Arch Gynecol Obstet (2009) 280:13–18 DOI 10.1007/s00404-008-0845-7

ORIGINAL ARTICLE

When should we perform prophylactic in elective cesarean cases?

Gokhan Yildirim · Kemal Gungorduk · Hamit Zafer Guven · Halil Aslan · Özgü Celikkol · Sinem Sudolmus · Yavuz Ceylan

Received: 9 August 2008 / Accepted: 3 November 2008 / Published online: 26 November 2008 © Springer-Verlag 2008

Abstract 1.77, 95% CI 0.51–6.16), and intensive care stay period Objective The aim of this study was to determine whether (P =0.16). the timing of prophylactic antibiotics at cesarean delivery Conclusions Time of prophylaxis application inXuences maternal and neonatal infectious morbidity. does not change maternal infectious morbidity in cesarean Study design This was a prospective, randomized trial. section deliveries. Preoperative prophylaxis application Four hundred patients that underwent elective cesarean sec- does not aVect neonate morbidity rates as stated in litera- tion between June and December 2007 formed the study ture. population. Eleven patients were excluded from the study because they needed transfusion during the cesarean sec- Keywords Antibiotic prophylaxis · Cesarean delivery · tion. The population was divided into two groups: Group A, Maternal/neonatal antibiotic prophylaxis was applied to 194 women before skin incision and Group B, antibiotic prophylaxis was applied to 195 women after umbilical cord clamping. The Introduction occurrence of endomyometritis/, infec- tion, febrile morbidity, total infectious morbidity, and neo- The most important in postpartum maternal natal complications were compared. infection is cesarean delivery [1]. Women undergoing Results There were 389 patients enrolled. No demo- cesarean section (C/S) are at 5- to 20-fold greater risk of graphic diVerences were observed between groups. No sig- infection than are women delivering vaginally. Infectious niWcant diVerence was found between the groups for total complications after cesarean delivery are an important infectious morbidity [relative risk (RR) 1.39, 95% conW- cause of maternal morbidity, and can prolong length of hos- dence interval (CI) 0.71–2.69] and endometritis (RR 1.40, pital stay. These include wound infection, postpartum 95% CI 0.43–4.51). There was no increase in neonatal sep- endometritis, and [2–4]. sis (RR 1.47, 95% CI 0.61–3.53), workup (RR 1.35, Antibiotic prophylaxis in women who undergo cesarean 95% CI 0.75–2.42), need for neonatal intensive care (RR delivery has been proven to be beneWcial in decreasing infectious morbidities both in high-risk women (e.g., labor- ing, after ), and low-risk patients (e.g., nonlaboring, intact membranes). The 60–70% reduc- G. Yildirim · K. Gungorduk · H. Z. Guven · H. Aslan · tion in endometritis and the 30–65% reduction in wound Ö. Celikkol · S. Sudolmus · Y. Ceylan infection rate prompted the Cochrane library to recommend Department of and Gynecology, Istanbul Bakirkoy Women and Children Hospital, prophylactic antibiotics to women who undergo both elec- Istanbul, Turkey tive and nonelective cesarean delivery [5–9]. First-generation cephalosporin antibiotics are the most & K. Gungorduk ( ) commonly used, agents and are usually administered Kartaltepe Mah. BÂtÂsÂkbaglarorta Sok., No: 13/7 BakÂrköy, Istanbul, Turkey following delivery of the infant after the cord is clamped e-mail: [email protected] [10]. Neither the use of broad-spectrum antimicrobials nor 123 14 Arch Gynecol Obstet (2009) 280:13–18 the administration of additional doses postoperatively has diVerence in postoperative , with  = 0.05. Four been shown to be superior to a single-dose cephalosporin hundred women who fulWlled the inclusion criteria were regimen [11]. randomly (two parts, block random using sealed, sequen- Optimal timing for prophylactic antibiotic administra- tially distributed envelopes to which the letters A and B had tion is based on animal studies that demonstrate a maxi- been allocated: the letter A to the antibiotic prophylaxis mum protective eVect when adequate tissue antibiotic before skin incision group and the letter B to the antibiotic levels are present prior to bacterial contamination [5]. How- prophylaxis after clamping umbilical cord group; the ever, many clinicians prefer to prevent exposure of the baby patients chose the envelops which were opened by the to antibiotics by starting them after the umbilical cord has investigator, and according to the letters, the group of been clamped. Because, if antibiotics are given before patients were determined) divided into two groups: 200 delivery of the infant and clamping of the umbilical cord, women represented the antibiotic prophylaxis before skin pre-existing neonatal infections might be masked, or there incision group (Group A), and 200 women the antibiotic might be an increased risk for the development of resistant prophylaxis after clamping umbilical cord group (Group organisms [12]. A recent randomized trails found no diVer- B). ence in neonatal septic workups or complications [13, 14]. One gram of cefazolin sodium (cephamezin 1 g There are few randomized controlled studies on timing EczacÂbaoÂ) was used for antibiotic prophylaxis during of antibiotic prophylaxis in cesarean section. Thigpen et al. cesarean section. One gram of cefazolin sodium was [13] concluded that there was no diVerence in infectious applied by mixing it with 50 cm3 normal saline for at least morbidity between preoperative antibiotics and those given 10 min, at most 45 min before skin incision or after clamp- at cord-clamp. On the other hand, Sullivan et al. [14] ing umbilical cord according to the determined prophylaxis reported decrease in both endomyometritis and total post method. No other antimicrobial agents were given unless a cesarean infectious morbidity in a prospective randomized postoperative