Antibiotic prophylaxis: Is there a difference? Sebastian Faro, MD, PhD, Mark G. Martens, MD, Hunter A. Hammill, MD, Gerry Riddle, MS, and Guillermo Tortolero, MD Houston, Texas Seven antibiotics, administered in 10 different regimens for prophylaxis, were randomly assigned to 1580 patients who were delivered by cesarean section. Cefazolin 1 gm, administered for three doses, served as the control group. Cefazolin 1 gm, cefazolin 2 gm, cefoxitin 1 gm, cefoxitin 2 gm, cefonicid 1 gm, cefotetan 1 gm, ceftizoxime 1 gm, ampicillin 2 gm, and piperacillin 4 gm were all administered in a single dose. Four antibiotics proved to be superior in preventing postpartum endometritis: ampicillin 2 gm (p = 0.03), cefazolin 2 gm (p = 0.005), piperacillin 4 gm (p = 0.0007), and cefotetan 1 gm (p = 0.0001). Single-dose cephalosporin antibiotic prophylaxis was found to result in approximately a twofold increase in Enterococcus faecalis colonization of the vagina (p < 0.01). This may be significant in patients in whom postpartum endometritis develops and who have failure of initial treatment with a broadspectrum cephalosporin, e.g., cefoxitin or cefotetan, or a combination such as clindamycin or metronidazole plus an aminoglycoside. Rupture of amniotic membranes for a half hour or more was associated with an increased risk for postpartum endometritis. The use of internal fetal monitoring was associated with an increased risk of soft tissue pelvic infection. (AM J OSSTET GVNECOL 1990;162:900-9.)

Key words: Antibiotic prophylaxis, enterococci, cefotetan, piperacillin, cefazolin, ampicillin Postpartum endometritis and wound infection are not uncommon complications after delivery by cesarean section, especially in women laboring with ruptured amniotic membranes for >6 hours. The incidence of postpartum endometritis ranges from a low of 5% in a typical private practice setting to as high as 85% in indigent patients. t. 2 The incidence of wound infection is dependent on many factors. These factors can be divided into two broad categories, namely, endogenous (obesity, chronic disease, vaginitis) and exogenous (poor surgical technique, improper preparation of surgical site, preoperative hospitalization > 1 day). One significant factor is whether the operative procedure is elective or emergency. The incidence of wound infection in elective versus emergency cesarean section is 3.8% versus 7.5%.3 Postpartum endometritis and wound infections are frequently polymicrobial, involving gram-negative and gram-positive aerobes and facultative and obligate anaerobes. 4 Wound infections occurring in patients who have not had prolonged labors tend to have a higher risk of Staphylococcus aureus as the primary pathogen, whereas those patients who have had a prolonged labor tend to have a higher incidence of polymicrobial wound infection. The bacteria involved reflect the micro flora of the lower genital tract. 5 From the Department of Obstetrics and Gynecology, Section of Infectious Diseases, Baylor College of Medicine. Presented by invitation at the Eighth Annual Meeting of the American Gynecological and Obstetrical Society, Hot Springs, Virginia, September 7-9, 1989. Reprint requests: Sebastian Faro, MD, PhD, Department of Obstetrics and Gynecology, Smith Tower, 6550 Fannin, Suite 701, Houston, TX 77030.

616118250

900

Antibiotic prophylaxis has proved to be effective in reducing the incidence of postcesarean soft tissue pelvic infection. Comparative studies have demonstrated that a single dose of antibiotic is as effective as administering three doses of antibiotics for prophylaxis. 6 Studies with broad-spectrum penicillins for prophylaxis also have been shown to be equally as effective.' These studies have not determined that one antibiotic is more efficacious than another, even when first- and secondgeneration cephalosporins, as well as penicillins, are compared, e.g., cefazolin, cefotetan, piperacillin, and cefoxitin. The present study was designed to address two issues: (1) In a comparison of the standard three doses of a first-generation cephalosporin (cefazolin) with a single dose of five cephalosporins and two penicillins, are one or more antibiotics superior in reducing the incidence of postcesarean pelvic infection? (2) Does a single dose of antibiotic exert select pressure on the vaginal microflora?

