Introduction
The objectives of this review question are to evaluate the effectiveness of antibiotic prophylaxis administered routinely to all babies after birth, or to well babies in whom maternal or fetal risk factors for early-onset neonatal infection have been identified, to prevent early-onset neonatal infection. Specific issues prioritised by the guideline development group (GDG) for consideration in this question were: potential differences in clinical management, depending on whether intrapartum antibiotics had been administered as prophylaxis to prevent early-onset neonatal infection or for maternal indications; which class of antibiotics to use; timing, route and frequency of antibiotic administration; dosage; and the impact of prematurity on clinical management.
The considerations regarding inclusion of evidence obtained using particular study designs are similar to those in the review question relating to intrapartum antibiotics (see Chapter 6). For the evaluation of clinical outcomes (such as prevention of early-onset neonatal infection) the GDG restricted consideration to evidence from randomised controlled trials (RCTs). For pharmacokinetic outcomes (for example incidence of therapeutic or toxic concentrations of a particular antibiotic) used to evaluate dosage regimens the GDG restricted consideration to evidence from RCTs where such evidence was available. Other comparative or non-comparative pharmacokinetic and pharmacodynamic studies were considered only when no relevant evidence from RCTs was identified. The GDG drew initial conclusions about effectiveness based on clinical outcomes reported in RCTs and then reviewed pharmacokinetic outcomes only for those antibiotics that it was considering recommending. As noted in Chapter 6, the rationale for considering pharmacokinetic outcomes in this guideline is that few antibiotics are licensed for use in pregnancy or in preterm babies; the GDG prioritised consideration of safe and effective dosage regimens in all of the review questions relating to antibiotic treatment.
Review question
In babies with maternal risk factors for early-onset neonatal infection is routine administration of antibiotics to the baby effective in preventing early-onset neonatal infection?
Existing NICE guidance
Intrapartum care (NICE clinical guideline 55, 2007) identified prelabour rupture of membranes (PROM) as a major obstetric risk factor for neonatal infection and evaluated various clinical management strategies following term PROM (although preterm PROM was outside the scope of the guideline). The guideline recommendations covered the clinical management and care of the baby after birth, including criteria for administration of antibiotics to the baby after birth. The guideline noted that babies who are asymptomatic at birth have a lower risk of developing neonatal sepsis. The guideline recommended a risk-based clinical management strategy for women with term PROM, which included the following elements:
If there are no signs of infection in the woman, do not give antibiotics to the woman or the baby, even if the membranes have been ruptured for over 24 hours.
Do not perform blood, cerebrospinal fluid or surface culture tests in an asymptomatic baby.
Observe asymptomatic term babies born to women with term PROM more than 24 hours before labour closely for the first 12 hours of life (at 1 hour, 2 hours and then 2 hourly for 10 hours). The observations should include: general wellbeing; chest movements and nasal flare; skin colour including perfusion, by testing capillary refill; feeding; muscle tone; temperature; heart rate; and respiration.
Offer immediate referral to a neonatal care specialist for a baby with any symptom of possible sepsis, or born to a woman who has evidence of chorioamnionitis.
Description of included studies
Six studies (all RCTs) were identified for inclusion for this review question (Auriti 2005; Hammerberg 1989; Hammerschlag 1980; Patel 1999; Pyati 1983; Siegel 1982).
Clinical outcomes reported in randomised controlled trials
Four RCTs evaluated the effectiveness of antibiotics in babies with risk factors for early-onset neonatal infection (Pyati 1983; Auriti 2005; Hammerberg 1989; Hammerschlag 1980).
The first study (Pyati 1983) evaluated the effectiveness of benzylpenicillin administered intramuscularly to babies with low birthweight (indicating a high risk of group B streptococcus [GBS] infection) within 60–90 minutes of birth and then every 12 hours for 72 hours compared to no administration of benzylpenicillin within 60–90 minutes of birth, but administration at 12 hours after birth and then every 12 hours for 72 hours.
The second study (Auriti 2005) evaluated the effectiveness of a single bolus of ampicillin plus netilmicin administered intravenously to babies with at least one risk factor for early-onset neonatal infection upon admission to the neonatal intensive care unit (NICU) compared to a 3-day course of ampicillin plus netilmicin administered intravenously in two divided doses. The risk factors for infection were:
history of prolonged rupture of membranes (more than 24 hours)
suspected chorioamnionitis (rupture of membranes more than24 hours, stained and foul amniotic fluid, maternal fever or leukocytosis)
proven maternal urinary tract infection
leukopenia (less than 5000 cells/mm3) or neutropenia (less than 1750 cells/mm3) at birth
presence on admission to the NICU of a central venous or arterial catheter, an endotracheal tube, pleural drainage or history of invasive resuscitation manoeuvres.
