1. ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, June 1990, p. 1194-1197 Vol. 34, No. 6
0066-4804/90/061194-04$02.00/0
Copyright © 1990, American Society for Microbiology
Randomized, Double-Blind Comparison of the Efficacies, Costs, and
Vaginal Flora Alterations with Single-Dose Ceftriaxone and
Multidose Cefazolin Prophylaxis in Vaginal Hysterectomy
H. GRANT STIVER,1* BERNARD 0. BINNS,/3 ROBERT C. BRUNHAM,2'4 NICHOLAS CHENG,1
DEANNE M. DEAN,5 ANITA M. GOLDRING,1 JEANNE B. WALKER,' ELSIE TAN,' AND JUDITH McLEOD2
Division of Infectious Disease, Department of Medicine, University of British Columbia, and Vancouver General
Hospital, Vancouver, British Columbia VSZ 1M9,1 Departments of Medicine,4 Medical Microbiology,2 and Obstetrics
and Gynecology,3 Health Sciences Centre, Winnipeg, Manitoba R3E 3J7, and Medical Department,
Hoffmann-La Roche Canada Ltd., Etobicoke, Ontario M9C 5J4,5 Canada
Received 11 October 1989/Accepted 15 March 1990
A comparison of efficacies, costs, and effects on vaginal microflora of one preoperative and three
postoperative 1-g doses of cefazolin versus those of one preoperative l-g dose of ceftriaxone was done with 65
and 73 women, respectively, undergoing elective vaginal hysterectomy. Patient infection rates were not
statistically different between the cefazolin group (six events in 6 of 73 patients [8.2%]) and the ceftriaxone
group (11 events in 9 of 65 patients [13.8%]). Side effects, including diarrhea, were minimal and similar
between the two groups. Significant shifts in the cervicovaginal microflora of the patients occurred postoper-
atively, with a marked increase in enterococci and a drop in nonenterococcal streptococci. No shifts among
aerobic, facultative gram-negative rods and staphylococci were observed. Among the anaerobes, a significant
decrease in the number of patients harboring nonsporulating, gram-positive rods and a less striking
concomitant increase in Bacteroides species and members of the family Peptococcaceae were noted. No
qualitative differences were noted between the two groups that received prophylactic therapy. Aside from
enterococci, cefazolin or ceftriaxone resistance among vaginal isolates (2103/ml) was minimal. Selection of
resistant isolates was not different between the treatment groups. We could not detect a difference between a
single l-g dose of ceftriaxone and multidose cefazolin for infection prophylaxis in patients undergoing a vaginal
hysterectomy. However, the total acquisition, preparation, and administration costs were greater for the
ceftriaxone regimen than they were for the cefazolin regimen. Cefazolin should therefore remain the drug of
choice for infection prophylaxis in uncomplicated vaginal hysterectomies.
Infectious morbidity following vaginal hysterectomy has vaginal hysterectomy, the hypothesis being that the single-
been variously reported to occur in from 24 to 52% of dose regimen would be as effective as the multidose regimen.
patients, and routine use of perioperative prophylactic anti-
biotics has been shown to reduce this significantly to be- MATERIALS AND METHODS
tween 6 and 10% (3). Most prophylactic antibiotic regimens
have included one preoperative dose followed by two or Both premenopausal and postmenopausal women admit-
three postoperative doses. A few studies have shown that ted to the Vancouver General Hospital and the Winnipeg
the preoperative dose is the critical one and that it may not Health Sciences Centre for vaginal hysterectomy were en-
be necessary to continue prophylactic antibiotics postoper- rolled in the study after signed informed consent was ob-
atively (6, 8). Besides cost-saving considerations, lower tained. Patients were randomly assigned to receive cefazo-
levels of antibiotics may result in less of an alteration in the lin, 1 g within 1 h preoperatively followed by three 1-g doses
cervicovaginal microbial flora and a lower level of selection at 8-h intervals postoperatively, or ceftriaxone, 1 g within 1
of resistant organisms. h preoperatively followed by three saline infusions at 8-h
Ceftriaxone is a broad-spectrum cephalosporin with a intervals postoperatively, in a double-blind manner. Patients
prolonged half-life in serum of approximately 8 h. Its spec- were excluded from enrollment in the study if they had
trum includes most gram-positive cocci, except enterococci, received any antibiotics within the previous 14 days, were
and most gram-negative rods, except Pseudomonas species receiving any anti-inflammatory agents, were allergic to
and members of the family Bacteroidaceae (MIC, usually penicillin or cephalosporins, or declined to give consent.
