ajog.org
Editorials
Antibiotic prophylaxis in obstetrics Scott A. Sullivan, MD; David Soper, MD
T
he use of antimicrobial prophylaxis for the prevention of maternal, fetal, and neonatal infections is well-established in obstetric practice. Prevention of chorioamnionitis, neonatal sepsis, urinary tract infections, endometritis, and other common infections reduces morbidity and mortality rates and health care costs. There are a number of examples of level I evidence that have been used to develop clinical recommendations that have improved outcomes. These can also provide obstetric units with important quality and safety measures. The prevention of early-onset group B Streptococcus (GBS) sepsis in neonates is one such example. The incidence of early-onset GBS sepsis has decreased nearly 80%, from 1.8 per 1000 births to 0.26 per 1000 births by 2008, through the wide-spread introduction of universal screening and intrapartum antimicrobial prophylaxis.1 These treatment guidelines have decreased overall neonatal mortality and morbidity rates. These are regularly updated by the Centers for Disease Control and Prevention in the face of new evidence or emerging bacterial resistance. Although there is a promising GBS vaccine under development, universal screening will remain the standard for the foreseeable future. Antimicrobial prophylaxis for cesarean delivery is an important preoperative safety and quality measure. Cesarean delivery is associated with significant risks of endometritis, wound infection, and other serious maternal infections. These risks are reduced with proper prophylaxis, both for elective and nonelective indications.2 This risk reduction ranges from 60-70% in pooled estimates from prospective studies. Recent prospective trials and metaanalyses demonstrate that properly dosed and preoperatively timed antibiotics can improve maternal infectious outcomes significantly.3,4 Preterm premature rupture of the membranes is associated with significant risks for maternal and fetal infectious morbidity, which includes chorioamnionitis and neonatal sepsis. Among patients who have ruptured membranes at 36 weeks’ gestation. They are to be commended for following the Preferred Reporting Item for Systematic Reviews and Metaanalyses statement and for their careful analysis. They report the intriguing conclusion that, in their analysis, the use of antimicrobial prophylaxis significantly reduced chorioamnionitis (2.9% vs 6.1%; relative risk [RR], 0.49; 95% confidence interval [CI], 0.27e0.91) and endometritis (0% vs 2.2%; relative risk, 0.12; 95% CI 0.02e0.62) in the subset of women whose membranes were ruptured for >12 hours. If these results are replicated in a sizeable prospective trial, this would be an important finding that would change current obstetric practice. The current study has several strengths that should be noted. The 5 trials that were included had a low risk of allocation bias by Cochrane Collaboration tool assessment. Intent-to-treat analysis was used, and both random and mixed effects models were used when appropriate. In addition, publication bias was not apparent by funnel plot analysis. Heterogeneity between studies was variable but generally was not significant. These are key elements that are needed to evaluate the reliability of a metaanalysis. Despite these strengths, we must exercise caution with these results, as the authors correctly point out. Neither chorioamnionitis nor endometritis were the primary outcomes of the metaanalysis. They were planned elements of the secondary outcomes. The authors included 20 secondary maternal and neonatal outcomes in total, so this observation could be from chance alone. All 5 trials included for analysis used a different antimicrobial prophylaxis regimen. None of the included trials were restricted to women who were GBS positive; the older trials were before universal GBS screening. There may have been obstetric management differences across 5 different countries and 26 years of study. Given these limitations, we do not agree with the authors that it is currently reasonable to start antimicrobial prophylaxis for patients who undergo prolonged inductions or for those likely to have PROM for >12 hours. We believe current American College of Obstetricians and Gynecologists guidelines for PROM and GBS should be followed until additional evidence is available.7 The authors of a recently published metareview on this subject report different findings.8 They included 4 of the 5 trials that were analyzed in the study of Saccone and Berghella,6 with a very comparable final number of studied MAY 2015 American Journal of Obstetrics & Gynecology
