Surgical-site infections are the second most common hospital-acquired infections in U.S. hospitals, and reducing these infections remains a considerable challenge to all health care institutions.1 The most important risk factor for postpartum infection in women is cesarean delivery.2 Rates of infectious morbidity after cesarean delivery, including postpartum endometritis and wound infection, range from 3–20%.3–5 Because cesarean delivery is one of the most common surgical procedures performed in the United States today, surgical-site infections after cesarean delivery are among the most frequently encountered hospital-associated infections in obstetrics and gynecology.
Antimicrobial prophylaxis at the time of surgery has reduced the rate of postpartum endometritis and wound infection after both nonelective and elective cesarean delivery.6 Antimicrobial agents for surgical prophylaxis often are administered after umbilical-cord clamping. Withholding antibiotic administration until cord clamping avoids transplacental delivery of antibiotics to the fetus, reducing the concern of masking neonatal infections or affecting microbiologic cultures of the neonate. However, the timing of administration has been recognized to influence the efficacy of antimicrobial prophylaxis.7 Animal studies demonstrate that adequate tissue levels of the antibiotic before bacterial colonization are required to achieve maximal benefit from antimicrobial therapy.8 Classen et al9 demonstrate that prophylactic antibiotics are most effective when administered within the hour before incision, with surgical infection rates rising if antibiotics are administered after the surgical incision. Despite evidence that the reduction of surgical-site infections is greater when antimicrobial prophylaxis is administered before rather than after skin incision, common obstetrical practice is to delay antimicrobial prophylaxis until clamping of the neonate’s umbilical cord.6 The objectives of this study are to estimate 1) whether administration of prophylactic antibiotics before skin incision is more effective in reducing surgical-site infections (postpartum endometritis and wound infections) after cesarean delivery than administration of antibiotic prophylaxis after umbilical-cord clamping and 2) whether administration of antimicrobial prophylaxis before skin incision adversely affects the rate of neonatal infections or the evaluation of the neonate.
MATERIALS AND METHODS
This study was performed at Magee-Womens Hospital of the University of Pittsburgh Medical Center, a tertiary care hospital in Pittsburgh, Pennsylvania. In November 2004, a hospital-wide guideline recommending a change in the timing of surgical antimicrobial prophylaxis for cesarean deliveries was instituted. Before this institutional practice change, antibiotics routinely were administered after umbilical-cord clamping. The new recommendations advised administering the surgical antimicrobial prophylaxis before skin incision. A single dose of cefazolin was the preferred antimicrobial agent used for surgical prophylaxis in cesarean delivery before and after the change in guidelines.
We performed a review of maternal and neonatal infectious outcomes after cesarean delivery, evaluating the effect of our institution’s new guidelines on the timing of surgical antimicrobial prophylaxis. We identified all women undergoing both elective and nonelective cesarean delivery at our institution and created two comparative groups—one group of women who delivered before and the other who delivered after the new guidelines were implemented. Group 1 consisted of women who received antimicrobial prophylaxis after umbilical-cord clamping (July 2002 to November 2004). After allowing a 6-month period for implementation of the new guidelines, group 2 was established and contained women who delivered at our institution between June 2005 and August 2007, when antimicrobial prophylaxis was administered before skin incision. This study was approved by the University of Pittsburgh’s institutional review board.
The data for this study were obtained by using two research databases. A perinatal database gathers information on every obstetric delivery at our institution and contains approximately 300 variables including demographic information, obstetrical and medical history, and events at the time of delivery and the immediate postpartum period. The Newborn Medicine Division maintains a separate neonatal database, containing information on all newborns receiving care in the neonatal intensive care unit (NICU). All women identified in the maternal database then were matched with their newborns in the neonatal database if their newborns were evaluated in the NICU. All newborns with suspected or confirmed infections are treated in the NICU. We identified 1,979 newborns who received care in the NICU, 954 born to women in group 1 and 1,025 in group 2. Maternal and neonatal demographic information and maternal past medical and obstetrical history were captured. Antepartum complications and intrapartum events including medical comorbidities and group B streptococcus colonization were collected. Infectious morbidities, including chorioamnionitis, postpartum endometritis, and wound infections were included based on International Classification of Diseases, 9th Revision, Clinical Modification coding. The neonatal database provided information on neonatal outcomes, including evaluation for infection, antibiotics, microbiologic tests, and confirmed or suspected infection.