2-16 ,ug/ml). Because both pharmacologic and clinical stud- Measurements of infectious morbidity included standard
ies confirm that ceftriaxone is efficacious in treating suscep- febrile morbidity, which was defined as two or more temper-
tible bacterial infections with once-daily dosing, we under- atures of 38°C or higher 4 h or more apart (based on 4-h
took a study of the efficacies, side effects, effects on vaginal measurements of oral temperature), excluding the first 24 h
bacterial colonization, and comparative costs of a single 1-g following the end of the surgical procedure, and clinical
preoperative dose of ceftriaxone versus those of a 1-g dose evidence of pelvic infection, which was defined as fever plus
of cefazolin preoperatively and three doses at 8-h intervals one or more of the following: purulent drainage from the
postoperatively in the prophylaxis of infection following vaginal cuff, abdominal pain associated with rebound tender-
ness and/or guarding, localized tenderness with a tender
adnexal mass on bimanual palpation, or bacteremia with or
without hypotension. Urinary tract infection was defined as
*
Corresponding author. symptoms of urgency, frequency, and dysuria with pyuria
1194

2. VOL. 34, 1990 CEFTRIAXONE VERSUS CEFAZOLIN IN VAGINAL HYSTERECTOMY 1195
(greater than or equal to eight leukocytes per high-power TABLE 1. Demographic data on prophylaxis groups
field) and bacteriuria of i103 CFU/ml on a catheter urine by the drug administered
specimen or .10W CFU/ml on a midstream urine specimen. No. aged: No. with concomitant:
Urinary tract colonization was recorded when a patient had Prophylactic Total no.
a positive catheter urine specimen of 2103 CFU/ml or a antibiotic of patients <50 yr 250 yr Medical Surgical
midstream urine specimen yielding i10' CFU/ml with or conditions procedure
without pyuria but did not have any of the symptoms of Cefazolin 73 40 33 13a 45b
urinary tract infection listed above. Patients were monitored Ceftriaxone 65 33 32 16C 47d
for a minimum of 6 weeks postoperatively for signs of a
Includes 10 patients with cardiovascular disease, 2 patients with neopla-
infection. sia, and 1 patient with diabetes mellitus.
Bacteriological measurements included a preoperative and b Includes 18 patients with anterior repair, 2 patients with posterior repair,
3- to 4-day postoperative swab of the vaginal vault and 23 patients with anterior and posterior repair, and 2 patients with oophorec-
catheter urine culture. In the event of postoperative fever, tomy or salpingo-oophorectomy.
c Includes 11 patients with cardiovascular disease and 5 patients with
blood and urine samples were obtained for culture and neoplasia.
antimicrobial susceptibility testing, including cefazolin and d Includes 14 patients with anterior repair, 6 patients with posterior repair,
ceftriaxone. In order to assess any qualitative shifts in 21 patients with anterior and posterior repair, and 6 patients with salpingo-
vaginal flora and the emergence of resistant organisms, the oophorectomy.