559
ajog.org
Editorials subjects, 2639 vs 2699. They excluded the fifth trial over concerns about randomization specifics. These authors reported no difference in chorioamnionitis and/or endometritis in patients with either an early induction of labor (12 hours) group who received antibiotics (RR, 0.34; 95% CI, 0.08e1.47). There are differences in statistical analysis between the 2 reviews, including the combination of chorioamnionitis and endometritis as 1 endpoint. However, the differences in outcome do illustrate some of the problems and limitations inherent to metaanalysis, including the subjective nature of choosing studies to include. We also cannot forget that we lack a properly powered analysis for neonatal outcomes. In Saccone and Berghella,6 there were only 12 perinatal deaths and 32 cases of neonatal sepsis in 2699 study subjects. These poor perinatal outcomes thankfully are rare at near term and beyond. However, it does not allow us to conclude that increased use of antimicrobial prophylaxis, specifically for PROM, is completely safe for the fetus and neonate. GBS may be a cautionary tale; although there has been undoubtedly neonatal benefit to universal screening, we have experienced changes in bacterial resistance. This had led to an increased use of vancomycin, a major tool against methicillin resistant Staphylococcus aureus and one that has been associated recently with potential renal risk in children. It is wise to consider all outcomes, intended or otherwise, when changing clinical recommendations. The 2 recent metaanalyses are in agreement on a number of findings. Perhaps most importantly, they agree on the primary one. In neither trial was the use of routine antimicrobial prophylaxis for all women with PROM at >37 weeks’ gestation beneficial for either maternal or neonatal outcomes. In Saccone and Berghella,6 there was no decrease in maternal endometritis, chorioamnionitis, total infection, and neonatal sepsis. This supports the current usual obstetric practice of not routinely administering antibiotics for PROM. Interestingly, both studies also report a significant increase in cesarean delivery among the patients who received antibiotics. In the current study, the authors posit that the increase seen in cesarean deliveries (RR, 1.32; 95% CI, 1.09e1.6) could be due to chance, given the multiple comparisons and no clear mechanism to explain the increase. The authors have done us a service in presenting diverse data from around the world in an elegant analysis. They also bring needed attention to the fact that additional data are needed for this and other obstetric antimicrobial prophylaxis questions. We commend them for their efforts and join with them in calling for additional research into PROM prophylaxis, ideally in a multicentered randomized clinical trial, with chorioamnionitis and endometritis as a primary outcome. Secondary endpoints ideally would include neonatal outcomes and a trial that is powered adequately to detect differences in neonatal sepsis, acidosis, and death. This likely would require significant resources and the involvement of a research consortium such as the National Institute of Child 560 American Journal of Obstetrics & Gynecology MAY 2015
Health and Human Development Maternal-Fetal Medicine Network. Given the need to balance the risks of bacterial resistance, maternal side-effects, and cost vs the potential decrease in morbidity, we believe they would be resources well spent. We would not stop with PROM in calling for new research into antimicrobial prophylaxis in obstetrics. There are many unanswered or incompletely answered questions that need new data. Examples include dosing of antimicrobial prophylaxis for obese patients who undergo cesarean delivery. What dose and for which size patients is still largely unknown. Another example is extended antimicrobial prophylaxis for cesarean delivery. Although early data are promising,9 the ideal dose and antimicrobial agent are not known. We need randomized controlled trials on the use of prophylactic antibiotics for third-/fourth-degree obstetric laceration repairs, exam-indicated cerclages, and manual removal of placentas. We need guidance for proper cesarean antimicrobial prophylaxis for