Addition of Parenteral Cefoxitin to Regimen of Oral
Antibiotics for Elective Colorectal Operations
A Randomized Prospective Study
DAVID J. SCHOETZ, JR., M.D., PATRICIA L. ROBERTS, M.D., JOHN J. MURRAY, M.D.,
JOHN A. COLLER, M.D., and MALCOLM C. VEIDENHEIMER, M.D.
The efficacy of cefoxitin, a perioperative parenteral antibiotic,
combined with mechanical bowel preparation and oral antibiotics
to prevent wound infections and other septic complications in
patients undergoing elective colorectal operations, was examined
in a prospective randomized study. All 197 patients who com-
pleted the study received mechanical bowel preparation and oral
neomycin/erythromycin base. In addition a perioperative par-
enteral antibiotic was given in three divided doses to 101 patients.
The other 96 patients received no parenteral antibiotics. The
overall incidence of intra-abdominal septic complications was
7.3% (7 of 96) in the control group (no cefoxitin) and 5% (5 of
101) in the treatment group (cefoxitin). This difference was not
statistically significant. The incidence of abdominal wound in-
fection was 14.6% in the control group and 5% in the treatment
group, a statistically significant difference (p = 0.02). The ad-
dition of perioperative parenteral cefoxitin greatly reduced the
incidence ofwound infections in patients undergoing elective co-
lorectal operations who had been prepared with mechanical bowel
cleansing and oral antimicrobial agents.
REPARATION TECHNIQUES OF the large intestine
for elective operations of the colon and rectum
have evolved so that the current standard of sur-
gical care mandates the use ofmechanical cleansing com-
bined with antimicrobial agents. Some debate exists re-
garding the relative importance of the preoperative ad-
ministration oforal antibiotics compared with the use of
parenteral antibiotics in the perioperative period.' Spe-
cifically the addition of parenteral antibiotics to the reg-
imen of mechanical cleansing and the use of oral anti-
biotics has not consistently reduced the number ofseptic
complications associated with elective colorectal proce-
dures.2'3 As a result of these inconsistencies and in light
ofincreased pressures for cost containment, surgeons are
being asked to reassess the common clinical practice of
combining oral and parenteral antibiotics as prophylaxis
in this situation.
Supported by a grant from Merck Sharp & Dohme, West Point, PA.
Address reprint requests to Dr. David J. Schoetz, Jr., Department of
Colon and Rectal Surgery, Lahey Clinic Medical Center, 41 Mall Rd.,
Burlington, MA 01805.
Accepted for publication October 30, 1989.
From the Department of Colon and Rectal Surgery, Lahey
Clinic Medical Center, Burlington, Massachusetts
Our aim was to determine the efficacy of the addition
ofa second-generation cephalosporin, cefoxitin (Mefoxin;
Merck Sharp & Dohme, West Point, PA), to the regimen
ofmechanical cleansing and the use oforal antibiotics in
a randomized prospective study. Cefoxitin is an antibiotic
with activity against gram-negative aerobes and anaerobes,
the most common pathogens isolated from infections as-
sociated with colorectal procedures.4 This feature makes
it one ofthe most commonly used prophylactic antibiotics
in colorectal procedures. Our study compares the rates of
wound infection with the rates ofintra-abdominal infec-
tions in two groups of patients. One group received ce-
foxitin in addition to mechanical preparation and oral
antibiotics, and the other group did not.
Materials and Methods
The study population consisted of 197 patients who
underwent elective colorectal operations at the Lahey
Clinic Medical Center between January 4, 1985, and
March 25, 1988. Prior approval for the study was obtained
from the Institutional Review Board ofthe Lahey Clinic
All patients 18 years of age and older who were un-
dergoing elective operations on the colon or rectum by
the abdominal route were potentially eligible. Previous
penicillin or cephalosporin allergy, use of oral or paren-
teral antibiotics within 72 hours before operation, use of
probenecid within 1 week before operation, and evidence
ofcolonic obstruction that would preclude adequate me-
chanical bowel preparation were criteria for exclusion. In
addition patients requiring parenteral antibiotic prophy-
laxis for cardiac valve disease and patients having in situ
foreign material, such as vascular grafts, prosthetic valves
or joints, and cardiac pacemakers, were not eligible.
