence of associations of an increased risk of a birth at 32 weeks or before, for women that took certain asthma medications, though most of this evidence is fairly weak statistically. For all women with asthma (whatever their medication), the extra number of births this early in pregnancy was plausibly between about 1 and about 5 per thousand births. Again a fairly small risk and a fairly small increase in risk, and there’s no way to be sure about cause and effect – but births this early in pregnancy can have a severe effect on the quality of life of the child.
“The researchers also found associations between having been prescribed asthma medication and having a stillbirth, though the risk of stillbirth is small. Of women without asthma in this study, only about 4 in a thousand experienced this sad event. For women with asthma, the study estimated that there would be between about 1 and about 4 extra stillbirths in every thousand births, and again that association was rather stronger if the medication had been discontinued during pregnancy. And again, we cannot be sure that the asthma is the cause of the stillbirths.
“The study found rather weaker evidence of associations between asthma, or taking certain asthma medications, and giving birth to babies who weighed less than they should have (given the timing in pregnancy when they were born). It found only extremely weak evidence of any association between asthma and having discontinued breast feeding by 6-8 weeks after birth.”
‘Medicines prescribed for asthma, discontinuation and perinatal outcomes, including breastfeeding: A population cohort analysis’ by Gareth Davies et al. was published in PLOS ONE at 19:00 UK time on Wednesday 9 December 2020.
DOI: 10.1371/journal.pone.0242489
Declared interests
Prof Seif Shaheen: “I have no COI.”
Prof Andrew Shennan: “I have no conflicts in this matter.”
Prof Kevin McConway: “I am a Trustee of the SMC and a member of the Advisory Committee, but my quote above is in my capacity as a professional statistician.”