pre- and postoperative vaginal vault swabs, which were
transported in B-D VACUTAINERS (Becton Dickinson compared with 11 events in 9 of 65 ceftriaxone-treated
Vacutainer Systems, Rutherford, N.J.), were cultured both patients, 2 of which occurred after discharge. This gave a
aerobically and anaerobically. At one center (Winnipeg total patient infection rate of 13.8% (P = 0.197 by the Fisher
Health Sciences Centre), all aerobic and anaerobic flora exact test). Operative site-related infection rates and the
were identified and evaluated semiquantitatively as 1+, 2+,
incidence of febrile morbidity alone occurred in 4 of 73
3+, or 4+ growth. At the other center (Vancouver General cefazolin-treated patients and in 7 of the 65 ceftriaxone-
Hospital), only predominant flora were identified, as follows. treated patients (P = 0.203 by the Fisher exact test). In the
Vaginal swabs were pressed out in 1 ml of phosphate- cefazolin group there were four patients with febrile morbid-
buffered saline, diluted to 10-4, and cultured aerobically and ity only and no patients with localized infection. In the
anaerobically at 37°C on antibiotic-free blood agar. To assess ceftriaxone group, five patients developed febrile morbidity
the predominant resistant organisms, a portion of the initial only and two acquired a local infection (one patient had a
saline suspension was diluted to 10' and plated on separate vaginal cuff abscess and one patient had a wound infection).
blood agar plates containing cefazolin and ceftriaxone at 32 Symptomatic urinary tract infections occurred more fre-
p.g/ml. Any emerging colonies were identified and defined as quently in the ceftriaxone group. There were four lower
being resistant to the respective agent. urinary tract infections in ceftriaxone-treated patients, three
Qualitative differences in microflora and resistance pat- of which were enterococcal and one of which was a ceftri-
terns were compared between all preoperative and postop- axone-susceptible strain of Escherichia coli. Two cefazolin-
erative specimens and between cefazolin and ceftriaxone treated patients developed cystitis; one case was caused by
preoperatively and postoperatively. an Enterobacter species that was resistant to cefazolin, and
Cost analysis. The relative cost of each regimen was the other was caused by an enterococcus. Asymptomatic
determined by totaling the pharmacy acquisition cost for urinary tract colonization occurred in six cefazolin-treated
each regimen ($34.00 [U.S. $28.55] for 1 g of ceftriaxone and patients (Escherichia coli [n = 1], Proteus mirabilis [n = 1],
$10.00 [U.S. $8.40] for 4 g of cefazolin), the pharmacy Acinetobacter calcoaceticus [n = 1], and enterococci [n =
preparation cost ($3.00 [U.S. $2.52] per dose), and the 3]). Likewise, six ceftriaxone-treated patients were colo-
nursing administration cost (by estimating 10 min per admin- nized, and interestingly, all patients were colonized with
istration at an average hourly nursing wage of $16.25 [U.S. enterococci.
$13.65]). Given the low incidence of infection and the number of
For statistical analysis we used the chi-square and Fisher patients enrolled in the study, the power of the study to
exact tests. Assuming that the infectious morbidity in pa- detect a 50% difference was less than 50%. Therefore, we
tients who received antibiotic prophylaxis for vaginal hys- could not conclude that there were no differences in the
terectomy would be 8 to 10%, the number of patients efficacies of the two prophylactic regimens.
required to detect a 25 or 50% difference between the two Clinical side effects of the antibiotics were rare. Phlebitis
groups was 2,468 and 697, respectively.
occurred in one cefazolin-treated patient, and three patients
RESULTS
TABLE 2. Comparison of infectious events in patients receiving
Clinical outcome. Demographic data on the patients by cefazolin or ceftriaxone prophylaxis for vaginal hysterectomy
prophylaxis group are given in Table 1. There were no No. with the following
significant differences in any of the parameters examined. infectious complications:
Pre- and postmenopausal women were evenly distributed Prophylactic
within and between each of the two groups at approximately regimen Febrile Cuff or Urinary
morbidity pelvic tract Total
a 1:1 ratio. Mean preoperative patient weights were 65.6 ±
only infection infection
12.2 kg for the cefazolin group and 66.1 + 11.1 kg for those
in the ceftriaxone group. Four-dose cefazolin 4 0 2 6
Table 2 describes the relative distributions of postopera- (n = 73)
tive infectious complications. Overall, infectious events oc- Single-dose ceftriaxone 5 2 4 ga
curred six times in 6 of 73 cefazolin-treated patients, all (n = 65)
during hospitalization, for a patient infection rate of 8.3%, a Two patietns each had two infectious events.