SCHOETZ AND OTHERS
This study was conducted as an open-label randomized
observer-blinded trial. Randomization was achieved by
means of a multiplicative congruent linear method of
pseudorandom number generation using a Hewlett Pack-
ard Series 41 program SSTI-04A (Rockville, MD). Blocks
consisted of 10 patients randomized to cefoxitin and 10
patients to no cefoxitin. Sealed opaque envelopes were
used and were prepared in the Department of Research,
where they were numbered consecutively. Consent for
entrance into the study was obtained in writing from the
patient by physicians in the Department of Colon and
Rectal Surgery and the Department of General Surgery.
Four of five attending surgeons (JAC, JJM, DJS, and
MCV) and all patients were not informed of the results
All patients received a standard mechanical bowel
preparation and oral antibiotics in a schedule as outlined
in Table 1. The control group received no parenteral an-
tibiotics. The treatment group received three doses of ce-
foxitin, 2 g each, with the first dose administered within
60 minutes before incision, the second dose in the recovery
room at 6 hours, and the third dose 6 hours thereafter.
Shaving ofthe abdomen was performed within 60 min-
utes before the incision was made. The skin was prepared
with povidone-iodine (Betadine; Purdue Frederick Co.,
Norwalk, CT) solution. Paper drapes were used in all in-
stances. Adhesive skin barriers and wound protectors were
used for patients without a pre-existing stoma. Fascial
closures were accomplished with a single layer ofabsorb-
able monofilament sutures using a continuous mass clo-
sure technique. Subcutaneous sutures were not used, and
the skin was closed with stainless steel clips. No attempt
was made to standardize operative technique.
Wounds were inspected daily. For purposes of this
study, wound infection was defined as purulent wound
drainage; the exudate either spontaneously drained or was
expressed after removal of skin staples over a clinically
suspect area ofthe wound. Cultures were obtained when-
Other complications of infection were recorded and
managed appropriately according to the clinical circum-
stance. Intra-abdominal sepsis was diagnosed in patients
who had a fever after operation and for whom no other
TABLE 1. Bowel Preparation
Citrate of magnesia, 10 fluid oz.-Afternoon
Saline enemas until clear-Evening
Citrate of magnesia, 10 fluid oz.-Morning
Operation at 8 A.M.
1 g, erythromycin base,
1 P.M., 2 P.M., II P.M.*
*Times amended for later operation.
TABLE 2. Comparison ofControl and Treatment Groups
n = 96
Age, median (range)
Chronic ulcerative colitis
Patients taking steroids
*Five patients had two diagnoses.
source of infection could be demonstrated on complete
investigation or a documented intra-abdominal source by
radiographic investigation, such as in the case ofabscesses
within the peritoneal cavity or anastomotic leakage.
Records were reviewed by one of the authors (PLR).
Follow-up was obtained by office visits 1 and 4 months
after operation. Late complications of infection were re-
The randomization code was broken after data had been
compiled. Statistical analysis was carried out by paired
two-tail t test (program BMDP-3D) and Miettinen's
modification of the Fisher exact test and the Mantel-
Haenszel chi test described by Rothman and Boices using
aCALC41 System (Tacit Logic Systems, Inc., Tigard, OR)
in a microcomputer with 80287 coprocessor. Differences
at the level oftwo-tail p < 0.05 were considered statistically
Of the 225 randomized patients, 28 were withdrawn
from the study: because ofimproper dose of study drug,
10 patients; because of change in operative strategy, 9
patients; because of ineligibility that was not recognized,
7 patients; and due to patient request, 2 patients. Data
from the remaining 197 patients (126 men and 71 women)
were analyzed. The median age was 55 years (range, 18
to 96 years). The control group (no cefoxitin) included
96 patients, and 101 patients received cefoxitin. Com-
parisons between the groups and the indications for op-
eration are given in Table 2. The actual operative pro-
cedures are summarized in Table 3 for both groups. Car-
cinoma was the most common indication for operation
in this series. Data for septic complications directly at-
tributable to the colorectal nature of the operation are
presented in Table 4.
Wound infection occurred in 14 patientsin the control
group (14.6%) and in 5 patients in the treatment group
CEFOXITIN FOR COLORECTAL OPERATIONS
TABLE 3. Operations Performed
Low anterior resection
Total colectomy, ileorectal
Small bowel and colon
(5%). This difference is statistically significant (p = 0.02).