infection was diagnosed. trial with use of preoperative cefazolin prophylaxis. Before cesarean section, a Foley was inserted, In this prospective randomized study, our aim was to and the abdomen cleaned with Providon-Iod solution. determine whether the timing of prophylactic antibiotics at Pfannenstiel incisions were done on all patients. Thereafter, cesarean delivery inXuences maternal and neonatal infec- the cesarean followed our standard procedure: transverse tious morbidity. lower uterine segment incision and delivery of the and . Suturing of the uterine incision was performed without exteriorization of the . The abdominal wall Methods was closed in two layers. Skin incisions were closed. The postpartum care for both groups was identical, and A randomized controlled study was conducted between included vital signs every 4 h, discontinuation of the Foley June 2007 and December 2007 in the Bakirkoy Maternity catheter and advancement of diet on the Wrst postoperative and Children Diseases Hospital in Istanbul, a tertiary care day. All patients in both groups were observed daily in center for obstetrics and gynecology and pediatrics. This order to assess the following variables: any sign of wound hospital has 16,000 deliveries annually. Women undergo- infection (, swelling, discharge or tenderness) ing elective cesarean section in the hospital without exclu- , uterine consistency and height, and peri- sion criteria were eligible for this study. Elective cesarean toneal reaction for . Clinical signs of urinary tract section was deWned as C/S performed before the presence infection were checked and urinalysis was performed. A of labor. The exclusion criteria were: use of antibiotics dur- complete count was assessed 24 h after delivery. The ing the last 24 h, pathology that should be treated with anti- patients were discharged on third postoperative day if there biotics, pre-existing maternal diseases (such as , was no infection or . Patients were followed collagen vascular disease, immune system problems), cho- for 6 weeks postpartum so that infectious complications in rioamnionitis, on admission, need of transfusion the puerperium period following abdominal delivery were before or during the cesarean section, ruptured membranes, all included. emergency cesarean section, and preterm cesarean section. The rates of postoperative infectious morbidity (endo- Women who satisWed the inclusion criteria were invited to metritis, wound infection, febrile morbidity, urinary tract participate, and they signed a consent form approved by the infection), estimated blood loss at surgery, and operative institutional review board. time were compared between the two groups. We also eval- In the year preceding this investigation, our institution uated neonatal outcomes including need for admission to had a 20% postcesarean endometritis rate. Using the the neonatal intensive care unit, Apgar score less than 7 at program “Med Calc,” we calculated that a sample size of 5 min, neonatal sepsis and sepsis workup. Neonatal sepsis 197 subjects would provide 80% power to detect a 50% was suspected if tachycardia and/or tachypnea, as well as 123 Arch Gynecol Obstet (2009) 280:13–18 15 an increased white count with bands, was present and was Table 1 Demographic characteristics of the study groups conWrmed by positive blood cultures. Length of stay, Group A Group B P value admission status, and decision to undertake a sepsis workup (n=194) (n = 195) were determined by the neonatologists. Febrile morbidity was deWned as a persistent fever of at Age § SD (year) 28.25 § 4.87 27.53 § 5.02 0.15 least 38°C for at least 24 h after surgery not associated with Gravidity § SD 2.57 § 1.05 2.45 § 1.05 0.25 lower abdominal or pelvic tenderness and no signs of infec- Parity § SD 1.14 § 0.67 1.11 § 0.68 0.59 tion elsewhere. Partial or total dehiscence or presence of GA at delivery § 38.32 § 0.94 38.24 § 0.69 0.31 purulent or serous wound discharge with induration, SD (weeks) warmth, and tenderness was considered as a wound infec- Body mass 31.98 § 2.89 31.96 § 2.29 0.93 W index § SD tion. Endometritis was de ned here as body temperature (kg/m2) greater than 38.5°C with concomitant foul-smelling dis- charge or abnormally tender uterus on bimanual examina- Data are presented as mean § SE tion. A diagnosis of urinary tract infection was only considered when urinary symptoms associated with signiW- cant (>1,000,000 organisms/ml) on culture of Results mid-stream specimen of urine were noted. Blood loss was estimated using a drop in hemoglobin concentration within There were a total of 5,596 deliveries during the study 24 h after the operation. period, of which 1,425 (25%) were cesarean sections. Four Med Calc 9.3 program was used for statistical analysis; hundred of these cesarean deliveries were included in this analysis was by intention to treat. Normal distribution of study; however, those who needed a blood transfusion continuous variables were assessed by Kolmogrov– (n = 11) were excluded, resulting in 389 women who com- Smimov test. Chi-square analysis was used for analysis of pleted the study and who were distributed into group A categoric variables, Student t test was used for normal dis- (n = 194) and group B (n = 195) (Fig. 1). In both groups no tributed variables in the analysis of continuous variables subjects were lost to follow-up. There were no signiWcant and Mann–Whitney U test was used for abnormally distrib- diVerences in demographics between women in either uted variables. Relative risk (RR) with 95% conWdence group participating in the study (Table 1). The indications interval (CI) were calculated. A P value of less than 0.05 for C/S are shown in Table 2. The indication for cesarean was considered statistically signiWcant. delivery was similar when two groups were compared.