Material and methods Patients were eligible to enroll in this prospective, open, randomized study according to the following criteria: (1) in labor for >2 hours, (2) afebrile, (3) delivered by cesarean section, and (4) no antibiotics received in the previous 7 days. Individuals known or suspected to have an allergy to penicillin or cephalosporins were ineligible to participate in this study. All eligible patients gave written informed consent and were randomized to receive either a cephalosporin or a penicillin according to a computer-generated schedule. Antibiotics were administered intravenously, imme-

Superiority of single-dose antibiotic prophylaxis

Volume 162 Number 4

diately after the umbilical cord was clamped. Cefazolin was used in three different dosage regimens, 1 gm xl, 2 gm xl, and 1 gm X 3, given 8 hours apart. Cefoxitin was administered in doses of 2 gm x 1 and 1 gm X 1. Ceftizoxime, cefonicid, cefotetan, ampicillin, and piperacillin were given in a single dose (Table I). Cefazolin, 1 gm x 3 doses, was chosen as the control group because this is a standard regimen that is widely used by practicing obstetricians. Patients were considered to have failure of antibiotic prophylaxis if postpartum endometritis developed. The criteria for establishing a diagnosis of postpartum endometritis were: (1) a temperature of 2:: 100.4 F, measured on two occasions at least 4 hours apart, occurring 24 hours after the last dose of antibiotic; (2) a white blood cell count 2::14,000/mm 3 or >10% increase in immature polymorphonuclear leukocytes; (3) tachycardia; (4) marked uterine tenderness. Patients must have met all four criteria to have a diagnosis of postpartum endometritis. Specimens for the culture, isolation, and identification of aerobes and facultative and obligate anaerobes were obtained from the vaginal fornices at the time the decision was made to perform a cesarean section. Sterile cotton-tipped applicators were used to circumferentially swab the vagina. The specimens were immediately placed in anaerobic transport vials containing brainheart infusion broth. A second specimen was obtained 24 hours after the last dose of antibiotic was administered. A Pipe lie tube was used to obtain specimens from the inner surface of the uterine fundus of patients with endometritis." The specimen was placed in an anaerobic transport vial (Baltimore Biologic Laboratories, Cockeysville, Md.) and taken to the laboratory to be processed for the isolation and identification of aerobes, facultative and obligate anaerobes, Chlamydia, Mycoplasma, and Ureaplasma. 9 , 10 A Pipe lie tube was used because tissue, not lochia, could be obtained. This yielded organisms that are adherent to cells, and all the bacteria listed earlier could be obtained from one specimen. Therefore individual transport media were not required, nor was there a need for using multiple sampling devices. Ten milliliters of venous blood was aspirated from the volar aspect of the upper extremity from patients with postpartum endometritis. The aspirated blood was divided into two equal portions to inoculate aerobic and anaerobic blood culture bottles. Urine specimens were obtained either from an indwelling catheter or by straight catheterization and inoculated on typical media for the isolation of uropathogens. The following blood studies were done: complete blood count, white blood cell differential count, serum electrolytes, blood urea nitrogen, prothrombin time, and partial thromboplastin time. Repeat tests were performed as indicated by the patient's condition. Statistical analysis was performed by the following

901

Table I. Number of patients by antibiotic Patients Antibiotic

Cefazolin Cefazolin Cefazolin Ceftizoxime Cefonicid Cefotetan Cefoxitin Cefoxitin Ampicillin Piperacillin

Dosage

n

I gm x 3 I gm

142 217 161 145 147 148 155 162 148 155

2 gm I gm I gm I gm I gm 2 gm 2 gm 4 gm

% 9.0

13.7 10.2

9.2 9.3 9.4

9.8

10.3

9.4

9.8

0

methods: X" test for categoric data and t test and oneway analysis of variance for continuous variability. McNemar's test was performed for dependent dichotomous variables. Relative risk with 95% confidence intervals was determined to measure the association between hours of ruptured membranes and risk of infection. Results