Because the number ofpatients taking steroids at the time
ofoperation was substantially higher in the control group,
this specific issue was addressed by analyzing data with
the Mantel-Haenszel chi test, controlling for steroid use
as a confounding variable. Again the incidence ofwound
infection was significantly higher in the control group
(p = 0.045).
Anastomotic leakage occurred in 4 of 78 (5.1%) anas-
tomoses in the control group and in 3 of 89 (3.4%) anas-
tomoses in the treatment group. This difference is not
statistically significant (p = 0.67). Of the seven instances
of leakage, four occurred in patients undergoing low an-
terior resection for either carcinoma ofthe rectum or vil-
lous adenoma; the other three occurrences were in patients
undergoing anastomosis for Crohn's disease.
Intra-abdominal sepsis was recognized in two different
forms. Pelvic cellulitis, characterized by fever, leukocy-
tosis, lower abdominal pain, and purulent drainage from
the ileal reservoir, occurred in two patients undergoing
construction of an ileoanal reservoir; one patient was in
the control group and the other patient was in the treat-
ment group. Cultures from the abdominal drain revealed
Escherichia coli in one patient and Staphylococcus aureus
and pseudomonas in the other patient. Based on these
cultures, specific antimicrobial therapy was administered
for 7 days without the need for operative intervention.
Pelvic abscesses occurred in two patients in the control
group (2%) and in one patient in the treatment group
(1%). Two patients required surgical intervention, and in
the third patient, the abscess resolved after spontaneous
drainage from a Hartmann pouch. The difference between
the two groups in overall rates of the incidence of intra-
abdominal sepsis is not statistically significant (p = 0.68).
No patient experienced substantial adverse effects from
either the oral antibiotics or the cefoxitin.
The results of bacteriologic study are outlined in Ta-
Mechanical cleansing of the colon before elective co-
lorectal operations has long been recognized as an essential
part of preoperative preparation.6 Because the colon is a
rich reservoir ofbacteria, purgation offecal contents sub-
stantially reduces the number of potentially pathogenic
With the advent of antimicrobial agents, oral prepa-
ration was added to mechanical cleansing in an attempt
to reduce further the incidence of septic complications.
In 1953 Poth7 reviewed the data on available drugs and
concluded that neomycin was an excellent oral agent be-
cause of its poor absorption from the gastrointestinal tract
and consequent low toxicity. He noted overgrowth ofyeast
in the intestine with its use. Studies by Cohn and Rives8
demonstrated that the addition ofintraluminal neomycin
and tetracycline decreased the incidence of septic and
anastomotic complications in an experimental model.
Nichols and associates9 documented appreciable reduc-
tion in intraluminal bacteria by the use oforal antibiotics.
In fact, as a result ofthis study, the authors9 demonstrated
the need for drugs directed at both aerobes and anaerobes.
Based on their results, a combination of neomycin and
erythromycin base in the dosage schedule depicted in Ta-
tively," and a meaningful reduction in the number of
septic complications resulted when these antibiotics were
used with mechanical cleansing.
Despite the fact that the neomycin/erythromycin base
1 was studied retrospectively'0 and applied prospec-
TABLE 4. Operation andSpecific Septic Complications
Control (No Cefoxitin)
VOl. 212 * NO. 2
Ann. Surg. * August 1990
TABLE 5. Bacteriology ofInfected Wounds
Control (no cefoxitin): 9 of 15 wounds
Gram-negative rods (other)
Treatment (cefoxitin): 5 of 5 wounds
(the Nichols-Condon prep) has become the most widely
used oral antibiotic regimen, any nonabsorbable combi-
nation or single agent that is highly effective against both
aerobic gram-negative rods and anaerobes has also been
shown to be effective for colorectal cleansing before op-
With the development of parenteral antibiotics, at-
tempts were made to substitute these agents for oral agents.
Results of studies up to 1983, as reviewed by Guglielmo
and associates,'3 were mixed. Antimicrobial agents with
a spectrum limited to either aerobes or anaerobes alone
are usually ineffective in reducing septic complications of
colorectal operations when administered parenterally as
a substitute for oral antibiotics.
Crucial to our study was the question ofthe benefit of
combining parenteral and oral antibiotics to confer ad-
ditional protection against septic complications compared
with the benefit ofeither regimen alone. In recent analyses
ofavailable studies, the majority conclusion has been that
'combining oral and parenteral drugs should be unnec-
essary in most patients.'"3 Unfortunately most ofthe an-
tibiotics administered parenterally were not active against
both aerobes and anaerobes, and, consequently, the con-
clusions ofthese analyses are questionable.