Fig. 1 Structure of the study (Group A, the antibiotic prophy- Total Cesarean Births During laxis before skin incision. Group Study Period B, the antibiotic prophylaxis (n:1425) after umbilical cord clamping.)

Group A 400 Group B (n=200) randomized (n=200)

(n=6) (n=5) women women excluded excluded

Group A Group B (n=194) (n=195)

Complete trial Complete trial (n=194) (n=195)

123 16 Arch Gynecol Obstet (2009) 280:13–18

Table 2 Indications for cesarean section When the rate of maternal infectious morbidity is reana- lyzed on an ITT basis by considering the 11 excluded can- Group A Group B W (n = 194) (n =195) didates, there was no change in signi cance (RR 1.55, 95% CI 0.83–2.87). There were no serious side eVects related to Previous cesarean delivery 168 (86.6%) 173 (88.7%) the use of cefazolin. Multiple 5 (2.6%) 5 (2.6%) The neonatal outcome variables are illustrated in Fetal macrosomia (>4,500 g) 8 (%4.1) 6 (3.1%) Table 5. There were no signiWcant diVerences observed Breech and malpresentation 13 (6.7%) 10 (5.1%) between two groups in birth weight, 5-min Apgar score, Placenta previa 0 (0%) 1 (0.5%) neonatal sepsis, sepsis workup, NICU admission, or NICU admission days.