A total of 1580 patients were enrolled into the study and randomized to one of 10 different antibiotic regimens (Table I). The difference between the largest treatment group, cefazolin 1 gm containing 217 patients, and the smallest group, cefazolin 1 gm x 3 containing 142 patients, is statistically significant. This is most likely due to the fact that the study is ongoing and the randomization is not complete. The population consisted of patients delivered on the Baylor obstetric service at Jefferson Davis and Lyndon Baines Johnson Hospitals. The study was started on Dec. 7, 1987, and this phase was concluded on July 1, 1989. The Jefferson Davis Hospital closed on May 30,1989, and all patients were transferred to the new facility, the Lyndon Baines Johnson Hospital. The entire hospital population consisted of the indigent patients of the Harris County Hospital district. There are between 15,000 and 16,000 deliveries per yer and the cesarean section rate is 20%. Hispanics (55.6%) made up the largest group, followed by blacks (25.9%), whites (17.2%), and others (1.3%). There were no significant statistical differences in the demographic and clinical baseline characteristics between treatment groups (Tables II and III). There was no statistical difference in the use of internal fetal monitoring between the treatment groups. Ampicillin, cefazolin 2 gm, cefotetan 1 gm, and piperacillin 4 gm, all administered as a single dose, were statistically superior to cefazolin 1 gm X 3 doses in preventing postpartum endometritis (Table IV). Among the patients receiving cefotetan, only 6.1 % had postpartum endometritis (X 2 = 20.6, P = 0.0001); in 8.4% of the patients receiving piperacillin an infection de-

902

Faro et al.

April 1990 Am J Obstet Gynecol

Table II. Baseline demographic data by prophylaxis group Prophylaxis group Characteristic

Race (%) White Black Hispanic Age (yr) (mean ± SD) Obstetric history Gravidity (mean ± SD) Parity (mean ± SD) Abortions (mean ± SD) Primary cesarean section (%) Repeat cesarean section (%) Brought to labor (%)

CeJazolin

CeJazolin

CeJazolin

1 gm

3

CeJoxitin

19m

2gm

17.3 24.3 57.9 23.6 ± 5.3

19.9 23.6 54.0 24.9 ± 5.7

17.9 27.9 52.1 24.6 ± 5.8

17.4 23.2 59.4 24.9 ± 6.0

2.3 ± i.7 1.9 ± 1.3 0.4 ± 1.1 66.4 33.6 21.2

2.4 ± 1.6 2.0 ± 1.3 0.4 ± 0.7 63.4 36.6 23.8

2.4 ± 1.7 2.0 ± 1.2 0.4 ± 0.8 76.8 23.2 19.0

2.5 ± 1.7 2.1 ± 1.4 0.4 ± 0.8 65.2 34.8 26.5

X

2gm

Table III. Baseline clinical characteristics by prophylaxis group Prophylaxis group Characteristic

Length of gestation (wk) (mean ± SD) Length of labor (hr) (mean ± SD) Length of rupture of membranes (hr) (mean ± SD) Operation time (min) (mean ± SD) Estimated blood loss (ml) (mean ± SD) Temperature (OF) (mean ± SD) White blood cell count (x 10') (mean ± SD) Hematocrit (%) (mean ± SD) Fetal monitor (%) None Internal-internal External-external External-internal