patients with hypersensitivities to beta-lactams. We also need significant new data and insights on combating emerging bacterial resistance in obstetrics, which threatens to make our current treatment guidelines obsolete or possibly even futile. The authors have provided an important insight into another indication for antimicrobial prophylaxis in obstetrics. We look forward to future prospective data on the risks for infection in PROM and the potential role for prophylaxis. Infectious disease research in obstetrics should remain a priority for the research community. REFERENCES 1. Schrag S, Verani J. Intrapartum antibiotic prophylaxis for the prevention of perinatal group B streptococcal disease: experience in the United States and implications for a potential GBS vaccine. Vaccine 2013;31(suppl 4):D20-6. 2. Smaill F, Grivell R. Antibiotic prophylaxis versus no prophylaxis for preventing infection after cesarean section. Cochrane Database Syst Rev 2014;10:CD007482. 3. Sullivan S, Smith T, Chang E, VanDorsten J, Soper D. Administration of cefazolin prior to skin incision is superior to cefazolin at cord clamping in preventing post-cesarean infectious morbidity: a randomized controlled trial. Am J Obstet Gynecol 2007;196:455.e1-5. 4. American College of Obstetricians and Gynecologists Committee on Obstetric Practice. Antimicrobial prophylaxis for cesarean delivery: timing of administration. Committee Opinion no. 465. Obstet Gynecol 2010;116:791-2. 5. Kenyon S, Boulvain M, Neilson JP. Antibiotics for preterm rupture of membranes. Cochrane Database Syst Rev 2014;12:CD001058. 6. Saccone G, Berghella V. Antibiotic prophylaxis for term or near-term premature rupture of membranes: metaanalysis of randomized trials. Am J Obstet Gynecol 2015;212:627.e1-9. 7. American College of Obstetricians and Gynecologists. Premature rupture of membranes. ACOG practice bulletin, no. 139. Obstet Gynecol 2013;122:918-30. 8. Wojciezek A, Stock O, Flenady V. Antibiotics for prelabour rupture of membranes at or near term. Cochrane Database Syst Rev 2014;10: CD001807. 9. Tita A, Rouse D, Blackwell S, Saade G. Emerging concepts in antibiotic prophylaxis for cesarean delivery: a systematic review. Obstet Gynecol 2009;113:675-82.
Editorials
Antibiotic prophylaxis in obstetrics Scott A. Sullivan, MD; David Soper, MD
T
he use of antimicrobial prophylaxis for the prevention of maternal, fetal, and neonatal infections is well-established in obstetric practice. Prevention of chorioamnionitis, neonatal sepsis, urinary tract infections, endometritis, and other common infections reduces morbidity and mortality rates and health care costs. There are a number of examples of level I evidence that have been used to develop clinical recommendations that have improved outcomes. These can also provide obstetric units with important quality and safety measures. The prevention of early-onset group B Streptococcus (GBS) sepsis in neonates is one such example. The incidence of early-onset GBS sepsis has decreased nearly 80%, from 1.8 per 1000 births to 0.26 per 1000 births by 2008, through the wide-spread introduction of universal screening and intrapartum antimicrobial prophylaxis.1 These treatment guidelines have decreased overall neonatal mortality and morbidity rates. These are regularly updated by the Centers for Disease Control and Prevention in the face of new evidence or emerging bacterial resistance. Although there is a promising GBS vaccine under development, universal screening will remain the standard for the foreseeable future. Antimicrobial prophylaxis for cesarean delivery is an important preoperative safety and quality measure. Cesarean delivery is associated with significant risks of endometritis, wound infection, and other serious maternal infections. These risks are reduced with proper prophylaxis, both for elective and nonelective indications.2 This risk reduction ranges from 60-70% in pooled estimates from prospective studies. Recent prospective trials and metaanalyses demonstrate that properly dosed and preoperatively timed antibiotics can improve maternal infectious outcomes significantly.3,4 Preterm premature rupture of the membranes is associated with significant risks for maternal and fetal infectious morbidity, which includes chorioamnionitis and neonatal sepsis. Among patients who have ruptured membranes at 36 weeks’ gestation. They are to be commended for following the Preferred Reporting Item for Systematic