HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by REID, V. C. Articles by FRY, E. P. Search for Related Content PubMed PubMed Citation Articles by REID, V. C. Articles by FRY, E. P.

Vaginal Preparation With Povidone Iodine and Postcesarean Infectious Morbidity: A Randomized Controlled Trial

From the Department of Obstetrics and Gynecology, School of Medicine; Department of Epidemiology, School of Public Health; and the North Carolina Program for Women’s Health Research, University of North Carolina, Chapel Hill, North Carolina.

Address reprint requests to: Katherine E. Hartmann, MD, PhD University of North Carolina, Chapel Hill North Carolina Program for Women’s Health Research Cecil G. Sheps Center for Health Services Research 725 Airport Road, CB #7590 Chapel Hill, NC 27599-7590 E-mail: khartman{at}med.unc.edu

	Abstract

Top Abstract Materials and Methods Results Discussion References

 
Objective: To determine whether vaginal preparation with povidone iodine before cesarean decreased the incidence of postpartum infectious morbidity.

Methods: Participants were randomly assigned to vaginal preparation with povidone iodine (n = 247) or no preparation (n = 251). Postpartum infectious morbidity included fever, defined as temperature of 38C or greater after the day of surgery; endometritis, defined as fever with abdominal or uterine tenderness and initiation of intravenous antibiotics; and wound separation, defined as disruption of the abdominal incision that required wound care. We calculated overall rates of postpartum infectious morbidity, relative risks (RR), and 95% confidence intervals (CI) for the effect of vaginal preparation. As designed and reported, the trial had at least 80% power to detect a 10% or greater absolute difference in rates of overall infectious morbidity, fever, and endometritis (two-tailed, = 0.05).

Results: There was no difference between groups in maternal age, parity, race, education, prior cesarean, type of anesthesia, labor before current cesarean, number of vaginal examinations during labor, internal monitoring, prophylactic antibiotic use, gestational age at delivery, or payment status. Excluding 68 women with chorioamnionitis, incidence of postoperative fever was 19.3%, endometritis 7.2%, and wound separation 7.0%. Vaginal preparation with povidone iodine before cesarean had no effect on risk for fever (RR 1.1, 95% CI 0.8, 1.6), endometritis (RR 1.6, 95% CI 0.8, 3.1), or wound separation (RR 0.6, 95% CI 0.3, 1.3).

Conclusion: Vaginal preparation with povidone iodine before cesarean had no effect on the incidence of fever, endometritis, or wound infection.

Infectious morbidity is the most frequent complication of cesarean delivery. Of women who have cesareans, 5–24% have clinically significant fevers; and 6–21% are diagnosed with uterine infections (endomyometritis or endometritis), 1–5% with more extensive pelvic infections including abscesses and 2–9% with breakdown of the surgical incision, most often caused by wound infection.^1–5 Death from infectious complications is rare, with an incidence of six deaths per 100,000 cesareans.⁶ However, the personal and economic costs are substantial. Compared with women who do not have postcesarean infection, women with infection have more discomfort and pain, prolonged intravenous access, longer hospital stays, potentially greater anxiety about their postoperative progress, and disappointment with their birth experience. For each additional day after cesarean, the average hospital charge is more than $1900, excluding diagnostic testing and antibiotic costs.⁷ Conservatively estimated, we spend more than half a billion dollars each year treating postcesarean infectious morbidity in the United States.^7,8

Postpartum endometrial, pelvic, and wound infections are most often polymicrobial. Vaginal organisms generally dominate; rectal and skin contaminants are also common.^9–11 During and after cesarean, the uterus is open to the vagina, markedly so if the cervix is dilated. In many cases, the surgeon’s hand, reaching below the infant’s head or presenting part, is in direct contact with the vagina. Vaginal bacterial flora have been cultured from the delivering surgeon’s glove in 79% (95% confidence interval [CI] 58%, 100%) of cesareans that follow labor.¹² In these cases, vaginal flora are delivered directly to the uterus, abdominal cavity, and the abdominal incision.

Cleansing of all body surfaces that could be in contact during a surgical procedure is a central tenet of aseptic technique. Although such sterile preparation does not actually result in a sterile field, it is intended to minimize the presence of bacterial and fungal organisms normally present on skin and mucous membranes. We hypothesized that by reducing the inoculum of organisms delivered to the uterus by preoperative vaginal preparation subsequent infectious morbidity would decrease. We conducted a randomized clinical trial to test this hypothesis. Specifically, our goal was to determine whether vaginal preparation with 10% povidone iodine before cesarean would reduce the rates of fever or endometritis.