3. 1196 STIVER ET AL. ANTiMICROB. AGENTS CHEMOTHER.
TABLE 3. Cervical microflora shifts observed in combined cefazolin and ceftriaxone groups before and after vaginal hysterectomy
No. of patients with the following organisms:
Anaerobes Aerobes
Any or all flora
Preoperative (n =47) 40 16 7 9 5 26 12 8 1 22 6
P value' <0.001 0.03 0.01 <0.01 o.02 <0.01
Postoperative (n =44) 17 25 18 9 2 26 28 1 0 7 6
Predominant flora
Preoperative (n =86) 49 8 9 9 11 21 5 9 6 25 8
P valuea <0.001 0.01 <0.01 <0.01
Postoperative (n =83) 15 12 9 13 15 19 16 0 0 15 3
a By the Fisher exact test or the chi-square test.
developed a rash, one of which was urticarial. No cefazolin- isolated from 89 of 133 patients from whom samples were
treated patients developed diarrhea. In the ceftriaxone obtained for culture, but these isolates were obtained from
group, no patients developed phlebitis, but three patients only 32 of 127 patients 3 to 4 days after surgery (chi-square,
developed a skin rash after drug administration, one of 45.4; P < 0.0001).
which was urticarial. Three patients developed diarrhea, and The aerobic cervicovaginal flora also underwent changes,
one of these patients had stool samples positive for Clostrid- particularly the streptococci. The most striking alteration
ium difficile toxin. was the increase in the frequency of isolation of enterococci.
The added cost of ceftriaxone at $34.00/g (U.S. $28.55/g) Enterococci were isolated from 17 of 133 (12.8%) patients
was not offset by the cost savings in nursing and preparation preoperatively, versus 44 of 127 (34.6%) patients postoper-
times ($5.70 [U.S. $4.80] versus $22.85 [U.S. $19.20]) in the atively (chi-square, 17.3; P < 0.001). This has been observed
four-dose cefazolin group. The total cost for each ceftriax- consistently in patients who receive preoperative cephalo-
one course was $39.71 (U.S. $33.35) compared with $32.83 sporin prophylaxis and is not an uncommon cause of post-
(U.S. $27.58) for each cefazolin course. operative urinary tract infections in obstetric and gyneco-
Cervicovaginal microflora shifts. Pre- and postoperative logic patients. In this study, enterococcal urinary tract
cervicovaginal culture rates were as follows: for predomi- infections or colonizations occurred in 13 patients, all of
nant flora (Vancouver General Hospital), samples from 46 of whom had postoperative enterococcal vaginal colonization.
47 patients in the cefazolin group were obtained for culture In contrast, 2 of 5 patients had urinary tract infections or
preoperatively, and samples from 46 patients were obtained colonizations with organisms other than enterococci and 29
for culture postoperatively. Of the 42 patients in the ceftri- of 109 patients had no urinary tract infections or coloniza-
axone group, samples were obtained from 40 of them for tions. The members of other strains of streptococci, includ-
culturing preoperatively and from 37 of them for culturing ing groups A and B, nonenterococcal group D, and viridans
postoperatively. For any and all growth (Winnipeg Health group, were reduced postoperatively.
Sciences Centre), samples from 25 and 24 of 26 cefazolin- In contrast to the distribution shifts in the aforementioned
treated patients were obtained for culture pre- and postop- organisms, there was no detectable difference preoperatively
eratively, respectively, and samples from 22 and 20 of the 23 versus postoperatively in either the genus composition or the
patients who received ceftriaxone were obtained for culture number of patients with facultative gram-negative rods or
pre- and postoperatively, respectively. staphylococci. A similar observation was made for Gardner-
Qualitative differences were noted between the pre- and ella vaginalis and Candida species.
