EXPERT REACTION: Some antibiotics prescribed during pregnancy linked with birth defects

The highly respected BMJ has published a paper in what is becoming a disturbing trend in medical research.

Data dredging of databases involving large numbers of pregnant women is being used to look for an association between common medical treatments and adverse pregnancy effects. 

These observational type studies are not sufficiently powered to prove causation and can unnecessarily fuel anxiety in pregnant women and sometimes even lead to denial of what can be important treatment options.

The difficulty separating association from causation is a well-known challenge when undertaking medical research and this research is no different.

The study released today, suggests a link between pregnant women having been prescribed a particular class of antibiotics and significant congenital anomalies in their babies. It found that first trimester exposure was associated with an increased risk of cardiovascular anomalies and exposure during the second or third trimester associated with an increased risk of genital anomalies.

The authors do go to pains to point out that the study was an analysis of data from an 'administrative database and did not directly involve patient participants'.

The macrolide class of antibiotics (erythromycin most commonly but includes azithromycin and clarithromycin) is important because they frequently used to treat women who are allergic to penicillin and sometimes the only class of drugs that particular bacteria are sensitive to or that are recommended by professional bodies. E.g. Erythromycin is the only antibiotic that has been shown to lengthen the duration of pregnancy in preterm women whose membranes have ruptured.

Recommendations regarding the use of this group of antibiotics vary around the world and there has previously been published, conflicting research around pregnancy exposure being associated with increased congenital malformations including cerebral palsy and epilepsy.

Until further research clarifies this question, women and doctors should observe what has always been best practice when it comes to the prescription of medication including antibiotics, during pregnancy.

The benefits of using a drug should be weighed against the risks to the mother and developing baby of not providing treatment and where a safer and available alternative exists, that drug should be preferred.

The study by Fan et al. reported in the BMJ today raises some important questions about the use of a particular group of antibiotics (the macrolides such as erythromycin, clarithromycin and azithromycin) in both early and late pregnancy. Whilst the overall rates are still relatively low, the study suggests a significant increase in the risk of major congenital malformations (especially cardiovascular) associated with use of these antibiotics in the first trimester, and an increase in genital tract malformations such as hypospadias with use later in pregnancy. The authors call for international collaborations to extend their work using larger datasets to see if the results are confirmed. The use of all drugs in pregnancy should always be considered to be a balance of risks and benefits, but where there may be choices in the type of antibiotic used, information from studies such as this can be helpful in guiding clinical practice. Until further evidence is available it would seem prudent to avoid where possible the use of macrolides in the first trimester of pregnancy, and to exercise caution in their use in later gestations, carefully balancing the benefits with these possible risks of causing harm to the fetus.

This study describes the effects on children in the UK when one of two types of antibiotic (the macrolide or penicillin classes) had been given to their mother during pregnancy.

It examined over 100,000 children whose mother was given one of these antibiotics, and compared them to over 135,000 children whose mother was not exposed to an antibiotic during pregnancy.

The researchers found that the macrolides, but not the penicillins, lead to an increase in the number of malformations in parts of the body, particularly the cardiovascular system, if they were taken during the first trimester.

The overall levels were still low, rising from a risk of 18 cases in every 1,000 children for no antibiotic exposure, to 28 in 1,000 if exposed to a macrolide. 

No effects were seen in other outcomes assessed, including levels of autism, cerebral palsy, epilepsy or ADHD.

Notably, this study only included children whose mothers were prescribed an antibiotic for a respiratory tract infection. These are one of the most common causes of unnecessary antibiotic use, as they are usually caused by viruses, which are not affected by antibiotics.

It highlights that antibiotics should only be taken when needed - they are not harmless and do have side effects. You don’t see healthy people taking anticancer drugs 'just in case', but for some reason it is acceptable for antibiotics.

The study adds to the evidence of potential long term detrimental outcomes - for example another study in 2018 (International Journal of Epidemiology, 2018, 561–571 doi: 10.1093/ije/dyx272) showed that the offspring of mothers who took antibiotics were more likely to be hospitalised due to an infection.

