Antibiotic Prophylaxis in Vascular Surgery
ALLEN B. KAISER, M.D., KARL R. CLAYSON, M.D., JOSEPH L. MULHERIN, JR., M.D., ALBERT C. ROACH, PHARM. D.,
TERRY R. ALLEN, M.D., WILLIAM H. EDWARDS, M.D., W. ANDREW DALE, M.D.
Preoperative and intraoperative antibiotic prophylaxis of in-
fection in peripheral vascular surgery has been widely used
although controlled studies have been lacking. A randomized,
prospective, double-blind study of cefazolin versus placebo
during 565 arterial reconstructive operations was performed at
this hospital from February 1976 through August 1977.
Among the 462 patients undergoing surgery of the abdominal
aorta and lower extremity vasculature, there was a highly
significant difference in the infection rates: 6.8% for placebo
recipients versus 0.9% for cefazolin recipients (p <.001). Of
the 18 infections, four involved vascular grafts and all four
graft infections occurred in the placebo group. Over 8% of
abdominal wounds of patients receiving placebo became in-
fected versus 1.2% of cefazolin patients (p <.05). Groin
wounds were infected infrequently, 1.1% for placebo patients
versus none for cefazolin patients. No infections occurred
among 103 brachiocephalic procedures. Skin antisepsis was
analyzed retrospectively. Infection rates were significantly
higher (p <.01) following hexachlorophene-ethanol versus a
povidone-iodine skin preparation. Adverse effects of cefazolin
were carefully monitored: no rash, phlebitis, or emergence
of resistant strains was observed. A brief perioperative course
of cefazolin and povidone-iodine skin antisepsis are recom-
mended in vascular reconstructive surgery of the abdominal
aorta and lower extremity vasculature.
T HE EFFICACY OF PERIOPERATIVE
prophylaxis of wound infections following pe-
ripheral vascular surgery has been widely debated.
Controlled studies have been lacking, in part because,
as pointed out by Szilagyi, the low wound infection
rate associated with peripheral vascular surgery would
require a study ofvery large size to achieve statistically
significant results.16 In the absence of such controlled
studies, proponents of pre-, peri-, and postopera-
tive antimicrobial prophylaxis have emphasized the
high morbidity and mortality associated with graft
infections,7 and the success of prophylaxis in un-
controlled reports,10 and the reproducible efficacy of
antimicrobials in experimental studies.1,13'19Opponents
of unrestricted prophylaxis point out that the rela-
tively low infection rate in peripheral vascular surgery
Presented at the Annual Meeting of the American Surgical As-
sociation April 26-28, 1978, Dallas, Texas.
Reprint requests: Dr. Kaiser, P.O. Box 380, St. Thomas Hospital,
Nashville, Tennessee 37202.
From the Division of Infectious Diseases and Departments
of Surgery and Medicine, St. Thomas Hospital and the
Vanderbilt University School of Medicine,
does not justify the needless exposure of the vast
majority of patients to potentially toxic antimicro-
bials.16 Additionally, results from occasional clinical
studies have suggested that prophylactic antimicrobials
may not be efficacious in peripheral vascular surgery.6
Thus, Simmons and Stoley commented vigorously
that "the prophylactic use of antibiotics should undergo
the greatest scrutiny since this common use (especially
in surgery) is supported by very few appropriately de-
signed, randomized, controlled clinical trials.15
In an effort to resolve this important issue, a pro-
spective, randomized, double-blind study evaluating
the efficacy of cefazolin versus placebo in preventing
infections in peripheral vascular surgery was per-
The study was conducted at this institution from
1976 through August
scheduled to undergo vascular surgery were considered
eligible for participation in the study if:
abdominal aortic or peripheral vascular surgery was
performed; 2) no preoperative area of "wet gangrene"
or cellulitis was present; 3) there was no history of
severe penicillin allergy (anaphylaxis, wheezing or
exfoliative dermatitis) or of cephalosporin allergy; 4)
no preoperative antimicrobials had been administered;
5) written informed consent was obtained. Two pa-
tients did not enter the study because of refusal to
participate; one patient was excluded because of a his-
tory of a severe penicillin reaction; and only 18 pa-
tients (3% of the total) were inadvertently omitted
from the study. During the 19 month evaluation 565
patients completed the study. Antimicrobials were
given within 24 hours postoperatively in seven patients
1977. All patients
0003-4932/78/0900/0283 $00.85 © J. B. Lippincott Company
KAISER AND OTHERS
8. Hoffert, P. W., Gensler, S. and Haimovici, H.: Infection Compli-
cating Arterial Grafts. Personal Experience with 12 Cases
and Review of the Literature. Arch. Surg., 90:427, 1965.
9. Hunt, T. K. (Guest Editor), Alexander, J. W., Burke, J. F. and
MacLean, L. D.: Antibiotics in Surgery (A Panel by Corre-
spondence). Arch. Surg., 110: 148, 1975.
10. Lennihan, Jr.: Prophylactic Antibiotics in Arterial Surgery: A
Personal Experience Covering 218 Operations. Del. Med. J.,
11. Lowbury, E. J. L., Lilly, H. A. and Ayliffe, G. A. J.: Preopera-
tive Disinfection of Surgeons' Hands: Use of Alcoholic
Solutions and Effects of Gloves on Skin Flora. Br. Med. J.,
12. Madhavan, T., Quinn, E. L., Freimer, E., et al.: Clinical
Studies of Cefazolin and Comparison with Other Cephalo-
sporins. Antimicrob. Agents Chemother., 4:525, 1973.
13. Moore, W. S., Rosson, C. T. and Hall, A. D.: Effect of
fection of Vascular Prostheses. Surgery, 69:825, 1971.
