American physicians prescribe too many antibiotics for the pregnant woman and the fetus in utero. Practice modifications in 1996 and revisions in 2002 to prevent group B streptococcus (GBS) infections in newborns, and in 2010 to reduce the incidence of postpartum maternal infection after caesarean section, have led to the use of predelivery antibiotics in ≥40% of women in labour (and hence to the fetus). Before 1996, few of these children in utero would have been exposed to antibiotics. These obstetric strategies contrast with concerns about overuse of antibiotics in children, and run counter to the Centers for Disease Control campaign that has reduced children's antibiotic prescription rates for office-related visits by 24%. Should these increases in antibiotic prescribing by obstetricians be a concern? We think so.
The immediate impact of these two strategies has been favourable. The incidence of disease due to GBS in the newborn is reduced when predelivery antibiotics are administered, and the frequency and severity of infection in mothers after caesarean section is also reduced, with few immediate complications. Maternal allergic reaction to these prophylactic antibiotics is rare, and the development of resistant bacterial infections in either mothers or newborns has not as yet been a focus in the literature.
A belief in the long-term safety of predelivery antibiotics has been maintained, in part because meaningful prospective statistics are not being collected. In the current reality of early discharge from the hospital, many mothers with postpartum infection have their first symptoms at home. When seen most often by a doctor in their office, a draining abdominal wound is labelled a ‘seroma’ and the woman is treated with oral antibiotics without a culture. As most women get better with this approach, only the rare woman with treatment failure requires hospitalisation. Long-term changes in bacteria causing postoperative infections will not be easily detected with this practice pattern.
A similar sense of the universal effectiveness of predelivery antibiotics for the fetus is undeserved. In most community hospitals, very low birthweight babies are transferred to tertiary centres where their care is not observed by the transferring physician. Outcomes are not always apparent. At least two studies have shown an increase in Escherichia coli sepsis in very low birthweight newborns, whose mothers received antibiotics to prevent newborn GBS infection. This trend has not been reported in term babies.
There are other potential concerns about predelivery use of antibiotics. Predelivery antibiotics affect the bacterial populations of the mother's birth canal and skin, which will be transmitted to the babies during and following delivery. Changes in the composition of the indigenous microbiota of newborns have the potential to influence childhood development and disease risk. The rapid increase in recent years of illnesses with onset in childhood (including asthma, type 1 diabetes, obesity and autism) suggests that an environmental cause could be present. The loss of one or more constituents of the indigenous microbiota after maternal antibiotic exposure could be a contributing factor; this hypothesis should be studied.
Changes in early life microbiota of children may matter. (For example, the loss of Helicobacter pylori, disappearing from the gastric microbiota in mice, has resulted in decreases in gastric T-cell populations.) This change provides a basis for the increases now being seen in childhood asthma, allergic rhinitis and skin allergies. Currently, few young children have H. pylori-mediated regulation of gastric adipokines, such as ghrelin and leptin, at a time in life when long-term adiposity is being programmed; H. pylori removal increases postprandial ghrelin levels.
In another example, farmers add antibiotics to the food and water of young animals for ‘growth promotion’. In that setting, antibiotics are changing early life metabolism; is this an analogue for how we are treating our children and their mothers just before birth? A recent study by Cho et al. showed that subtherapeutic antibiotic therapy to young mice increases adiposity. There were substantial taxonomic changes in the microbiome, changes in copies of genes involved in the metabolism of carbohydrates to short-chain fatty acids, increases in colonic short-chain fatty acid levels, and alterations in the regulation of the hepatic metabolism of lipids and cholesterol. Clearly, antibiotic use in farm animals for growth, now banned in the USA, involved long-term use of low-dosage antibiotics. Predelivery antibiotics are short-term, but are given at a critical time when newborn acquisition of gut bacteria is just beginning. With the current (and growing) use of antibiotics affecting early life development, are we spawning a population of children at risk of being less healthy than previous generations? These issues should trigger discussion about current guidelines for care of pregnant women. What can we do?