This study by Heng Fan and colleagues examines the risks associated with the use of macrolide antibiotics in pregnancy and congenital malformations, such as heart defects, in offspring. 

The macrolide class includes widely used antibiotics such as Erythromycin, Clarithromycin and Azithromycin. These antibiotics are often prescribed to treat common bacterial genito-urinary, upper respiratory and soft tissue infections. They have a slightly broader spectrum of effectiveness, compared to penicillin, and are often used for similar indications in patients who are allergic or sensitive to penicillin. 

The researchers examined a large clinical database in the UK. They identified more than 1 million babies borne over the 26 years period. Then they looked at which drugs were prescribed to the mothers while they were pregnant; specifically, those who were prescribed either a course of penicillin or one of the macrolides, once in their pregnancies. The safety of penicillin in pregnancy is well established and by comparing outcomes of children who were exposed to penicillin in utero to the ones who were exposed to macrolides, researchers eliminated several biases and made their conclusions much more robust and reliable. 

About 1 in 10 pregnancies in the study group of over one million were exposed to either penicillin or macrolides. The majority were prescribed Penicillin - 92 per cent -, while macrolides were prescribed to 8 per cent of pregnant mothers. To put this in perspective, in the more than 1 million pregnancies included in the analysis, about 90 in 1,000 were prescribed Penicillin and 10 in 1,000 were prescribed one of the macrolide antibiotics. 

The results were examined based on the stage of pregnancy at which the antibiotics were prescribed. The major significant finding of this study is that the use of macrolide antibiotics in the first trimester, between the beginning of pregnancy and 13 weeks, is associated with an increased risk of cardiovascular malformations. The magnitude of this increased risk is in the order of about 60 per cent. 

While this is a significant and clinically important finding, an absolute risk of congenital, and specifically cardiac, malformations was found to be small. The authors confirmed that penicillin use does not increase the risk of congenital malformations. Nevertheless, it is important to stress that congenital malformations may be caused by a variety of factors, many of which are still unknown. It is not, therefore, surprising that in the penicillin-exposed group, the risk of congenital malformations was not zero, but rather was consistent with the background population risk of about 1.7 per cent. The use of macrolides was associated with an increased absolute risk of about 2.2 per cent. This effect was stronger if macrolides were used in the beginning of pregnancy, before 13 weeks gestation, where the risk rose to 2.8 per cent. These findings are consistent with the period of major organ development (organogenesis) taking place in the first 13 weeks of pregnancy. This is the time in pregnancy when the foetus is most vulnerable to the detrimental effects of maternal exposure to teratogenic substances (substances that may cause congenital malformations).

Other significant findings that deserve a mention include a small increase of genital malformations in the macrolide exposed pregnancies, especially in male offspring, mostly confined to the specific condition of hypospadias. This also confirmed previous studies that indicate that genital development may be susceptible to disruptive effects, including in late pregnancy.

On a positive note, it was reassuring to see that exposure to macrolides in pregnancy does not increase the risk of neurodevelopmental conditions, including cerebral palsy, epilepsy, attention deficit hyperactivity disorder (ADHD) and autism spectrum disorder (ASD).

Wider implications
Pregnancy, especially the period of organ formation, usually designated as the first trimester, is a period of unique vulnerability of a foetus to the disruptive influence of medications that may be prescribed to a pregnant woman. Therefore, the general advice should be to use medications during pregnancy only if clinically indicated. This population-based study demonstrates a possible impact of common antibiotics, used in pregnancy to treat a variety of ailments, on the risk of congenital malformations, specifically of the cardiovascular system. While the study cannot prove that macrolides cause these malformations, the association is sufficiently strong, and the methods used in the study are robust enough, to advise caution. Macrolides should only be used in pregnancy if no safer alternatives are available and after a full and frank discussion with a prospective mother about the perceived benefits and possible risks. Women who have taken macrolides before discovering that they are pregnant can be reassured that even though there appears to be an increased risk of cardiovascular congenital malformations, the absolute risk of these abnormalities remains small.