	Materials and Methods

Top Abstract Materials and Methods Results Discussion References

 
Between May 1996 and September 1998, we enrolled 501 women who had cesarean births at the University of North Carolina Women’s Hospital, a tertiary care center with over 2800 births per year. Patients were admitted from 13 health department or federally funded rural health clinics, our resident physician and faculty clinics, and from a large health maintenance organization. The study was open to all women admitted to labor and delivery who were mentally competent to consent for their cesareans. We did not exclude subjects on the basis of language. Exclusion from enrollment was related solely to medical contraindications to the vaginal preparation required for the trial, including highly emergent cesarean; allergy to povidone iodine, iodine, or shellfish; bleeding placenta previa; and active genital herpes.

After informed consent was obtained and while in the operating room, women were assigned to the intervention or control group. To make assignments, we used a computer-generated, permuted block randomization schedule, with block sizes of six and eight. Individual assignments were sealed in sequentially numbered opaque envelopes, flap side down, taped firmly to the abdominal preparation packs kept in a warmer immediately outside the operating rooms.

Materials used for vaginal preparation were consistent with those used for vaginal preparation in other settings at the University of North Carolina Hospitals. For participants assigned to the intervention, a sterile container of 4 x 4-in gauze was opened and saturated with a 10% povidone iodine surgical scrub solution. Resident physicians were responsible for preoperative preparation, which was done after subjects received adequate anesthesia. Residents were instructed to place a Foley catheter first, if required, then to fold and grasp one or two of the saturated gauze sponges at a time with a sterile sponge stick, and to prepare the field from the vaginal apex to the introitus with attention to the anterior, posterior, and lateral aspects of the vaginal walls. Three passes were required. Controls did not receive sham preparation. After preparation, all participants’ care was routine without other study-related intervention or evaluation.

After participants were discharged, one of four physicians, masked to randomization assignment, reviewed the admission record and abstracted information in the following categories: exposures related to risk of infection, such as number of vaginal examinations and time of rupture of membranes; characteristics necessary to define outcome, such as oral temperature and clinical examination; and treatment for diagnosed or suspected infections. Demographic characteristics and medical history were obtained from the perinatal database of the Department of Obstetrics and Gynecology. Two members of the same team of physicians reviewed participants’ entire charts a minimum of 3 months after discharge to search systematically for evidence of complications after discharge, predominantly wound-healing problems. When a complete abstract packet was assembled, a maternal-fetal medicine specialist and an obstetrician-gynecologist independently assigned diagnostic outcome codes. Discrepant diagnoses, affecting less than 3% of participants, prompted joint review by those two individuals and a third obstetrician-gynecologist. All discrepancies were resolved as oversights of available information. Of 501 women enrolled, three lacked chart materials required to assign outcome. Their randomization assignment was not known to abstractors who sought the missing chart materials. All three were in the intervention group and were excluded from the data analyses.

Operational definitions for outcomes were established a priori. The principal objective measure was fever, defined as any temperature of 38C or greater, after the day of surgery. Febrile morbidity was defined as postoperative fever on 2 or more calendar days, excluding the day of surgery. Endometritis was defined as postoperative fever, with a physician’s note indicating uterine or abdominal pain or tenderness, preceding an order for broad-spectrum, intravenous antibiotics. We required a statement in the chart that the antibiotic treatment was for uterine or pelvic infection, and that the chart and laboratory studies indicated no other apparent source of serious infection. Uncomplicated cystitis was not considered a serious infection. Wound separation was defined by a chart note during admission or other subsequent related care that reported separation of the operative incision requiring intervention, such as dressing changes or operative debridement.

We classified other infectious morbidity as follows: clinical chorioamnionitis was defined as a temperature elevation during labor followed by an order for broad-spectrum, intravenous antibiotics for the treatment of presumed chorioamnionitis. Similarly, the definition of clinical endometritis differed from the definition of endometritis in that a temperature of 38C or greater was not required. We required only that a physician’s note recorded a temperature elevation after the first postoperative day that prompted his or her concern because of uterine or abdominal pain or tenderness, for which broad-spectrum, intravenous antibiotics were ordered. Pulmonary diagnoses included atelectasis, defined by a physician’s note of lung examination findings and fever interpreted as atelectasis, and pneumonia, defined as a physician’s note of lung examination findings consistent with pneumonia accompanied by radiologic confirmation. Urinary tract infections were defined as uncomplicated if the patient had a positive urine culture obtained by catheterization, without fever or costovertebral angle tenderness, for which oral antibiotics were prescribed, and defined as complicated if she had a positive urine culture obtained by catheterization, with note of fever or costovertebral angle tenderness, for which intravenous antibiotics were prescribed. Diagnoses expected to be more rare, such as abscess or pelvic thrombophlebitis, required clinical diagnosis in addition to radiologic or laboratory evidence, as appropriate.