No statistically signiWcant diVerences between the two groups were observed with regard to preoperative hemoglo- Discussion bin level, preoperative hematocrit level, preoperative tem- perature, postoperative hemoglobin and hematocrit level, The beneWts of prophylactic antibiotics in the prevention of mean operative time, estimated blood loss, or total hospital serious infection and febrile morbidity after Cesarean sec- stays (Table 3). tion have been well documented in the literature [5, 9]. The frequency and causes of infectious morbidity in the Such beneWts appear to be applicable to both emergency two groups are shown in Table 4. In the antibiotic prophy- and elective Cesarean sections alike. There is still insuY- laxis before skin incision group (Group A), 8.8% had post- cient data to determine the optimal timing of antibiotic operative infectious morbidity as compared with 11.8% in administration (i.e., preoperative administration vs. admin- the antibiotic prophylaxis after clamping umbilical cord istration after the umbilical cord is clamped), although a group (Group B); however, this diVerence did not attain sta- study of non-obstetric surgical procedures demonstrated tistical signiWcance (RR 1.39, 95% CI 0.71–2.69). The inci- that prophylaxis was most eVective when administered dence of endomyometritis was 5/195 (2.6%) in the group within 2 h of the start of the procedure [15–17]. A, and 7/195 (3.6%) in the group B. These diVerences were However, in the case of cesarean delivery, preoperative not statistically diVerent (RR 1.40, 95% CI 0.43–4.51). antibiotic dosing is associated with a substantial plasma

Table 3 Surgical Group A (n = 194) Group B (n = 195) P value characteristics Preoperative hematocrit level § SD (%) 33.34 § 1.18 33.23 § 1.26 0.59 Preoperative hemoglobin level § SD (g/l) 11.48 § 0.69 11.43 § 0.79 0.46 Postoperative hematocrit level § SD (%) 30.17 § 0.97 30.04 § 0.92 0.18 Postperative hemoglobin level § SD (g/l) 9.91 § 0.50 9.78 § 0.59 0.02 Estimated blood loss § SD (ml) 656.29 § 190.54 668.92 § 203.57 0.52 Preoperative temperature § SD (°C) 36.89 § 0.22 36.91 § 0.18 0.19 Hospital stay § SD (days) 2.30 § 1.09 2.39 § 1.18 0.46 Data are presented as Operative time § SD (min) 36.63 § 2.66 37.12 § 3.89 0.14 mean § SE

Table 4 Infectious morbidity Group A (n = 194) Group B (n = 195) P value Relative risk (95% CI)

Total infectious morbidity 17 (8.8%) 23 (11.8%) 0.32 1.39 (0.71–2.69) Febrile morbidity 9 (4.6%) 7 (%3.6) 0.60 0.76 (0.29–2.09) Wound infection 6 (3.1%) 8 (%4.1) 0.59 1.34 (0.45–3.93) Endometritis 5 (2.6%) 7 (3.6%) 0.56 1.40 (0.43–4.51) Urinary tract infection 3 (1.5%) 5 (2.6%) 0.47 1.67 (0.39–7.11) 0 (0%) 0 (0%) Septic pelvic thrombophlebitis 0 (0%) 0 (0%) Respiratory tract infection 0 (0%) 0 (0%) Overall infectious morbidity 23 (11.9%) 27 (13.8%) 0.65 1.19 (0.65–2.16)

123 Arch Gynecol Obstet (2009) 280:13–18 17

Table 5 Neonatal Wndings Group A (n = 201) Group B (n = 198) P value Relative risk (95% CI)