Cefoxitin is a second-generation cephalosporin that is
bactericidal against a wide range of potential colonic
pathogens, including gram-positive cocci, gram-negative
bacilli, and anaerobes.' Hoffmann and associates'4 re-
ported a small series in which cefoxitin was randomly
added to one half of a group of patients undergoing me-
chanical preparation and oral kanamycin (Kantrex; Bris-
tol Laboratories, Syracuse, NY) therapy. They'4 dem-
onstrated appreciable reduction in rates ofwound infec-
tion with the addition ofparenteral cefoxitin. A potential
weakness of this study'4 was the deletion of an oral an-
tianaerobic agent in the control group. Consequently the
favorable results may have been the result ofthis omission
rather than because of the addition of cefoxitin.
Our study was designed to compare numbers of septic
complications in two groups ofpatients undergoing elec-
tive colorectal operations by means of the abdominal
route. Specifically cefoxitin was added in a randomized
prospective manner to the most widely used oral antibiotic
combination that has been proved highly effective. Results
indicate that the addition ofcefoxitin confers appreciably
more protection against the development of wound in-
fection (5% versus 14.6%; p = 0.02). On the other hand,
anastomotic leakage and intra-abdominal sepsis were not
affected by addition of cefoxitin. The former is usually
the result ofa technical failure and would not be expected
to be altered substantially by prophylactic antibiotics; the
latter might be expected to be affected, but the number
of occurrences in the two groups is too small to draw
statistically valid conclusions.
Our data support the addition of perioperative paren-
teral cefoxitin to the regimen ofmechanical cleansing and
the use oforal neomycin/erythromycin base to reduce the
risk ofabdominal wound infection in patients undergoing
elective colorectal operations.
The authors thank Elton Watkins, Jr., M.D., Sias Surgical Research
Unit at the Lahey Clinic, for randomization and data analysis.
1. Keighley MRB, Arabi Y, Alexander-Williams J, et al. Comparison
between systemic and oral antimicrobial prophylaxis in colorectal
surgery. Lancet 1979; i:894-897.
2. Barber MS, Hirschberg BC, Rice CL, Atkins CC. Parenteral anti-
biotics in elective colon surgery? A prospective, controlled clinical
study. Surgery 1979; 86:23-29.
3. Condon RE, Bartlett JG, Greenlee H, et al. Efficacy of oral and
systemic antibiotic prophylaxis in colorectal operations. Arch Surg
4. Neu HC. Cefoxitin: an overview of clinical studies in the United
States. Rev Infect Dis 1979; 1:233-239.
5. Rothman KJ, Boice JD Jr. Epidemiologic Analysis with a Program-
mable Calculator, New edition. Boston: Epidemiology Resources,
6. Nichols RL, Condon RE. Preoperative preparation of the colon.
Surg Gynecol Obstet 1971; 132:323-337.
7. Poth EJ. Intestinal antisepsis in surgery. JAMA 1953; 153:1516-
8. Cohn I Jr, Rives JD. Protection of colonic anastomoses with anti-
biotics. Ann Surg 1956; 144:738-752.
9. Nichols RL, Condon RE, Gorbach SL, Nyhus LM. Efficacy ofpre-
operative antimicrobial preparation ofthe bowel. Ann Surg 1972;
10. Nichols RL, Broido P, Condon RE, et al. Effect of preoperative
neomycin erythromycin intestinal preparation on the incidence
of infectious complications following colon surgery. Ann Surg
11. Clarke JS, Condon RE, Bartlett JG, et al. Preoperative oral antibiotics
reduce septic complications of colon operations: results of pro-
spective, randomized double-blind clinical study. Ann Surg 1977;
12. Keighley MRB. Prevention and treatment ofinfection in colorectal
surgery. World J Surg 1982; 6:312-320.
13. Guglielmo BJ, Hohn DC, Koo PJ, et al. Antibiotic prophylaxis in
surgical procedures: a critical analysis ofthe literature. Arch Surg
14. Hoffmann CEJ, McDonald PJ, Watts JM. Use ofperioperativece-
foxitin to prevent infection after colonic and rectal surgery.Ann
Surg 1983; 193:353-356.
SCHOETZ AND OTHERS