Children exposed to macrolides during the first trimester of pregnancy had an increased risk of major malformations, compared with first-trimester penicillin exposure, according to an observational study.
Creatas Images
Use of antibiotics is common in pregnancy, and macrolides commonly are used if a penicillin allergy is reported.
Hypospadias and other genital malformations also were more likely with exposure during any trimester to macrolides, although this association lost significance when limited to the first trimester. The researchers did not identify any associations with macrolides exposure and neurodevelopmental disorders.
The observational study could not establish causality, but the researchers calculated an estimate of likely excess malformations if the association were found to be causal: “For every 1,000 mothers prescribed macrolides instead of penicillins during the first trimester, an additional 4.1 children would have cardiovascular malformations,” Heng Fan, a PhD student at the University College London, and colleagues wrote in the BMJ. “The corresponding figures for prescriptions during any trimester and genital malformations would be 1.7.”
The researchers used records from the U.K. Clinical Practice Research Datalink to analyze outcomes in 104,605 children born between 1990 and 2016 to mothers who received at least one prescription of erythromycin, clarithromycin, azithromycin, or penicillin monotherapy between their fourth week of pregnancy and delivery. Women prescribed any known teratogenic medications were excluded.
The majority of the mothers (92%) had been prescribed penicillin once, and 8% were prescribed a macrolide antibiotic once during pregnancy.
The researchers tallied and calculated the children’s risk of major malformations; cerebral palsy; epilepsy; ADHD; autism spectrum disorder; and any nervous, cardiovascular, gastrointestinal, genital, or urinary malformations. The children were tracked through a median 6 years of age.
In comparing risk of malformations or neurodevelopmental disorders among children, the researchers chose to compare exposure to macrolides and penicillin to reduce the likelihood of confounding by indication for infections. (They also included two negative control groups: unexposed siblings and women prescribed antibiotics before conception.) The authors acknowledged, however, that residual confounding still may occur “if macrolides were prescribed for specific indications (e.g., chlamydia), or when potential risk factors for malformations or neurodevelopmental outcomes differed between treatment groups.”
The overall rate of malformations was 22 per 1,000 children prenatally exposed to macrolides (28 in first trimester and 20 in second or third trimester) and 17 per 1,000 children prenatally exposed to penicillin. The risk and type of malformations varied, however, according to the trimester.
The researchers made adjustments to account for differences in a wide range of maternal factors: age at delivery, calendar year of delivery, alcohol misuse, illegal drug use, tobacco use, obesity, hypertension, diabetes, anxiety, depression, and epilepsy. They also adjusted for parity, multiples, and chronic medical treatments, as well as genitourinary tract infections or STIs during pregnancy, both of which are linked to preterm labor.
Compared with children exposed to penicillin during the first trimester of pregnancy, risk of malformations was 1.6 times greater in those exposed to macrolides in the first trimester (risk ratio, 1.55; 28 vs. 18 per 1,000). Erythromycin exposure in the first trimester also was linked to a 50% greater likelihood of any major malformation compared with penicillin (RR, 1.5; 27 vs. 18 per 1,000).
Cardiovascular malformations in particular were more likely in those exposed to macrolides (11 per 1,000), compared with penicillin (7 per 1,000) in the first trimester (RR, 1.62). Meanwhile, genital malformations, primarily hypospadias, occurred more frequently in children whose mothers were prescribed macrolides (5 per 1,000), compared with penicillin (3 per 1,000) in any trimester (RR, 1.58).
No increased risk of major malformations was associated with macrolides prescribed only in the second or third trimester, although a borderline significant association existed with gastrointestinal malformations. The authors also found no links between macrolides exposure and increased risk of cerebral palsy, epilepsy, ADHD, or autism spectrum disorder.
The findings did not change in several sensitivity analyses, including one that restricted analysis to antibiotics prescribed only for respiratory tract infections.
Dr. Fan and associates discussed several potential biological mechanisms for causation, including the arrhythmic effect of macrolides that may relate to cardiovascular malformations or contribute to fetal hypoxia. They noted that “macrolide prescribing during pregnancy warrants caution,” and recommend including on drug safety labels “that there is uncertainty about the safety of macrolides, including erythromycin” and alternative antibiotics should be used when possible.
Iris Krishna, MD, MPH, assistant professor of maternal-fetal medicine at Emory University, Atlanta, agreed with the study authors that use of macrolides in the first trimester warrants further investigation, and if an appropriate alternative antibiotic is available, then it should be preferentially considered when treating infections in the first trimester.
“However, if macrolides are the only treatment option, pregnant women can be reassured that the absolute risk of a birth defect is low, and this should not discourage them from taking a macrolide when needed as untreated infections pose a greater risk in pregnancy,” she said in an interview.
“This study does not establish that macrolide antibiotics cause birth defects, but it suggests a potential association. Previous studies examining the use of macrolides, such as erythromycin, have not demonstrated a consistent pattern of birth defects, and heart defects identified were classified as mostly mild. The authors suggest that the potential biologic mechanism based on rat models may be that macrolides might induce fetal cardiac arrhythmias and short-term fetal hypoxia. This study was underpowered to examine macrolide exposure for specific malformations. To avoid underpowered comparisons, the authors’ categorized malformations by organ systems, so the spectrum of cardiac defects is unclear,” commented Dr. Krishna, who also is a member of the Ob.Gyn. News editorial advisory board.
“Current recommendations for macrolide antibiotic use in pregnancy in the second and third trimester of pregnancy, and in particular when used for obstetric indications, such as prelabor rupture of membranes to prolong the latency period to delivery, should not be altered based on the findings of this study,” she concluded.
The research was funded by Child Health Research CIO Trust, the China Scholarship Council, Health Data Research UK, and the National Institute for Health Research. Dr. Fan and associates had no industry disclosures. Dr. Krishna had no relevant financial disclosures.
SOURCE: Fan H et al. BMJ. 2020;368:m331.