The third study (Hammerberg 1989) evaluated the effectiveness of piperacillin plus placebo versus ampicillin plus amikacin administered to babies with risk factors for sepsis within 7 days of birth. The risk factors for sepsis were:
prolonged rupture of membranes (more than 18 hours; more than 36% of babies in both treatment groups were born following prolonged rupture of membranes)
intrapartum maternal fever (more than 38°C)
foul-smelling amniotic fluid.
The study also included babies with the following clinical signs or laboratory findings consistent with sepsis (these were also described as risk factors for sepsis by the study authors):
apnoea (cessation of breathing for more than 15 seconds resulting in bradycardia and cyanosis)
poor perfusion (capillary refill time more than 5 seconds)
ratio of immature to total neutrophils (I:T ratio) more than 0.2.
The fourth study (Hammerschlag 1980) evaluated the effectiveness of erythromycin eye ointment versus 1% silver nitrate eye drops administered immediately after birth for the prevention of neonatal chlamydial conjunctivitis and respiratory tract infection in babies born to women who tested positive for Chlamydia (Chlamydia trachomatis) in the third trimester of pregnancy.
Two further RCTs evaluated the effectiveness of antibiotics given to all babies shortly after birth (Patel 1999; Siegel 1982). The first study (Patel 1999) evaluated the effectiveness of benzylpenicillin administered intramuscularly to babies within 1 hour of birth versus no benzylpenicillin for the prevention of GBS infection. The second study (Siegel 1982) evaluated the effectiveness of benzylpenicillin administered intramuscularly to babies within 1 hour of birth versus topical tetracycline ointment for the prevention of gonococcal eye infections (caused by N gonorrhoeae), GBS colonisation and systemic GBS disease.
Pharmacokinetic and pharmacodynamic studies
Based on the GDG’s initial consideration of clinical outcomes reported in RCTs, the pharmacokinetics and pharmacodynamics of benzylpenicillin were prioritised for evaluation, but no RCTs or studies of other designs reporting pharmacokinetic outcomes associated with benzylpenicillin treatment were identified for inclusion.
Evidence statements
In babies at risk of early-onset GBS infection(birthweight 501–2000 g) who received intramuscular benzylpenicillin within 60–90 minutes of birth and then every 12 hours for 3 days, there was no difference in the rate of culture-proven early-onset GBS infection or mortality compared with babies who received no early treatment (moderate quality evidence).
In preterm babies (less than 32 weeks’ gestation) with risk factors for (or laboratory evidence of) early-onset neonatal infection there was no difference in mortality among those who received a single bolus of ampicillin plus netilmicin compared with those who received a 3-day course of the same antibiotics (low quality evidence).
In babies aged less than 7 days with risk factors for (or clinical signs or laboratory evidence of) sepsis who received piperacillin plus placebo there were no differences in blood culture-proven sepsis, mortality from infection, mortality during antibiotic treatment or up to 1 week after treatment, renal impairment or hepatic impairment compared with babies who received ampicillin plus amikacin (low quality evidence).
In babies born to Chlamydia-positive mothers, application of erythromycin eye ointment immediately after birth was more effective than 1% silver nitrate eye drops in reducing the incidence of culture-proven chlamydial eye infection. There was, however, no difference in the incidence of culture-proven chlamydial nasopharyngeal infection or pneumonia among the two groups (low quality evidence).
Babies who received intramuscular benzylpenicillin prophylaxis within 1 hour of birth for the prevention of GBS infection had lower rates of blood culture-proven early-onset sepsis, but not mortality, compared with babies who received no benzylpenicillin prophylaxis at birth (low quality evidence).
Babies who received intramuscular benzylpenicillin prophylaxis within 1 hour of birth for the prevention of gonococcal eye infections had lower rates of culture-proven early-onset benzylpenicillin-susceptible bacterial infections and culture-proven early-onset GBS infection, but not culture-proven early-onset benzylpenicillin-resistant bacterial infections, nor mortality associated with any early-onset infection, compared with babies who received topical tetracycline eye ointment (low quality evidence).