postoperative vaginal vault swabs from both groups com- The selection of cervicovaginal organisms that were resis-
bined (Table 3). There was a general decrease in the number tant to the drug used for prophylaxis was not common, with
of anaerobic isolates obtained postoperatively from patients the exception of enterococci. Among 17 patients with en-
from the Vancouver center but not in those from the terococci preoperatively, 11 received cefazolin and 6 re-
Winnipeg Center. Since the prevalence of these strains was ceived ceftriaxone. Among 44 patients with vaginal entero-
not different preoperatively, this probably reflects a decrease cocci postoperatively, 19 received cefazolin and 25 received
in the counts below the 104/ml detection limit at the Vancou- ceftriaxone (pre- versus postoperative differences were not
ver center. Among the anaerobes, there was a sharp drop in significant). Of 46 cefazolin-treated patients, 7 patients har-
the number of nonsporulating, gram-positive rods, which bored cefazolin-resistant, nonenterococcal vaginal organ-
accounted for the majority of anaerobic strains isolated isms postoperatively (anaerobic lactobacillus [n = 1], facul-
preoperatively, and an increase in the number of isolates of tative gram-negative rods [n = 21, coagulase-negative
the families Bacteroidaceae and Peptococcaceae. Preoper- staphylococci [n = 3]), whereas 3 of 46 patients harbored
atively, anaerobic nonsporulating, gram-negative rods were these types of organisms preoperatively (anaerobic lactoba-

4. VOL. 34, 1990 CEFTRIAXONE VERSUS CEFAZOLIN IN VAGINAL HYSTERECTOMY 1197
cilli [n = 2], Bacteroides bivius [n = 1], aerobic corynebac- recently received cephalosporins and who may harbor resis-
teria [n = 1]). Of 37 ceftriaxone-treated patients, 4 had tant flora, agents such as ceftriaxone or cefotaxime may be
ceftriaxone-resistant organisms as part of their predominant useful. Taking into account the pharmacy's acquisition cost
vaginal flora postoperatively (Propionibacterium species [n for ceftriaxone, it is still not as cost-effective as a four-dose
= 1], coagulase-negative staphylococci [n = 2], Enterobac- cefazolin regimen, even though less preparation and nursing
ter cloacae [n = 1], and Corynebacterium species [n = 1]). time is required. Our regimen of four cefazolin doses is less
Of 42 patients from whom samples were obtained for culture conventional than that reported in most studies, in which
preoperatively, none had nonenterococcal ceftriaxone-resis- one preoperative and two 8-h postoperative doses of cefazo-
tant organisms as part of their predominant microbial flora. lin were used. The latter would make the cost difference
between the two regimens even greater. Although caution is
DISCUSSION advised in using agents for prophylaxis that may be needed
to treat actual infections (5), there is little evidence suggest-
We could not detect a significant difference between a ing that this is associated with postoperative treatment
single dose of 1 g of ceftriaxone given preoperatively and one failures in obstetrics and gynecology.
pre- and three postoperative doses of 1 g of cefazolin in
preventing infectious morbidity following vaginal hysterec- ACKNOWLEDGMENTS
tomy. However, as mentioned previously, given the number
of patients enrolled in this study, the power of the study was We thank the attending surgeons and house staffs of the Depart-
ments of Obstetrics and Gynecology, University of British Colum-
low and a beta error may have precluded the detection of a bia and University of Manitoba, for assistance in this study and
significantly greater efficacy of cefazolin over that of ceftri- A. W. Chow for advice.
axone, given the observed differences in the rates of infec- This work was supported by a grant from Hoffmann-La Roche
tion. We did not compare these therapies with single-dose Canada, Ltd.
cefazolin therapy. The latter has been shown to be an
effective prophylaxis regimen in patients undergoing vaginal LITERATURE CITED
hysterectomy compared with one preoperative and two 1. Briautigam, H. H., H. Knothe, and R. Rangoonwala. 1988.
postoperative doses of either cefazolin itself (10) or cepha- Impact of cefotaxime and ceftriaxone on the bowel and vaginal
loridine (6) or a single preoperative dose of cefoxitin or flora after single-dose prophylaxis in vaginal hysterectomy.
cefotaxime (4) or cefonicid (10). Thus, it is the least expen- Drugs 35(Suppl. 2):163-168.