14. Roberts, Jr., N. J. and Douglas, Jr., R. G.: Gentamicin Use
and Pseudomonas and Serratia Resistance: Effect of a Surgi-
cal Prophylaxis Regimen. Antimicrob. Agents Chemother.,
15. Simmons, H. E. and Stolley, P. D.: This is Medical Progress?
Trends and Consequences of Antibiotic Use in the United
States. JAMA, 227:1023, 1974.
16. Szilagyi, E. D., Smith, R. F., Elliott, J. P. and Vrandecic,
M. P.: Infection in Arterial Reconstruction with Synthetic
Grafts. Ann. Surg., 176:321, 1972.
17. White, J. J. and Duncan, A.: The Comparative Effectiveness
of lodophor and Hexachlorophene Surgical Scrub Solutions.
Surg. Gynecol. Obstet., 135:890, 1972.
18. Willerth, B. M. and Waldhausen, J. A.: Infection of Arterial
Prostheses. Surg. Gynecol. Obstet., 139:446, 1974.
19. Wilson, S. E., Wang, S. and Gordon. H. E.: Perioperative
Antibiotic Prophylaxis Against Vascular Graft Infection.
South. Med. J., 70:(S-1)68. 1977.
DR. WILEY F. BARKER (Los Angeles, California):
want to compliment Drs. Mulherin and Dale and their associates
for presenting the objective data to justify what many of us have
been doing on the basis of their prior advice several years ago,
with good apparent results, but most of us haven't had the docu-
I would like to submit the data which Mr. H. H. G. Eastcott, of
St. Mary's London, provided me with last week when I showed him
In 1976, St. Mary's Hospital had the first readout of a computer-
ized analysis ofabout seven prior years ofwork on abdominal aortic
aneurysm grafting. Their practice had been established as follows.
On the basis of the concept that a patient who had pathogens
growing in his nasopharynx would be more at risk for infection,
they had obtained cultures of the nose and throat, before operation.
patients with positive cultures
appropriate antibiotics on the regimen that Dr. Dale and his asso-
ciates have suggested; namely, given at the time of on-call medica-
tion. Other patients received none.
(Slide) As you will see, there were five graft infections in the
patient group that seemed to be less at risk, whereas there were
no graft infections in the 76 from which pathogens were grown.
Whether this is really a justifiable assumption or not, I'm not
Since this data came out in 1976, St. Mary's has gone to the treat-
ment of all patients on the regimen outlined by Dr. Dale, using
floxicillin. They now have 139 patients in their treated series, and
have had only one instance ofan infected graft; the patient is thought
to have been already infected at the time of operation, as the
patient was moribund, with a leaking aneurysm, in established
I have one question for Dr. Dale and his associates. I did not
clearly understand whether graft infections were counted separately
from the graft and wound infections which may have occurred
together. Can you clarify that finally in your discussion?
DR. ROBERT EDWARD CONDON (Wood, Wisconsin): This cer-
tainly has been a controversial subject. Prior to this morning's
presentation, I think it's fair to say that there was a small body of
experimental evidence which supported the concept of antibiotic
prophylaxis in connection with vascular grafting procedures, but
the only prospectively organized, blinded, controlled trial which
had been conducted in this area was the report of Evans and
Pollack, which included a small subset of vascular patients in a
much larger clinical study. The numbers were small, and the in-
fection rates were low. Those authors, when comparing cephalori-
dine with placebo, were unable to demonstrate that there was any
significant difference between treated and nontreated patients.
So I think it's not only Dr. Barker, but many others of us, who
are grateful to Dr. Dale and his colleagues for bringing forward
this morning this well-designed, prospective. blinded, well-con-
trolled clinical trial.
But I would like to raise one issue. The infection rate in the
placebo group over all was about 6%. Surprisingly, many of those
infections involved, not the groin wound, but the abdominal
wound where the infection rate was about 8%. Even if you eliminate
patients who got the ineffective hexachlorophene/alcohol prep.
the overall infection rate is still about 4% in the placebo group.
These infection rates seem to me to be a little bit high for clean
elective surgery, and since the conclusion of the study supporting
the administration of prophylactic antibiotics really depends pri-
marily on the infection rate in the control group, I'd appreciate
it if Dr. Mulherin or Dr. Dale could give us some further informa-
tion about the infection rate in the placebo group, in comparison
with their previous experience.
Is the infection rate experienced in this study representative
of their previous experience with vascular procedures. or is
some kind of an unusual phenomenon, related only to the study.
and perhaps not truly representative of the infection risk in pa-
tients undergoing vascular grafts?
prophylaxis in vascular surgery is a nearly universal practice. and al-
though originally its use was purely empirical, and although it is
still often abused, there is a respectable body of evidence in sup-
port of its rational, selective employment. The ultimate proof of its
value-that is, a randomized, prospective, double-blind evalua-
tion-has, however, been missing. The report we have just heard
is an account of the first attempt to provide this proof. a circum-
stance that clearly shows its great importance.
I had the opportunity to read the text of this report. for which
I am grateful to the authors, and in reading it
but ask myself the question: Has it, in fact. succeeded in providing
the definitive assessment of the value of these drugs, an assessment
we have so keenly been looking for?
There is no question that the study was devised on sound statisti-
cal principles, conducted with great care and reported with candor.
Nevertheless, the report has features that have aroused some con-
cern with respect to its ultimate meaning.
Time allows only the briefest indication of the more readily
visible problems of this type. I find it regrettable that the authors
included in the overall statistical treatment of the results the trivial
degrees of infection. What one
the incidence of infectious involvement of the prosthetic implant.
Inclusion ofthe minor infectious complications diluted the statistical
material, and, in addition, introduced a potential source ofnumerical
error, since noninfectious healing complications are often im-
possible to distinguish from primary infection.
I could not help
is exclusively interested