CeJazolin

CeJazolin

CeJazolin

1 gm

CeJoxitin

19m

2gm

38.9 ± 2.9

38.9 ± 2.9

38.6 ± 3.2

39.1 ± 2.5

14.8 ± 12.8

12.5 ± 10.3

12.2 ± 9.2

11.9 ± 9.8

9.2 ± 24.1

7.0 ± 7.6

7.2 ± 7.2

8.4 ± 20.6

54.2 ± 19.5

52.1 ± 16.2

52.3 ± 17.7

52.5 ± 18.5

747 ± 343

693 ± 264

753 ± 240

759 ± 332

98.4 ± 1.0 11.4 ± 3.8

98.5 ± 1.0 10.7 ± 3.4

98.6 ± 0.8 11.3 ± 3.9

98.6 ± 0.9 10.9 ± 3.9

35.4 ± 4.1

35.6 ± 3.3

35.9 ± 3.8

35.6 ± 4.1

10.1 39.9 41.1 8.9

5.8 46.0 46.0 2.2

4.3 46.6 42.8 6.3

veloped (X 2 = 21.5, P = 0.0007); among those receiving cefazolin 2 gm 10.6% of the patients became infected (X 2 = 22.9, P = 0.005); 12.8% of the patients receiving ampicillin 2 gm had infection (X 2 = 24.9, P = 0.03). In comparing infected with noninfected patients, the infected group differed only in that the duration of ruptured membranes was longer in the group that became infected, 12.3 versus 8.1 hours (p = 0.01) (Table V). In comparing the group of laboring patients with regard to the status of the amniotic membranes and risk of infection, it was found that the relative risk of infection increased with the duration of ruptured membranes, from 1.47 (95% confidence interval, 0.85 to 2.53) among patients with V2 to 3 hours of ruptured membranes to 2.81 (95% confidence interval, 1.9 to 4.15) among those with >12 hours of ruptured mem-

X

3

2gm

6.6 47.4 41.4 4.6

branes (Table VI), The use of internal monitoring, that is, a fetal scalp electrode and intrauterine pressure catheter, as compared with no monitoring, external monitoring only, and external pressure monitoring plus a fetal scalp electrode was associated with an increased risk of postpartum endometritis (p = 0.0001). There were no significant differences in the total number of bacterial isolates when preantibiotic and postantibiotic specimens were compared. In comparing preantibiotic and postantibiotic specimens no differences were noted among gram-positive aerobic isolates in any of the antibiotic groups (Fig. I). Patients receiving a penicillin showed approximately a twofold increase in the rate of isolation of gram-negative facultative anaerobes after treatment (Fig. 2). Except for piperacillin and ceftizoxime, antibiotic prophylaxis did