We assumed a 5% infection rate (endometritis or wound infection) in the post–cord-clamping group.10 To have 90% power to detect a 30% reduction in infections, 4,226 women would be needed in each group. We estimated that 27 months of data from each time period would be required to provide the sample size. Differences between the two groups of women and their neonates were evaluated using Fisher’s exact, χ2, and Student’s t-tests, where appropriate. Data were presented as mean±standard deviation for continuous variables and as frequency (percentage) for categorical variables. Multivariable logistic regression modeling then was used to determine which factors were independently associated with the development of postpartum endometritis and wound infection. Odds ratios (ORs) for independent risk factors were calculated along with the corresponding 95% confidence intervals (CIs). Models were developed using forward stepwise regression based on the likelihood ratio test statistic. χ2 for linear trend was used to evaluate whether the rates of postpartum endometritis decreased over time (divided by 3-month intervals) within each group of women. Generalized estimating equations were used to evaluate differences in neonatal outcomes between the two groups to account for correlation in multiple-gestation pregnancies. An exchangeable working correlation matrix was specified, and modified sandwich estimates of the variance were calculated. Statistical inference was based on the generalized Wald test statistic.11 All statistical analyses were considered significant at the level of .05.
The study population included 9,010 women undergoing cesarean delivery—4,229 in group 1 (prophylactic antibiotics administered after umbilical-cord clamping) and 4,781 in group 2 (prophylactic antibiotics administered before skin incision).The mean age and racial distribution were similar between the two study groups (Table 1). Approximately one half of women in each group were primiparous. The percentage of patients who received care from the resident teaching service was lower in group 2 compared with group 1 (14.9% compared with 18.9%, P<.001).
Maternal and obstetrical variables were compared between the two study groups (Table 2). Women receiving antibiotics before skin incision were less likely to smoke or to have a prior cesarean delivery. Mean body mass index was similar between the two groups. Similar proportions of women in each group were laboring before cesarean delivery. Group B streptococcus colonization was more common in group 2 than in group 1 (24.4% compared with 22.2%, P=.05). Chorioamnionitis was less common in women receiving prophylactic antibiotics before skin incision than in women who received antibiotics after umbilical-cord clamping (5.6% compared with 10.3%, P<.001).
The rates of postpartum maternal infections were compared between the study groups (Table 3 and Fig. 1). Among women receiving prophylactic antibiotics after umbilical-cord clamping, 164 (3.9%) developed postpartum endometritis. The rate of postpartum endometritis in women receiving prophylactic antibiotics before skin incision was 2.2%, representing a 44% decrease in the rate of endometritis compared with women receiving antibiotic prophylaxis after umbilical-cord clamping (unadjusted OR 0.56, 95% CI 0.44–0.72). After adjusting for variables associated with endometritis in the study population (maternal race, age, parity, chorioamnionitis, trial of labor, gestational age, maternal hypertension, and resident teaching service [data not shown]), administration of antimicrobial prophylaxis before skin incision remained associated with a reduction in endometritis (OR 0.61, 95% CI 0.47–0.79). Similarly, there was a reduction in the rate of wound infections among women receiving prophylactic antibiotics before skin incision compared with those receiving antibiotics after umbilical-cord clamping (2.5% compared with 3.6%, P<.01) After controlling for variables associated with wound infection (maternal race, maternal age, multiple gestation, and resident teaching service [data not shown]), there remained a statistically significant reduction in wound infections among women receiving antibiotics before skin incision (OR 0.70, 95% CI 0.55–0.90).
Intrapartum antibiotic therapy, as is common for Group B streptococcus colonization and chorioamnionitis, may affect postpartum infection rates. The observed protective effect of surgical antimicrobial prophylaxis before skin incision did not change when women with chorioamnionitis or Group B streptococcus colonization were removed from the analysis. Excluding women with chorioamnionitis, endometritis rates in women receiving antimicrobial prophylaxis before skin incision were reduced by nearly 50% (OR 0.51, 95% CI 0.38–0.69) and wound infection rates were reduced by 30% (OR 0.71, 95% CI 0.55–0.92). When women with chorioamnionitis or Group B streptococcus colonization were excluded, women receiving antimicrobial prophylaxis before skin incision had lower rates of endometritis (OR 0.54, 95% CI 0.38–0.75) and wound infection (OR 0.72, 95% CI 0.55–0.96) than did women receiving antimicrobial prophylaxis after umbilical-cord clamping.