As an indication of severity of infectious complications, we also evaluated the number of postoperative days with a fever, the average duration of postoperative antibiotic administration, and length of postoperative hospitalization. We are confident that, under ordinary circumstances, women with postdischarge complications returned for care at the University of North Carolina. We have no reason to suspect that detection of complications varied by intervention status.

Sample size of 500 subjects was based on two factors: we wanted to be able to detect a 10% absolute difference in risk of each outcome; eg, fever rates of 22.5% compared with 12.5%, or endometritis rates of 15% compared with 5%; and we wanted to enroll a sufficient number of participants to accommodate an estimated 10–15% of participants who would be excluded after meeting diagnostic criteria for chorioamnionitis after systematic chart review. As designed and reported, the study had a greater than 80% power to detect such a difference if the outcome investigated was no more common than 23% (two-tailed ² with = 0.05). Analysis was done using Stata Statistical Software (Stata Corporation, College Station, TX). Continuous data were inspected for normal distribution and compared using a t test. The association between vaginal preparation and dichotomous outcomes was investigated by calculating relative risk (RR) and 95% confidence interval (CI) based on the Pearson ² test. The Mantel-Haenszel test of homogeneity was used to evaluate potential heterogeneity in stratified analyses of ratio measures.

	Results

Top Abstract Materials and Methods Results Discussion References

 
The average age of participants in the trial analysis was 27 years; 48% were white; 64% had Medicaid coverage or no insurance, and 61% had had a previous birth. Table 1 summarizes key characteristics of participants by intervention assignment, excluding women who met criteria for chorioamnionitis. One hundred four participants had clinical diagnoses of infectious morbidity for which they received intravenous antibiotics during their obstetric admissions. Sixty-eight women had clinical chorioamnionitis. When those 68 women with antepartum infection were included, estimates of the effect of vaginal preparation were not meaningfully different from those presented. However, their inclusion distorted the absolute rates of postpartum fever and infectious complications by inflating the denominator. As a result, we planned to exclude them from further reports of outcomes, leaving 217 participants assigned to receive vaginal preparation and 213 assigned to the control group (Figure 1).

View larger version (17K): [in this window] [in a new window]   Figure 1. Trial profile, indicating the participants assigned to receive vaginal preparation and those assigned to the no preparation control group.

Seven percent of women (n = 30) had wound separations; four cases occurred among those who also had endometritis. The majority (19 of 30) had complete superficial wound disruption down to the level of the fascia. There were no cases of fascial dehiscence. Excluding the four women who had endometritis, six of 26 women with wound separations (23%) had a fever during their hospital stay. The rate of wound complications was lower in the vaginal preparation group (5.5% compared with 8.4%); however, the difference is not statistically significant. In contrast with the possible increased risk for endometritis among those who did not labor, if there is a protective effect of vaginal preparation on risk of wound breakdown, it is among those who did not labor (P < .27 for test of homogeneity).

Similar proportions of the intervention and control groups (16.6% and 16.9%, respectively) received postoperative antibiotics (P = .93). Length of postoperative stay did not differ by intervention group. Women in the preparation group had an average stay of 3.38 days compared with 3.31 in the no preparation group (P = .68); median postoperative stay was 2.9 days in both groups. Each group had three participants with stays of 10 days or longer. Thus, no measure of infectious morbidity or severity was meaningfully associated with vaginal preparation.

	Discussion

Top Abstract Materials and Methods Results Discussion References

 
This randomized trial of vaginal preparation with povidone iodine before cesarean delivery clearly illustrates the need to pursue objective evidence of benefit from seemingly logical interventions. Similar vaginal preparation is routine for other surgical procedures in which communication between the lower genital tract and the uterus or abdominopelvic cavity could result. The preparation solution used has been shown to decrease markedly the total number of aerobes and anaerobes recovered from the vaginal pool compared with normal saline.¹³ Ten minutes after vaginal preparation in nonpregnant women, the total organism count was reduced below detectable levels, with a gradual return to roughly half the original concentration of organisms after 2 hours.¹³ Not only did we find no benefit, but the risk of infection among women who had vaginal preparation without labor suggests there is potential for harm.