Health economics profile
The GDG planned to conduct a cost effectiveness analysis comparing different strategies for identifying and treating babies at risk of early-onset neonatal infection or with symptoms and signs of early-onset neonatal infection. However, no published health economic analyses were identified in relation to this review question, and no clinical evidence was identified to inform development of a health economic model specifically for the guideline.
Evidence to recommendations
Relative value placed on the outcomes considered
The GDG members prioritised the following clinical outcomes in the baby as they considered these would be reported consistently across study populations and would aid their decision making by most strongly reflecting early-onset neonatal infection:
failure of prevention of neonatal infection
mortality (the GDG prioritised this outcome where mortality due to infection was specified)
duration of hospital stay
neonatal adverse events
long-term outcomes in the baby
resistance among neonatal flora.
Consideration of clinical benefits and harms
The priority outcome for the GDG was the prevention of early-onset neonatal infection and the associated benefit of reduced morbidity and so the GDG made its decisions primarily based upon this outcome. The GDG considered the additional benefits or harms associated with other clinical outcomes where the findings for early-onset neonatal infection were equivocal, where there was evidence of reduced incidence of early-onset neonatal infection following treatment with more than one type of antibiotic and where there was clear evidence of statistically significant harms. The GDG also considered the spectrum of antibiotic activity and the potential broader harm of antibiotic resistance.
Consideration of net health benefits and resource use
The costs associated with intravenous administration of antibiotics to the baby are the cost of the antibiotics themselves, the equipment and staff costs needed to set up and perform the intravenous infusion, and the need for a hospital stay. The GDG noted that, as a general principle, prevention of infection is cost effective. The guideline review of maternal risk factors used to specify indications for maternal intrapartum antibiotic prophylaxis or postnatal antibiotic prophylaxis in the baby reflected the GDG’s opinion that the benefits of universal treatment experienced by the few would be outweighed by the potential harms of exposing many women and babies to antibiotics unnecessarily. Such practice would also promote the potential broader harm of promoting antibiotic resistance.
Quality of evidence
The GDG noted that only a few RCTs were identified for inclusion in the guideline review and that only one RCT provided evidence of even moderate quality. The evidence for all other outcomes was of very low or low quality. No evidence was identified at all relating to duration of hospital stay, long-term outcomes in the baby or resistance among neonatal flora.
No pharmacokinetic studies specific to benzylpenicillin in babies with or without risk factors were identified for inclusion.
Neonatal risk factors
Evidence came from only one RCT in which low birthweight was used as an indicator of GBS in a population of predominately black babies. The diagnostic test accuracy of this proxy measure of the risk of GBS infection is unclear, and there were no significant differences in GBS infection or mortality due to early-onset neonatal infection between those babies who received early treatment with benzylpenicillin and those who did not.
Maternal risk factors
Evidence came from only one RCT in which babies with the maternal risk factor of a positive Chlamydia test result were given localised prophylaxis to prevent eye infection immediately after birth; the antibiotic used was either erythromycin (an antibiotic) in the form of eye ointment or 1% silver nitrate (a disinfectant) in the form of eye drops. Significantly fewer babies who received erythromycin eye ointment developed conjunctivitis compared with those who received silver nitrate eye drops. However, there were no significant differences in the incidence of nasopharyngeal or respiratory infections (pneumonia) due to Chlamydia.
Combinations of maternal and neonatal risk factors
Evidence came from two RCTs that examined the effects of different antibiotics or different regimens involving the same combination of antibiotics. The GDG highlighted the absence of evidence with regard to risk factors in placebo-controlled (or no-treatment controlled) studies.
In the first RCT there were no statistically significant differences within 7 days of birth for any sepsis, mortality or toxicity outcome when administration of piperacillin plus placebo was compared with ampicillin plus amikacin in babies with any of six maternal or neonatal risk factors for sepsis.
In the second RCT there was no statistically significant difference in mortality caused by early-onset infection in preterm babies (less than 32 weeks of gestation) who had any of five maternal or neonatal risk factors for infection following prophylaxis with a single intravenous bolus dose of ampicillin plus netilmicin compared to a 3-day course of intravenous ampicillin plus netilmicin.
Universal administration of antibiotics
Two quasi-randomised clinical trials in newborn babies evaluated universal prophylaxis with intramuscular benzylpenicillin within 1 hour of birth compared with no treatment or treatment with tetracycline eye ointment, respectively.