Reviews and Metaanalyses statement and for their careful analysis. They report the intriguing conclusion that, in their analysis, the use of antimicrobial prophylaxis significantly reduced chorioamnionitis (2.9% vs 6.1%; relative risk [RR], 0.49; 95% confidence interval [CI], 0.27e0.91) and endometritis (0% vs 2.2%; relative risk, 0.12; 95% CI 0.02e0.62) in the subset of women whose membranes were ruptured for >12 hours. If these results are replicated in a sizeable prospective trial, this would be an important finding that would change current obstetric practice. The current study has several strengths that should be noted. The 5 trials that were included had a low risk of allocation bias by Cochrane Collaboration tool assessment. Intent-to-treat analysis was used, and both random and mixed effects models were used when appropriate. In addition, publication bias was not apparent by funnel plot analysis. Heterogeneity between studies was variable but generally was not significant. These are key elements that are needed to evaluate the reliability of a metaanalysis. Despite these strengths, we must exercise caution with these results, as the authors correctly point out. Neither chorioamnionitis nor endometritis were the primary outcomes of the metaanalysis. They were planned elements of the secondary outcomes. The authors included 20 secondary maternal and neonatal outcomes in total, so this observation could be from chance alone. All 5 trials included for analysis used a different antimicrobial prophylaxis regimen. None of the included trials were restricted to women who were GBS positive; the older trials were before universal GBS screening. There may have been obstetric management differences across 5 different countries and 26 years of study. Given these limitations, we do not agree with the authors that it is currently reasonable to start antimicrobial prophylaxis for patients who undergo prolonged inductions or for those likely to have PROM for >12 hours. We believe current American College of Obstetricians and Gynecologists guidelines for PROM and GBS should be followed until additional evidence is available.7 The authors of a recently published metareview on this subject report different findings.8 They included 4 of the 5 trials that were analyzed in the study of Saccone and Berghella,6 with a very comparable final number of studied MAY 2015 American Journal of Obstetrics & Gynecology
559
ajog.org
Editorials subjects, 2639 vs 2699. They excluded the fifth trial over concerns about randomization specifics. These authors reported no difference in chorioamnionitis and/or endometritis in patients with either an early induction of labor (12 hours) group who received antibiotics (RR, 0.34; 95% CI, 0.08e1.47). There are differences in statistical analysis between the 2 reviews, including the combination of chorioamnionitis and endometritis as 1 endpoint. However, the differences in outcome do illustrate some of the problems and limitations inherent to metaanalysis, including the subjective nature of choosing studies to include. We also cannot forget that we lack a properly powered analysis for neonatal outcomes. In Saccone and Berghella,6 there were only 12 perinatal deaths and 32 cases of neonatal sepsis in 2699 study subjects. These poor perinatal outcomes thankfully are rare at near term and beyond. However, it does not allow us to conclude that increased use of antimicrobial prophylaxis, specifically for PROM, is completely safe for the fetus and neonate. GBS may be a cautionary tale; although there has been undoubtedly neonatal benefit to universal screening, we have experienced changes in bacterial resistance. This had led to an increased use of vancomycin, a major tool against methicillin resistant Staphylococcus aureus and one that has been associated recently with potential renal risk in children. It is wise to consider all outcomes, intended or otherwise, when changing clinical recommendations. The 2 recent metaanalyses are in agreement on a number of findings. Perhaps most importantly, they agree on the primary one. In neither trial was the use of routine antimicrobial prophylaxis for all women with PROM at >37 weeks’ gestation beneficial for either maternal or neonatal outcomes. In Saccone and Berghella,6 there was no decrease in maternal endometritis, chorioamnionitis, total infection, and neonatal sepsis. This supports the current usual obstetric practice of not routinely