The effect of timing of antimicrobial prophylaxis on neonatal outcomes was compared between the two study groups (Table 4). During this study, there were 1,979 newborns who received care in the NICU (954 from group 1 and 1,025 from group 2). The mean birth weight of neonates born to women in group 2 was similar to the weight of those born to women in group 1 (2,214±994 g compared with 2,134±1,038 g, P=.13). Mean gestational age was greater in group 2 (34 weeks ±4 days compared with 33 weeks ±5 days, P=.01). Length of stay in the NICU was not different among neonates in each study group. Neonatal infections were stratified into early-onset (within 3 days of birth) and late-onset (7 days of life or later). Culture-proven infections within the first 3 days of life were similar in both groups, occurring in 12 neonates born to women in group 1 (1.3%) and seven (0.7%) in group 2. Culture-proven late-onset neonatal infections were less common among neonates born to women receiving antibiotic prophylaxis before skin incision than in those born to women receiving antibiotic prophylaxis after umbilical-cord clamping (1.8% compared with 5.7%, P<.001). The proportion of neonates treated for presumed infection (clinical presentation consistent with infection but microbiologic cultures negative) was similar between the two study groups (24.1% in group 1, 22.2% in group 2). Further, there was no difference in the proportion of neonates without infection who initially received a short course of empiric antibiotics after birth while undergoing an infectious work-up (commonly referred to as “rule-out sepsis”) among women receiving antibiotics after umbilical-cord clamping compared with women receiving antibiotics before skin incision (531 [55.7%] compared with 587 [57.3%], P=.69).
We examined the organisms recovered from neonates with early-onset neonatal infections because these infections may have been acquired peripartum, when surgical antibiotic prophylaxis may have its greatest effect on isolation of microorganisms. Twelve organisms were recovered from neonates in group 1 and seven organisms from those in group 2. There were fewer cases of group B streptococcus infections (one compared with five) but more infections due to Escherichia coli (four compared with two) in neonates born to women receiving antibiotic prophylaxis before skin incision compared with women receiving antibiotics after umbilical-cord clamping. Organisms infrequently isolated from group 1 included: Haemophilus influenzae (n=1), coagulase-negative Staphylococcus (n=3), and viridans Streptococcus (n=1); from group 2: Enterobacter species (n=1) and Listeria monocytogenes (n=1). No statistically significant change in recovery of organisms was observed between the two study groups.
We verified that the new institutional guidelines were adopted by our institution’s physicians by performing a manual audit of the medical records of the first 115 cesarean deliveries in July 2002 (group 1) and the first 115 cesarean deliveries in January 2006 (group 2). All audited charts of women in group 1 indicated that the antibiotics were administered after umbilical-cord clamping. Ninety-two percent (106/115) of the audited charts of women in group 2 indicated antibiotic administration within 60 minutes before skin incision, confirming that the recommended change in timing of antimicrobial prophylaxis was implemented at our institution.
We have demonstrated that administration of antimicrobial prophylaxis for cesarean delivery before skin incision, compared with after umbilical-cord clamping, is associated with a nearly 40% reduction in postpartum endometritis (3.9% to 2.2%) and a 30% reduction in wound infection (3.6% to 2.5%). An estimated 1.7 million hospital-acquired infections occur each year in the United States.1 Surgical-site infections account for 20% of all hospital-acquired infections; these hospital-acquired infections continue to represent an important cause of morbidity and mortality.12 Antimicrobial prophylaxis, long used to prevent infection in several surgical procedures, is being incorporated widely as a performance measure.13 Cesarean delivery is a major risk factor for postpartum infection.14 In women undergoing elective and nonelective cesarean delivery, antimicrobial prophylaxis is an important strategy to reduce surgical-site infections.6 As the proportion of women delivering by cesarean exceeds 30% of all deliveries in the United States, strategies that reduce surgical-site infection risks will have a substantial effect on the morbidity associated with cesarean delivery.15
The objective of antibiotic prophylaxis is to administer an antimicrobial agent that both provides coverage against commonly encountered organisms and achieves adequate tissue and serum drug levels throughout the duration of the operation.16 Burke demonstrated the importance of establishing tissue levels of antibiotics before bacterial contamination.8 Data from human studies reinforce the influence of antibiotic timing on infection rates. Lowest rates of wound infections were observed when antibiotics were administered within 2 hours before skin incision, with infection rates increasing when antibiotics were delayed until after skin incision.9 The National Surgical Infection Prevention Project encourages the administration of antimicrobial agents within 60 minutes before incision.16 Although the historical practice of antimicrobial prophylaxis in obstetrics after umbilical-cord clamping contradicts this basic tenet of surgical prophylaxis, our study is consistent with antimicrobial prophylaxis principles supporting preoperative administration of antimicrobials for cesarean deliveries.