In the first study (comparison with no treatment), statistically significantly fewer babies who received benzylpenicillin had clinical sepsis, identified sepsis, GBS or a positive GBS blood culture, or died due to clinical or identifiable sepsis. In subgroup analyses performed according to gestational age, statistically significantly fewer term babies who received benzylpenicillin had clinical or identified sepsis, but there was no statistically significant difference for preterm babies. However, statistically significantly fewer preterm babies who received benzylpenicillin had a positive GBS blood culture, whereas there was no statistically significant difference for term babies. There were no statistically significant differences in mortality from clinical or identifiable sepsis in either preterm or term babies, although the finding for all babies showed a significant protective effect of benzylpenicillin.
In the second study (comparison with tetracycline eye ointment), significantly fewer babies who received benzylpenicillin developed bacterial infections, benzylpenicillin-susceptible bacterial infections, GBS infection and meningitis.
Other considerations
The GDG did not identify any equalities issues requiring attention in this review question, although the group drew a careful distinction between preterm and term babies.
Recommendations from existing guidelines (including Intrapartum care, NICE clinical guideline 55 [2007], which covers immediate care of babies born to women with term PROM, including indications for antibiotic treatment) were also reviewed, and the GDG discussed the need for recommendations to prevent early-onset neonatal infection in the babies covered by this guideline.
Key conclusions
The GDG defined ‘routine’ administration of antibiotics as antibiotic administration to every newborn baby. If intrapartum antibiotic prophylaxis was indicated but not received (for example because the baby was born before antibiotics could be administered to the woman) then the GDG considered that the baby would still be regarded as having a risk factor for early-onset neonatal infection.
The GDG noted that risk factors should not be counted twice during assessment of the baby after birth, especially with regard to whether or not intrapartum antibiotic prophylaxis was indicated, and whether or not intrapartum antibiotic prophylaxis was received.
Neonatal risk factors
The GDG chose not to expand its recommendation (in Chapter 6) regarding intrapartum antibiotic prophylaxis for GBS infection to include postnatal administration of antibiotics.
Maternal risk factors
The GDG considered that systemic rather than localised antibiotic prophylaxis would be more effective in preventing chlamydial infections. The GDG noted that current practice is to treat women who test positive for Chlamydia with a single dose of azithromycin, and the group was satisfied that this would provide prophylactic cover for their babies too.
Although no evidence relating to gonococcal infections was identified for inclusion in the guideline review, the GDG noted that current practice is to treat women who test positive for Gonococcus (N gonorrhoeae) with a single dose of an appropriate antibiotic, and the group was satisfied that this would provide prophylactic cover for their babies too.
Combinations of maternal or neonatal risk factors
The GDG did not regard either of the RCTs identified for inclusion regarding antibiotic prophylaxis for combinations of maternal and neonatal risk factors as being very relevant to the UK setting. The group noted that piperacillin (which contains ampicillin plus a pharmaceutical agent that is active against Pseudomonas species) is currently marketed in the UK only in combination with tazobactum, and so it would be an unlikely choice of antibiotic. Furthermore, all the antibiotics evaluated in the RCT involving piperacillin were broad-spectrum, and the GDG did not wish to recommend the use of broad-spectrum antibiotics unless absolutely necessary because of the risk of promoting antibiotic resistance.
Universal administration of antibiotics
The GDG noted that in the 1980s, when one of the RCTs relating to universal administration of antibiotics was conducted, intrapartum antibiotic prophylaxis may not have been established clinical practice. The GDG chose not to expand its recommendation (in Chapter 6) regarding intrapartum antibiotic prophylaxis to cover postnatal administration of antibiotics for the prevention of early-onset neonatal infection.
Having considered all the evidence identified for inclusion, the GDG concluded that it was important that babies with no maternal or neonatal risk factors should not receive antibiotics routinely. Such practice will avoid exposing babies unnecessarily to potential adverse effects of antibiotics, and it will minimise the risk of antibiotic resistance developing further. Thus, the GDG’s recommendation was that antibiotics should not be given routinely to babies without risk factors, clinical indicators or laboratory evidence of possible infection.
Recommendations
View in own window
| Number | Recommendation |
|---|
| Avoiding routine use of antibiotics in the baby |
|---|
| 31 | Do not routinely give antibiotic treatment to babies without risk factors for infection or clinical indicators or laboratory evidence of possible infection. |
Research recommendations
No research recommendations were identified for this review question.