Birth weight § SD (g)* 3263.75 § 505.86 3232.92 § 500.26 0.53 – 5-Minute Apgar score § SD* 9.08 § 0.71 9.06 § 0.78 0.79 – Neonatal Sepsis (%) 9 (4.4%) 13 (6.3%) 0.38 1.47 (0.61–3.53) NICU admission (%) 4 (2%) 7 (3.4%) 0.35 1.77 (0.51–6.16) Sepsis workup 23 (11.2%) 30 (14.6%) 0.30 1.35 (0.75–2.42) NICU length of stay (days)a 8.25 § 2.62 5.66 § 2.58 0.16 – a Data are presented as mean § SE level in the neonate. Because this therapeutic drug level in [20] reported that more intense research for sepsis was car- the newborn may alter blood culture results and, thus, per- ried in neonates of case groups with preoperative prophy- haps delay, or mask the diagnosis of neonatal sepsis, it is laxis by neonatologists. But in our study, no advantage or common practice to delay antibiotics until the baby is deliv- disadvantage of one to another in terms of infectious mor- ered and the umbilical cord clamped. This sequence pur- bidity of neonate was found. Neonatal sepsis and need for portedly achieves maternal levels suYcient to reduce intensive care were found to be similar in both groups. Sim- infectious morbidity without any transfer of the antibiotic ilar Wndings were demonstrated elsewhere [13, 14]. to the infant. The strength of this investigation was that it was a large Does the administration of antibiotic prophylaxis before prospective randomized trial of patients with similar mater- skin incision really decrease infectious morbidity? In a nal demographics, and a similar intrapartum proWle of study assessing prophylaxis application time with 303 events that have been associated with the subsequent devel- cases, endometritis and wound infection rates were found to opment of postcesarean infectious morbidity. In addition, be similar between both case groups. No diVerence was the completion of this investigation in a single institution found between two groups in terms of neonatal sepsis, and with the same team of surgeons probably increases the the need for neonatal intensive care unit [13]. Wax et al. validity of our results. However, our study has several limi- [18] showed infectious morbidity to be similar between two tations. We could not research the incidence of vaginal col- groups in their study of 80 cases with antibiotic prophylaxis onization by group B , so we could not cure application before and after cord clamping. Gordon et al. these patients. Despite these limitations, it seems that time [19] reported postcesarean endometritis rates to be similar of antibiotic prophylaxis application does not change compared to the group with preoperative antibiotic admin- maternal infectious morbidity in cesarean deliveries. Preop- istration in their studies assessing antibiotic prophylaxis erative prophylaxis application does not aVect neonate time in 114 pregnant women in 1979. Cunnnigham et al. morbidity rates as stated in literature. Well designed [20] did not Wnd any diVerence between postcesarean endo- randomized controlled studies including more cases are metritis and neonatal sepsis in the study of 600 cases where needed for examination of the eVect of prophylaxis applica- they applied prophylaxis preoperatively and after cord tion time on maternal and neonatal results. clamping. Like the others in our study, no diVerence was found between two groups in terms of endometritis, wound ConXict of interest statement None. infection and other maternal infectious morbidity rates But, in 2007, Sullivan et al., reported that administration of pro- phylactic cefazolin prior to skin incision resulted in a References decrease in both endomyometritis and total postcesarean 1. Gibbs RS, Hunt SE, Schwartz RW (1973) A follow-up study on infectious morbidity, compared with administration at the prophylactic antibiotics in cesarean section. Am J Obstet Gynecol time of cord clamping. We have no explanation for the U7:419–422 diVerence between the studies. But Sullivan et al. have only 2. Yokoe DS, Christiansen CL, Johnson R et al (2001) excluded need for emergent cesarean delivery and diabetic of and surveillance for postpartum infections. Emerg Infect Dis 7:837–841 patients; however, patients with rupture of membranes have 3. Ramsey PS, White AM, Guinn DA et al (2005) Subcutaneous tis- been included into study population. As a result, study sue reapproximation, alone or in combination with drain, in obese groups are not homogeneous. women undergoing cesarean delivery. Obstet Gynecol 105:967– Another question is that does administration of antibiotic 973 4. Henderson E, Love ES (1995) Incidence of hospital-acquired prophylaxis before skin incision or after cord clamping infections associated with cesarean section. J Hosp Infect 29:245– really eVects neonatal outcomes? Only Cunningham et al. 255. doi:10.1016/0195-6701(95)90271-6 123 18 Arch Gynecol Obstet (2009) 280:13–18