Recent antimicrobial trials in women undergoing cesarean delivery arrived at conflicting conclusions. A randomized controlled trial comparing antimicrobial prophylaxis among 303 laboring women demonstrated a nonstatistically significant reduction in endometritis (relative risk [RR] 0.67) and wound infection (RR 0.84) in women administered antimicrobials before skin incision compared with after cord clamping.17 Although the authors concluded that there was no difference in maternal infectious morbidity, it is noteworthy that enrollment did not meet the authors’ predetermined recruitment goals to achieve power. A subsequent randomized controlled trial by Sullivan et al showed a statistically significant reduction in endometritis (RR 0.22) and a nonsignificant reduction in wound infection (RR 0.4).10 Lower surgical-site infection rates were seen among 516 women after a policy change to administer antibiotics before incision.18 Our results add considerably to the emerging data that administration of antimicrobial prophylaxis before skin incision, compared with after umbilical-cord clamping, is associated with a reduction in rates of postpartum endometritis and wound infection.
Concern about adversely affecting neonatal infectious morbidity or the evaluation of the neonate was the rationale for the widespread practice of withholding antibiotic prophylaxis until umbilical-cord clamping. Previous studies were underpowered to evaluate the effect of the timing of antimicrobial prophylaxis on neonatal infectious outcomes.10,17 Our study of nearly 2,000 newborns receiving care in the NICU found that the timing of maternal antibiotics had no effect on the occurrence of early-onset newborn infection. The rates of culture-proven early-onset neonatal infections were similar (1%) between study groups. Importantly, the timing of prophylaxis had no effect on the number of neonates evaluated for infection. Furthermore, there was no difference in the rate of neonates with early-onset presumed infection (received antimicrobial therapy for 7 days for presumed infection) between the post–cord-clamping group (n=230, 24.1%) and the preincision group (n=227, 22.2%). Although the rate of culture-proven late-onset infection was lower in neonates born to women receiving antibiotics before skin incision, the reason for this interesting observation is unclear. Overall, our study demonstrates that antimicrobial prophylaxis administered before skin incision has no adverse effect on neonatal infections or on the evaluation of the neonate.
We recognize that a limitation of this study is that it retrospectively compares two time periods, during which time other variables may have influenced the maternal and neonatal outcomes. To our knowledge, there were no other changes to the health care delivered to women undergoing cesarean delivery at our institution or to neonates during the two time periods, increasing the likelihood that timing of antimicrobial administration was related to the observed decrease in infectious complications. Biases of the study design include the possibility of confounding variables as well as the accuracy of a database and the use of International Classification of Diseases, 9th Revision, Clinical Modification classifications for the definitions of our outcomes. A further limitation is the study’s generalizability because we report a single institution’s experience. Two recent randomized controlled trials enrolled 357 and 303 participants, inadequate to address neonatal outcomes.10,17 Our study examined nearly 2,000 newborns who were evaluated and treated in the NICU. Despite our large cohort, the study is underpowered for the neonatal outcomes (eg, difference of death at or before 3 days of life 2.4–1.8%); a study of more than 10,000 neonates admitted to the NICU is needed to achieve adequate power. A large randomized controlled trial powered for neonatal outcomes would address many of these limitations.
In summary, the administration of antimicrobial prophylaxis before skin incision at cesarean delivery is associated with a reduction in surgical-site infections. Importantly, the administration of antimicrobials before skin incision had no negative effect on neonatal infection rates, the evaluation of the neonate, or treatment for presumed neonatal infection. This study should provide clinicians with reassurance of the safety and efficacy of preoperative antimicrobial prophylaxis in women undergoing cesarean delivery.
