Some antibiotics, when used in pregnancy, could cause birth defects.
New research has found a link between macrolide antibiotics - such as erythromycin, clarithromycin and azithromycin - and an increased risk of heart and genital malformations.
The drugs are commonly used when patients are allergic to penicillin.
Scientists in the UK looked at data from more than 180,000 children born in the UK between 1990 and 2016. Of those, 8632 were children born to mothers who took macrolide antibiotics during their pregnancy, and 95,973 penicillin.
The scientists found compared to children whose mothers were prescribed penicillin in the first trimester, those exposed to macrolide antibiotics were more than 50 percent more likely to have "any major malformation... and specifically cardiovascular malformations".
There appeared to be no increased risk from taking either macrolide antibiotics in the second and third trimesters, compared to penicillin.
The overall numbers of defects were low - 28 per 1000 for macrolide antibiotic exposure in the first trimester, compared to 18 for penicillin, 11 and seven for cardiovascular malformations, and five and three for genital malformations.
But the scientists behind the research, published on Thursday in The BMJ, say doctors and mums-to-be should exercise caution.
"If feasible, alternative antibiotics should be prescribed until further research is available."
MedSafe's advice on erythromycin, clarithromycin and azithromycin says there have been, until now, no decent research on the effects of macrolide antibiotics on pregnancy. It also notes while animal studies have been done, those findings don't always match what happens in humans.
"The physician should not prescribe clarithromycin to pregnant women without carefully weighing the benefits against risk, particularly during the first three months of pregnancy," Medsafe's advice on clarithromycin says, for example, and "erythromycin should be used by women during pregnancy only if clearly needed".
It's not clear why macrolide antibiotics carry more risk than penicillin, the researchers said, as their study was only observational - it can see the effects, but not explain them.
They also looked at possible links between macrolide antibiotics and neurodevelopmental disorders such as autism, but found none.