5. Smaill F, Hafmeyr GJ (2002) Antibiotic prophylaxis for cesarean in the uninfected laboring gravida: a randomized clinical trial. cesarean section. The Cochrane Database of Systematic reviews, Am J Obstet Gynecol 192:1864–1871. doi:10.1016/j.ajog.2004. Issue 3. Art No: CD 000933. doi:10.1002/14651858.CD000933 12.063 6. Gibbs RS, St Clair PJ, Castillo MS, Castaneda YS (1981) Bacteri- 14. Sullivan SA, Simith T, Chang E, Hulsey T, Vandorsten JP, Soper ologic eVects of antibiotic prophylaxis in high-risk cesarean D (2007) Administration of cefazolin prior to skin incision is supe- section. Obstet Gynecol 57:277–282 rior to cefazolin at cord clamping in preventing postcesarean infec- 7. American College of Obstetricians and Gynecologists (2003) tious morbidity: a randomised, controlled trial. Am J Obstet Prophylactic antibiotics in labor and delivery:ACOGPractice Gynecol 196:455e1–455e5 Bulletin No. 47. The College, Washington, DC 15. Classen DC, Evans RS, Pestotnik SI, Horn SD, Menlove RL, 8. Chelmov D, Ruehli MS, Huang E (2001) Prophylactic use of Burke JP (1992) The timing of prophylactic administration of anti- antibiotics for nonlaboring patients undergoing cesarean delivery biotics and the risk of surgical wound infection. N Engl J Med with intact membranes: a meta-analysis. Am J Obstet Gynecol 326:281–286 184:656–661. doi:10.1067/mob.2001.111303 16. Fiore MT, Pearlman MD, Chapman RL, Bhatt-Mehta V, Faix RG 9. Rudge MV, Atallah AN, Peraçoli JC, Tristao Ada R, Mendonça (2001) Maternal and transplacental pharmacokinetics of cefazolin. Neto M (2006) Randomized controlled trial on prevention of post- Obstet Gynecol 8:1075–1079. doi:10.1016/S0029-7844(01) cesarean infection using penicillin and cephalothin in Brazil. Acta 01629-5 Obstet Gynecol Scand 85:945–948 17. American Society of Health-System Pharmacists (1999) ASHP 10. Noyes N, Berkeley AS, Freedman K, Ledger W (1998) Incidence therapeutic guidelines on antimicrobial prophylaxis in surgery. of postpartum endomyometritis following single-dose antibiotics Am J Health Syst Pharm 56:1839–1888 prophylaxis with either /sulbactam, cefazolin or cefote- 18. Wax JR, Hersey K, Philput C, Wright MS, Nichols KV, Eggleston tan in high-risk cesarean section patients. Infect Dis Obstet Gyne- MK, Simith JF (1997) Single dose cefazolin prophylaxis for post- col 6:220–223 cesarean infections: before and after cord clamping. J Matern Fetal 11. Burke JF (1961) The eVective period of preventive antibiotic Med 6:61–65. doi:10.1002/(SICI)1520-6661(199701/02)6:1< action in experimental incisions and dermal lesions. Surgery 61::AID-MFM13>3.0.CO;2-P 50:161–168 19. Gordon HR, Phelps D, Blanchard K (1979) Prophylactic cesarean 12. Edwards RK, Clark P, Sistrom CL, DuV P (2002) Intrapartum anti- section antibiotics: maternal and neonatal morbidity before and biotic prophylaxis 1: relative eVects of recommended antibiotics after cord clamping. Obstet Gynecol 53:151 on gramnegative . Obstet Gynecol 100:534–539. 20. Cunningham FG, Leveno KJ, DePalma RT, Roark M, Rosenfeld doi:10.1016/S0029-7844(02)02096-3 CR (1983) Perioperative antimicrobials for cesarean delivery: 13. Thigpen B, Hood WA, Chauhan S, Bufkin L, BoWll S, Magann E, before or after cord clamping? Obstet Gynecol 62:151 Marrison SC (2005) Timing of prophylactic antibiotic administration

123