While this trial was under way, Rouse et al¹⁴ published a study of chlorhexidine vaginal irrigation in labor in which they found no overall reduction in maternal infectious morbidity. However, among women who subsequently had a cesarean birth (n = 116), there was an RR of 0.7 (95% CI 0.4, 1.4) for endometritis among women who received chlorhexidine irrigation during labor. They observed possible increased risk only in the group in which irrigation was done before rupture of membranes (n = 329); among those women, the authors found a twofold increase in maternal infection (95% CI 0.9, 4.4). It is difficult to postulate mechanisms to explain this counterintuitive increase in risk among those without rupture in the study by Rouse et al¹⁴ and those without labor in our study (RR 3.9; 95% CI 0.8, 18.0). Those observations might be statistical artifacts.

The central limitation of our study was the rarity of outcome events. At inception, the hospital rate of postcesarean endometritis, based on discharge summaries, was 18%; 10% of those who had cesareans had wound-healing complications, and our estimate for rates of postoperative fever was 20–25%. Rates in this study were low, closer to those low-risk populations described in the literature. Strict definitions are not likely the reason that we observed fewer cases of infection; hospital rates also decreased during that period. As a result, an absolute decrease of 10% in infection-related outcomes was unlikely, and we did not have adequate power to examine smaller variations in the outcomes of endometritis and wound separation by labor status. We postulate that the sharp decline in infection rate during the study might be related to increasingly uniform use of antibiotics for group B streptococcus prophylaxis for women with a history of positive cultures and for those with unknown status. Guidelines supporting intrapartum administration of antibiotics to patients at increased risk of delivering an infant who develops group B streptococcus were issued jointly by the ACOG and the American Academy of Pediatrics in June 1996.¹⁵ Forty percent of our cohort without chorioamnionitis received antepartum antibiotics. In addition, 75% of those without chorioamnionitis received a dose of prophylactic antibiotics after cord clamping. In total, 91% of participants received some antibiotic coverage at or before cesarean.

Other concerns include the lack of masking of care providers to intervention status. We assumed that if providers were aware of vaginal preparation status, they would tend toward treating fewer patients in the preparation group and treating more women in the no preparation group. That bias would favor finding the intervention effective. However, women with endometritis would eventually become apparent among those who were not initially treated, so any effect introduced by lack of masking would depend on overtreatment of the control group. Requiring a temperature of 38C or greater and considering only open wounds as wound-healing complications provided additional protection, as those are objective findings.

Ours was an intent-to-treat analysis. We did not seek validation that preparation was done if it was assigned and similarly did not seek evidence of elective crossover in those assigned to no preparation. Nonetheless, the presence of at least one other person in the operating room (the circulating nurse who opened the abdominal preparation kit) who knew the participant’s status improved the probability that the correct preparation was done.

We know of one other trial of vaginal preparation with 10% povidone iodine solution before cesarean that is unpublished and hope to identify others (Loffredo M, Schemmer G, Wade R, personal communication). Meta-analysis could clarify the association of vaginal preparation with possible decrease in wound-healing complications and allow investigation of effect modification by labor status. We do not assert that vaginal preparation does not provide benefit. A different antibacterial solution or a different method of preparing or timing of the preparation might have merit and warrants investigation.

Twenty percent of pregnancies in the United States are currently delivered by cesarean, totalling more than 800,000 per year.^16–19 Conservatively estimated, 80,000 to 120,000 women will have infection and wound-healing complications. Examination of risk factors for postpartum and postcesarean infection must continue as care patterns change, to refine our understanding of mechanisms that facilitate clinically important infection. Then we must seek evidence that preventive interventions are effective.

	Footnotes

 
PII S0029-7844(00)01087-5

Received February 14, 2000. Received in revised form June 19, 2000. Accepted July 7, 2000.