One approach would be to develop safe strategies that limit the use of antibiotics in women in labour. In the preguideline era, GBS sepsis in the highest risk newborn population (very small premature infants) occurred in 5.9 cases per 1000 population. To protect one small premature child from GBS infection, over 100 others are exposed to antibiotics. It is also clear that risk of infection differs. For example, the incidence of GBS infection in the newborn is greater in women who have previously delivered an infant with a GBS infection, have a history of first-trimester GBS bacteriuria, are delivering a premature (<37 weeks of gestation) baby, have prolonged (>18 hours) rupture of membranes, and/or are African-American, compared with women with uncomplicated pregnancies delivering at term. Also, whether or not GBS is present in the vaginal or intestinal tract, women who undergo elective caesarean section who are not in labour, and whose membranes are intact are at very low risk, and antibiotic prophylaxis had not been recommended for this group. Instead, it is now recommended in the USA to prevent postpartum maternal infection, and in consequence is now always being given.
For women at term with a positive GBS culture at 35–37 weeks, alternatives to maternal antibiotics should be considered. Although not adequately studied, a dilute hexachlorophene vaginal douche at the time of admission for labour is one alternative, but this strategy is flawed, for it still affects the birth canal microbiota. A possibly more appealing strategy is the performance of a rapid (polymerase chain reaction; PCR) testing for GBS presence at the time of admission to the labour and delivery unit. A negative PCR would eliminate antibiotics for women in premature labour and those women who were culture-positive at 35–37 weeks, who had cleared the bacterium. The fact that the PCR assay currently takes from 1 to 2 hours would eliminate its use in women in active labour expected to deliver in <6 hours, but it would eliminate the use of antibiotics for some women. Further technological research is needed to improve the speed of this assay while maintaining its sensitivity. A more satisfying approach would be the development of a vaccine to promote the elimination of the GBS carrier state in the mother. This is currently under investigation.
The concept of a single policy for giving antibiotics before caesarean section should be questioned because the risk for postoperative maternal infection varies widely. The lowest-risk group consists of women undergoing caesarean section with membranes intact, who are not in labour. Although a large (9432 women) observational study showed that prophylactic antibiotics reduced the risk of postoperative infection with statistical significance, what is the clinical significance? The strategy of employing prophylactic antibiotics in this population means that 1000 women receive antibiotics to prevent six cases of endometritis and 4.4 cases of abdominal wound infection.10 About 100 mothers and their babies are exposed to antibiotics to prevent each maternal infection. A group with even lower risk could be identified by using the index-scoring method of the American College of Surgeons in which maternal risk factors can add two points to the score (Table 1). A pregnant woman with a score of 0 could safely forego prophylactic antibiotics. The highest-risk group of women undergoing caesarean section consists of those in labour with ruptured membranes with risk index scores of 1 or 2. To date, the best reported approach to prophylaxis in this high-risk population has been a combination of two antibiotics given after cord clamping, which eliminates fetal exposure to the antibiotics. Such alternative approaches are examples that would reduce use of predelivery antibiotics, with minimal adverse impact upon mothers and newborns.
|A patient with an American Society of Anesthesiologists preoperative assessment score of 3, 4, or 5.|
|An operation classified as contaminated or dirty infected. [Does not apply to most obstetric patients.]|
|Operations lasting over T hours: for caesarean delivery T = 1 (where T is the accepted length of time for an uncomplicated caesarean delivery).|
In summary, in our well-intended zeal to reduce infectious complications of delivery, we may have (paradoxically) impacted the microbial milieu for the next generation of newborns. The growing concern about long-term consequences of early life exposures to antibiotics requires the contemplation of new strategies for risk stratification of pregnant women and the deployment of new approaches. Long-term studies of current newborns need to be implemented so that the magnitude of the problem can be documented.
Disclosure of interests
WJ Ledger and MJ Blaser are on the advisory board for Procter and Gamble.
Contribution to authorship
WJ Ledger and MJ Blaser share authorship equally.
Details of ethics approval