Superiority of single-dose antibiotic prophylaxis

Volume 162 Number 4

903

Prophylaxis group Cefoxitin

19m

Cefotetan

Ceftizoxime

Cefonicid

Ampicillin

Piperacillin

16.3 29.4 52.5 24.9 ± 5.3

13.5 27.0 58.8 24.4 ± 6.1

13.8 31.7 53.1 24.1 ± 5.8

22.4 19.7 57.1 24.8 ± 6.1

16.2 28.4 54.1 24.6 ± 5.7

17.4 24.5 56.1 25.1 ± 5.5

2.5 ± 1.9 2.1 ± 1.3 0.4 ± l.l 63.6 36.4 23.1

2.3 ± 1.4 2.0 ± 1.2 0.3 ± 0.8 64.2 35.8 21.8

2.4 ± 1.7 2.1 ± 1.4 0.3 ± 0.7 72.4 27.6 27.1

2.4 ± 1.5 1.9 ± 1.0 0.5 ± 1.3 67.3 32.7 23.1

2.3 ± 1.4 1.9 ± l.l 0.4 ± 0.7 66.2 33.8 23.0

2.4 ± 1.5 2.0 ± 1.8 0.3 ± 0.7 65.2 34.8 25.3

Prophylaxis group Cefoxitin

Cefotetan

Ceftizoxime

Cefonicid

Ampicillin

Piperacillin

39.1 ± 2.9

38.9 ± 2.9

38.8 ± 3.3

38.9 ± 2.8

39.1 ± 3.0

38.8 ± 2.4

13.8 ± 11.9

12.5 ± 10.7

13.5 ± 8.9

13.9 ± 12.8

12.8 ± 11.4

13.7 ± 10.2

9.0 ± 12.5

10.1 ± 28.6

9.8 ± 12.2

10.1 ± 29.2

8.5 ± 10.7

8.0 ± 9.8

53.5 ± 18.7

52.1 ± 15.1

56.2 ± 20.1

54.5 ± 19.3

51.8 ± 15.9

51.4 ± 21.5

720 ± 235

720 ± 215

769 ± 308

744 ± 244

744 ± 252

709 ± 212

98.6 ± 1.0 10.7 ± 3.4

98.5 ± 0.9 10.6 ± 3.3

98.7 ± 0.9 11.4 ± 5.3

98.5 ± 0.8 11.3 ± 3.9

98.6 ± 0.9 11.5 ± 4.6

98.5 ± 1.2 10.7 ± 3.8

35.1 ± 4.4

36.2 ± 3.8

36.1 ± 3.6

35.3 ± 3.4

35.0 ± 4.0

35.3 ± 3.6

7.6 41.4 44.6 6.4

7.6 49.3 38.2 4.9

5.6 46.5 43.8 4.2

5.0 39.3 50.0 5.7

7.7 44.4 45.1 2.8

5.8 41.6 45.5 7.1

19m

not appear to have an effect on anaerobic colonization (Fig. 3). However, when individual isolates were compared, the only bacterium that showed a significant difference between preantibiotic and postantibiotic isolation rates was Enterococcus Jaecalis. Overall, approximately 39% of patients were colonized by E. Jaecalis before administration of antibiotic and 62% of patients were colonized by this bacterium after prophylaxis. Comparing individual antibiotic groups, significant statistical differences in pretreatment and posttreatment isolation rates were observed for cefazolin 1 gm X 1 (P < 0.001), cefazolin 2 gm X 1 (P < 0.0001), cefazolin 1 gm X 3 (P = 0.04), cefoxitin 2 gm X 1 (p = 0.0002), cefoxitin 1 gm X 1 (p = 0.0001), cefotetan (p = 0.0002), ceftizoxime (p = 0.0063), and cefonicid (p = 0.0018). No difference in the preisolation and

postisolation rates of enterococci were observed in patients given ampicillin or piperacillin (Fig. 4). Although there were differences in the isolation rates of Lactobacillus before and after antibiotic, they did not appear to correlate with the development of postpartum endometritis (Fig. 5). No patients experienced an adverse reaction from receiving any antibiotic. Specifically, there were no anaphylactoid reactions or bleeding episodes caused by administration of any antibiotic.

Comment Antibiotic prophylaxis has been demonstrated to be effective in reducing the incidence of postpartum endometritis in patients delivered by cesarean section. Various antibiotics, both cephalosporins and penicillins,

904

Faro et al.

April 1990 Am J Obstet Gynecol

100~------~=-----------------------------~

80 60 %

40

20

o

AN-1

AN-2 AN-3 ME-1

ME-2

• Pre Antibiotic

TAN

ZOX MONO AMP

o Post

PIP

Antibiotic

Fig. 1. Percentage of gram-positive aerobic isolates obtained from the vagina in patients receiving antibiotic prophylaxis. Antibiotics given were as follows: cefazolin I gm x I (AN-i), 2 gm x I (AN-2), and I gm x 3 (AN-3); cefoxitin 2 gm x I (ME-I) and I gm x I (ME-2); cefotetan (TAN); ceftizoxime (ZOX); cefonicid (MONO); ampicillin (AMP); and piperacillin (PIP).

Table IV. Rate of infection by antibiotic prophylaxis group Infection

I

Table V. Demographic data on infected versus noninfected patients

No infection

I

%

n

32

22.5

110

77.5

44 17 24 27 9 26 22 19 13

20.3 10.6 15.5 16.7 6.1 17.9 15.1 12.8 8.4

173 144 131 135 139 119 124 129 142

79.7 89.4 84.5 83.3 93.9 82.1 84.9 87.2 91.6

Antibiotic

n

Cefazolin I gm x 3 Cefazolin I gm Cefazolin 2 gm Cefoxitin I gm Cefoxitin 2 gm Cefotetan I gm Ceftizoxime I gm Cefonicid I gm Ampicillin 2 gm Piperacillin 4 gm

%

Characteristic

p

Value

NS 0.005 NS NS 0.0001 NS NS 0.03 0.0007

including the new combinations containing [3-lactamase inhibitors, have been tested and all have been shown to be effective. The present study compared traditionally utilized antibiotics for prophylaxis, cefazolin and cefoxitin, and compared them with other agents, newer [3-lactams such as cefotetan, ceftizoxime, and piperacillin. Cefonicid was chosen because of its long half-life, 4V2 hours, and because it has been shown to be suitable for singledose prophylaxis. II Cefonicid was found to be equivalent to cefoxitin for prophylaxis in patients undergoing vaginal hysterectomy. In the present study cefonicid given in a single dose was less effective than a single dose of cefazolin 2 gm, ampicillin 2 gm, piperacillin 4 gm, or cefotetan I gm in preventing postpartum endometritis in patients delivered by cesarean section. The failure rate in patients receiving cefonicid was 15.1 % and was no different from that of cefoxitin 1 gm (15.5%) or 2 gm (16.7%). Cefazolin 1 gm, administered once and for three