sive regimen and has been advocated as the prophylactic 2. Chow, A. W. 1982. Antimicrobial therapy of gynaecologic
infections: an overview. J. Antimicrob. Chemother. 9(Suppl.
regimen of choice in patients undergoing vaginal hysterec- A):139-147.
tomy (4, 5). Despite this evidence, however, single-dose 3. Duff, P., and R. C. Park. 1980. Antibiotic prophylaxis in vaginal
cefazolin has not been approved for prophylaxis of infec- hysterectomy: a review. Obstet. Gynecol. 55(Suppl.):193S-
tions in patients undergoing vaginal hysterectomy either in 202S.
Canada or the United States (Canadian Health Protection 4. Hemsell, D. L., R. E. Bawdon, P. G. Hemsell, B. J. Nobles, E. R.
Branch and the U.S. Food and Drug Administration, per- Johnson, and M. C. Heard. 1987. Single-dose cephalosporin for
sonal communications, 1989). Some evidence exists that prevention of major pelvic infection after vaginal hysterectomy:
with cefotaxime prophylaxis, postoperative bacteriuria is cefazolin versus cefoxitin. Am. J. Obstet. Gynecol. 156:1201-
less (0.9%) with seven 1-g 12-h doses than it is with a single 1205.
5. Hemsell, D. L., M. C. Heard, B. J. Nobles, R. E. Bawdon, and
preoperative dose (7). P. G. Hemseli. 1987. Single-dose prophylaxis for vaginal and
The shifts observed in cervicovaginal flora agree with abdominal hysterectomy. Am. J. Obstet. Gynecol. 157:498-501.
those found in previous studies, with some minor differ- 6. Lett, W. J., R. Ansbacher, B. L. Davison, and W. N. Otterson.
ences. For example, we did not observe vaginal Candida 1977. Prophylactic antibiotics for women undergoing vaginal
overgrowth with ceftriaxone as has been reported previously hysterectomy. J. Reprod. Med. 19:51-54.
(1), nor did we see a significant increase in resistant aerobic 7. McDonald, P. J., R. Sanders, J. Turnidge, P. Hakendorf, P.
and anaerobic flora with ceftriaxone, as have other investi- Jolley, H. McDonald, and 0. Petrucco. 1988. Optimal duration of
gators who have examined resistance in fecal flora (9). This cefotaxime prophylaxis in abdominal and vaginal hysterectomy.
may partly be explained by our use of a 1-g instead of a 2-g Drugs 35(Suppl. 2):216-220.
8. Mendelson, J., J. Portnoy, J. R. DeSaint Victor, and M. Gelfand.
dose or our quantitative detection limit of 2103 resistant 1979. Effect of single and multidose cephradine prophylaxis on
organisms per ml. infectious morbidity of vaginal hysterectomy. Obstet. Gynecol.
Neither cefazolin nor ceftriaxone provided particularly 53:31-35.
good anaerobic coverage. This does not seem to be a 9. Michlea-Hazehpour, M., R. Auckenthaler, J. Kunz, and J. C.
requirement for effective prophylaxis in patients undergoing Pechere. 1988. Effect of a single dose of cefotaxime or ceftriax-
vaginal hysterectomy, even though, if an established infec- one on human faecal flora. A double blind trial. Drugs 35(Suppl.
tion develops, vaginal anaerobes are often involved (2, 11). 2):6-11.
For example, postoperative infection rates are no different 10. Soper, D. E., and A. L. Yarwood. 1987. Single dose antibiotic
when single-dose cefazolin and single-dose cefoxitin, a good prophylaxis in women undergoing vaginal hysterectomy.
Obstet. Gynecol. 69:879-882.
antianaerobic agent are used (4). 11. Sweet, R. L., and W. J. Ledger. 1979. Cefoxitin: single agent
We agree with previous suggestions that single-dose anti- therapy of mixed aerobic-anaerobic pelvic infections. Obstet.
microbial prophylaxis is desirable. In patients who have Gynecol. 54:193-198.