administering antibiotics for PROM. Interestingly, both studies also report a significant increase in cesarean delivery among the patients who received antibiotics. In the current study, the authors posit that the increase seen in cesarean deliveries (RR, 1.32; 95% CI, 1.09e1.6) could be due to chance, given the multiple comparisons and no clear mechanism to explain the increase. The authors have done us a service in presenting diverse data from around the world in an elegant analysis. They also bring needed attention to the fact that additional data are needed for this and other obstetric antimicrobial prophylaxis questions. We commend them for their efforts and join with them in calling for additional research into PROM prophylaxis, ideally in a multicentered randomized clinical trial, with chorioamnionitis and endometritis as a primary outcome. Secondary endpoints ideally would include neonatal outcomes and a trial that is powered adequately to detect differences in neonatal sepsis, acidosis, and death. This likely would require significant resources and the involvement of a research consortium such as the National Institute of Child 560 American Journal of Obstetrics & Gynecology MAY 2015
Health and Human Development Maternal-Fetal Medicine Network. Given the need to balance the risks of bacterial resistance, maternal side-effects, and cost vs the potential decrease in morbidity, we believe they would be resources well spent. We would not stop with PROM in calling for new research into antimicrobial prophylaxis in obstetrics. There are many unanswered or incompletely answered questions that need new data. Examples include dosing of antimicrobial prophylaxis for obese patients who undergo cesarean delivery. What dose and for which size patients is still largely unknown. Another example is extended antimicrobial prophylaxis for cesarean delivery. Although early data are promising,9 the ideal dose and antimicrobial agent are not known. We need randomized controlled trials on the use of prophylactic antibiotics for third-/fourth-degree obstetric laceration repairs, exam-indicated cerclages, and manual removal of placentas. We need guidance for proper cesarean antimicrobial prophylaxis for patients with hypersensitivities to beta-lactams. We also need significant new data and insights on combating emerging bacterial resistance in obstetrics, which threatens to make our current treatment guidelines obsolete or possibly even futile. The authors have provided an important insight into another indication for antimicrobial prophylaxis in obstetrics. We look forward to future prospective data on the risks for infection in PROM and the potential role for prophylaxis. Infectious disease research in obstetrics should remain a priority for the research community. REFERENCES 1. Schrag S, Verani J. Intrapartum antibiotic prophylaxis for the prevention of perinatal group B streptococcal disease: experience in the United States and implications for a potential GBS vaccine. Vaccine 2013;31(suppl 4):D20-6. 2. Smaill F, Grivell R. Antibiotic prophylaxis versus no prophylaxis for preventing infection after cesarean section. Cochrane Database Syst Rev 2014;10:CD007482. 3. Sullivan S, Smith T, Chang E, VanDorsten J, Soper D. Administration of cefazolin prior to skin incision is superior to cefazolin at cord clamping in preventing post-cesarean infectious morbidity: a randomized controlled trial. Am J Obstet Gynecol 2007;196:455.e1-5. 4. American College of Obstetricians and Gynecologists Committee on Obstetric Practice. Antimicrobial prophylaxis for cesarean delivery: timing of administration. Committee Opinion no. 465. Obstet Gynecol 2010;116:791-2. 5. Kenyon S, Boulvain M, Neilson JP. Antibiotics for preterm rupture of membranes. Cochrane Database Syst Rev 2014;12:CD001058. 6. Saccone G, Berghella V. Antibiotic prophylaxis for term or near-term premature rupture of membranes: metaanalysis of randomized trials. Am J Obstet Gynecol 2015;212:627.e1-9. 7. American College of Obstetricians and Gynecologists. Premature rupture of membranes. ACOG practice bulletin, no. 139. Obstet Gynecol 2013;122:918-30. 8. Wojciezek A, Stock O, Flenady V. Antibiotics for prelabour rupture of membranes at or near term. Cochrane Database Syst Rev 2014;10: CD001807. 9. Tita A, Rouse D, Blackwell S, Saade G. Emerging concepts in antibiotic prophylaxis for cesarean delivery: a systematic review. Obstet Gynecol 2009;113:675-82.