Age (yr) Gravidity Parity Gestational age (wk) Length of labor (hr) Length of rupture of membranes (hr) Operating time (min) Blood loss (ml) Temperature (ml) C F) White blood cell count HCT(%) Preoperative Postoperative Fetal monitor External-external Internal-internal External-internal Primary cesarean section Repeat cesarean section

Infection*

N oninfection*

23.2* 2.1 1.7 38.8 15.4 12.3

24.8* 2.5 2.1 38.9 12.8 8.1

52.8 781 101.4 11,800

53.1 727.9 98.5 10,900

35.3 29.2

35.6 30.3

79 122 14 193 40

(11.8%) (17.9%) (16.9%) (18.3%) (7.7%)

592 558 69 864 482

(88.2%) (82.1%) (83.1 %) (81.7%) (92.3%)

*All values represent means. doses, resulted in 20.3% and 22.5%, respectively, of the patients having postpartum endometritis. However, cefazolin administered in a single 2 gm dose was effective in reducing the incidence of postpartum endometritis. The failure rate among patients receiving a single 2 gm dose of cefazolin was lO.6%, which was statistically better (X 2 = 22.96, P = 0.005) than that of cefazolin 1 gm X 3 doses. This difference in failure rates between the different doses of cefazolin is difficult to explain since the basic difference is probably a slightly higher serum concentration of the drug level when 2 gm is given. Theoretically, this may provide slightly longer serum levels and perhaps be more effective against organisms that are moderately sensitive to ce-

Superiority of single-dose antibiotic prophylaxis

Volume 162 Number 4

905

70~--------------------------------------~

60 50

40 %

30 20 10

o

AN-l

AN-2

AN-3 ME-l

ME-2

TAN

ZOX MONO AMP

o Post

• Pre Antibiotic

PIP

Antibiotic

Fig. 2. Percentage of gram-negative facultative anaerobic bacteria isolated from the vagina in patients before and 24 hours after administration of prophylactic antibiotics. For abbreviations see Fig. I legend.

60

40 %

20

o

AN-l

AN-2

AN-3 ME-l

ME-2

• Pre Antibiotic

TAN

l

ZOX MONO AMP

o Post

PIP

Antibiotic

Fig. 3. Percentage of anaerobic bacteria isolated from the vagina in patients before and 24 hours after administration of prophylactic antibiotics. For abbreviations see Fig. I legend.

fazolin, which results in a reduction of the inoculum size. The three-dose regimen was less effective; perhaps the prolonged serum concentration of the drug may have suppressed the sensitive organisms, thus allowing the more resistant bacteria to increase in numbers and therefore become the dominant organisms. Ampicillin, administered in a single 2 gm dose, had been previously shown to be effective when used for prophylaxis. In the present study ampicillin 2 gm also was found to be effective and the failure rate was 12.8%. A possible explanation for the slightly higher failure rate in this study may be that approximately 60% of the isolates of Escherichia coli from patients with postpartum endometritis are resistant to ampicillin. E. coli is one of the most frequently isolated gramnegative facultative anaerobes isolated from patients with postpartum endometritis and therefore is an important bacterium. If the patient population is colo-

Table VI. Relationship between ruptured amniotic membranes and postpartum endometritis Patients Length of rupture of membranes

o hr

0.5-3 hr 4-6 hr 7-12 hr >12 hr

Relative risk of infection

95%

n

%

Mean

Confidence interval

43 22 29 60 77

(30.5) (11.1) (11. 7) (24.1) (22.6)

1 1.4 1.75 1.77 2.42

0.85-2.53 1.15-3.11 1.27-2.9 1.9-4.15

nized predominantly by E. coli that are resistant to ampicillin, use of this antibiotic would result in selection for this bacterium. This would provide an opportunity for E. coli to become the dominant organism.

906

Faro et al.

April 1990 Am J Obstet Gynecol

80r------------------------------------------. 60

%

40 20

p =

o

AN-1

AN-2

Antibiotic prophylaxis: is there a difference?

Seven antibiotics, administered in 10 different regimens for prophylaxis, were randomly assigned to 1580 patients who were delivered by cesarean secti...
966KB Sizes 0 Downloads 0 Views