	References

Top Abstract Materials and Methods Results Discussion References

 
1. von Mandach U, Huch R, Malinverni R, Huch A. Ceftriaxone (single dose) versus cefoxitin (multiple doses): Success and failure of antibiotic prophylaxis in 1052 cesarean sections. J Perinatal Med 1993;21:385–97.[Medline]

2. Galask RP, Benigno BB, Cunningham FG, Elliot JP, Makowski E, McGregor J, et al. Results of a multicenter comparative study of single-dose cefotetan and multiple-dose cefoxitin as prophylaxis in patients undergoing cesarean section. Am J Surg 1988;155:86–90.[Medline]

3. McGregor JA, Gordon SF, Krotec J, Poindexter AN. Results of a randomized, multicenter, comparative trial of a single dose of cefotetan versus multiple doses of cefoxitin as prophylaxis in cesarean section. Am J Obstet Gynecol 1988;158:701–6.[Medline]

4. Duff P, Robertson AW, Read JA. Single-dose cefazolin versus cefonicid for antibiotic prophylaxis in cesarean delivery. Obstet Gynecol 1987;70:718–21.[Medline]

5. Roex AJ, Puyenbroek JI, van Loenen AC, Arts NF. Single-versus three-dose cefoxitin prophylaxis in cesarean section: A randomized clinical trial. Eur J Obstet Gynecol Reprod Biol 1987;25:293–8.[Medline]

6. Sachs BP, Yeh J, Acker D, Driscoll S, Brown DA, Jewett JF. Cesarean section-related maternal mortality in Massachusetts, 1954–1985. Obstet Gynecol 1988;70:385–9.

7. Raube K, Merrell K. Maternal minimum-stay legislation: Cost and policy implications. Am J Public Health 1999;89:922–3.[Abstract/Free Full Text]

8. Resnik E, Harger JH, Kuller JA. Early postpartum endometritis. J Reprod Med 1994;39:467–72.[Medline]

9. Watts H, Eschenbach D, Kenny G. Early postpartum endometritis: The role of bacteria, genital mycoplasmas, and Chlamydia trachomatis. Obstet Gynecol 1989;73:52–60.[Abstract/Free Full Text]

10. Faro S, Phillips LE, Martens MG. Perspectives on the bacteriology of postoperative obstetric-gynecologic infections. Am J Obstet Gynecol 1988;158:694–700.[Medline]

11. Gilstrap LC, Cunningham FG. The bacterial pathogenesis of infection following cesarean section. Obstet Gynecol 1979;53:545–9.[Abstract/Free Full Text]

12. Yancey MK, Clark P, Duff P. The frequency of glove contamination during cesarean delivery. Obstet Gynecol 1994;83:538–42.[Medline]

13. Monif GR, Thompson JL, Stephens HD, Baer H. Quantitative and qualitative effects of povidone-iodine liquid and gel on the aerobic and anaerobic flora of the female genital tract. Am J Obstet Gynecol 1980;137:432–8.[Medline]

14. Rouse DJ, Hauth JC, Andrews WW, Mills BB, Maher JE. Chlorhexidine vaginal irrigation for the prevention of peripartal infection: A placebo-controlled randomized clinical trial. Am J Obstet Gynecol 1997;176:617–22.[Medline]

15. American College of Obstetricians and Gynecologists. Prevention of early-onset Group B stretococcal disease in newborns. ACOG committee opinion no. 173. Washington DC: American College of Obstetricians and Gynecologists, 1996.

16. Curtin SC, Park MM. Trends in the attendant, place, and timing of births, and in the use of obstetric interventions: United States, 1989–97. National Vital Statistics Reports. Vol. 47, no. 27. Atlanta: Centers for Disease Control and Prevention, December 2, 1999.

17. Statistical abstract of the United States: 1998. Washington, DC: United States Bureau of the Census, 1998:83.

18. Graves E, Gillum B. National hospital discharge survey: Annual summary 1994. Vital and health statistics—series 13: Data from the National Health Survey. Hyattsville, Maryland: National Center for Health Statistics, 1997;13:39,128 (i–v).

19. Graves E, Owings M. 1995 summary: National hospital discharge survey. Advance data from vital and health statistics. Vol. 291. Hyattsville, Maryland: National Center for Health Statistics, 1997: 7–8.

This article has been cited by other articles:

				R. V. Starr, J. Zurawski, and M. Ismail Preoperative Vaginal Preparation With Povidone-Iodine and the Risk of Postcesarean Endometritis Obstet. Gynecol., May 1, 2005; 105(5): 1024 - 1029. [Abstract] [Full Text] [PDF]

				C. Pitt, L. Sanchez-Ramos, and A. M. Kaunitz Adjunctive Intravaginal Metronidazole for the Prevention of Postcesarean Endometritis: A Randomized Controlled Trial Obstet. Gynecol., November 1, 2001; 98(5): 745 - 750. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by REID, V. C. Articles by FRY, E. P. Search for Related Content PubMed PubMed Citation Articles by REID, V. C. Articles by FRY, E. P.