1.Klevens RM, Edwards JR, Richards CL Jr, Horan TC, Gaynes RP, Pollock DA, et al. Estimating health care-associated infections and deaths in U.S. hospitals, 2002. Public Health Rep 2007;122:160–6.
2.Gibbs RS. Clinical risk factors for puerperal infection. Obstet Gynecol 1980;55:178S–84.
3.Yokoe DS, Christiansen CL, Johnson R, Sands KE, Livingston J, Shtatland ES, et al. Epidemiology of and surveillance for postpartum infections. Emerg Infect Dis 2001;7:837–41.
4.Henderson E, Love EJ. Incidence of hospital-acquired infections associated with caesarean section. J Hosp Infect 1995;29:245–55.
5.Miner AL, Sands KE, Yokoe DS, Freedman J, Thompson K, Livingston JM, et al. Enhanced identification of postoperative infections among outpatients. Emerg Infect Dis 2004;10:1931–7.
6.Smaill F, Hofmeyr GJ. Antibiotic prophylaxis for cesarean section. The Cochrane Database of Systemic Reviews 2002, Issue 3. Art. No.: CD000933. DOI: 10.1002/14651858.CD000933.
7.Page CP, Bohnen JM, Fletcher JR, McManus AT, Solomkin JS, Wittmann DH. Antimicrobial prophylaxis for surgical wounds. Guidelines for clinical care [published erratum appears in Arch Surg 1993;128:410]. Arch Surg 1993;128:79–88.
8.Burke JF. The effective period of preventive antibiotic action in experimental incisions and dermal lesions. Surgery 1961;50:161–8.
9.Classen DC, Evan RS, Pestotnik SL, Horn SD, Menlove RL, Burke JP. The timing of prophylactic administration of antibiotics and the risk of surgical-wound infection. N Engl J Med 1992;326:281–6.
10.Sullivan SA, Smith T, Chang E, Husley T, Vandorsten JP, Soper D. Administration of cefazolin prior to skin incision is superior to cefazolin at cord clamping in preventing postcesarean infectious morbidity: a randomized, controlled trial [published erratum appears in Am J Obstet Gynecol 2007;197:333]. Am J Obstet Gynecol 2007;196:455.e1–5.
11.Hardin JW, Hilbe JM. Generalized estimating equations. Boca Raton (FL): Chapman & Hall/CRC; 2003.
12.Anderson DJ, Kaye KS, Classen D, Arias KM, Podgorny K, Burstin H, et al. Strategies to prevent surgical site infections in acute care hospitals. Infect Control Hosp Epidemiol 2008;29:S51–61.
13.Fry DE. Surgical site infections and the surgical care improvement project (SCIP): evolution of national quality measures. Surg Infect (Larchmt) 2008;9:579–84.
14.Chaim W, Bashiri A, Bar-David J, Shoham-Vardi I, Mazor M. Prevalence and clinical significance of postpartum endometritis and wound infection. Infect Dis Obstet Gynecol 2000;8:77–82.
15.Hamilton BE, Martin JA, Ventura SJ. Births: preliminary data for 2006. National vital statistics reports: vol 56 no 7. Hyattsville (MD): National Center for Health Statistics; 2007.
16.Bratzler DW, Houck PM, Surgical Infection Prevention Guidelines Writes Workgroup, American Academy of Orthopaedic Surgeons, American Association of Critical Care Nurses, American Association of Nurse Anesthetists, et al. Antimicrobial prophylaxis for surgery: an advisory statement from the National Surgical Infection Prevention Project. Clin Infect Dis 2004;38:1706–15.
17.Thigpen BD, Hood WA, Chauhan S, Bufkin L, Bofill J, Magann E, et al. Timing of prophylactic antibiotic administration in the uninfected laboring gravida: a randomized clinical trial. Am J Obstet Gynecol 2005;192:1864–8.
© 2009 The American College of Obstetricians and Gynecologists
18.Kaimal AJ, Zlatnik MG, Cheng YW, Thiet M-P, Connatty E, Creedy P, et al. Effect of a change in policy regarding the timing of prophylactic antibiotics on the rate of postcesarean delivery surgical-site infections. Am J Obstet Gynecol 2008;199:310.e1–5.