Wound infection after arterial surgical procedures

Eur J VascSurg 6, 494-498 (1992)

Wound Infection after Arterial Surgical Procedures F. J. G. van H i m b e e c k 1, L. A. A. van K n i p p e n b e r g 1, M. C. G. H. N i e s s e n 2 and A. J. A. van Griethuysen 2

Departments of 1Surgery and 2Microbiology, Canisius Wilhelmina Hospital, Nijmegen, The Netherlands During the period October 1983 to March 1987, 603 patients who underwent arterial surgical procedures were studied to determine the incidence and treatment of wound infections. Bypass procedures were performed in 395 patients (65.5%), in which autogenous vein was used for 158 grafts (26%), synthetic Dacronfor 216 grafts (36%), and umbilical vein for 21 grafts (3.5%). Thrombo-endarterectomies, embolectomies and patch-grafts were performed in 208 patients (34.5%). An Infection Control Nurse examined and registered the wounds. The definition of wound infection used in our study is equivalent to Szilagyi grade II infection. Vascular surgery is classified as clean surgery, the clean wound infection rate being a useful measurement to evaluate preventative measures and surgical technique. Thirty-one patients (5.1%) developed a wound infection as a postoperative complication. The overall incidence of wound complications including haematoma and seroma following arterial reconstruction was 13 %. The site of wound infection was predominantly the groin. The most common pathogen was Staphylococcus aureus which was found in 17 patients (2.8%). All infections resolved without further surgical intervention. The influence of possible aetiologicalfactors is considered and the importance of prophylactic antibiotics and good surgical technique is stressed. Key Words: Arterial procedures; Clean wound infection rate; Preventive measures; Surgical technique.

Introduction

Materials and M e t h o d s

Wound infections are a frequent postoperative complication and are of particular interest to the surgeon and microbiologist as they are an important cause of morbidity, longer hospital admission and excess hospital costs. This is particularly true in vascular surgery where infection may lead to the loss of an extremity or death. 1_7 Vascular surgery is classified as clean surgery. According to the 1985 Centres for Disease Control (C.D.C.) guidelines for the prevention of surgical wound infection, the clean w o u n d infection rate is 1-5% .8 A zero infection rate is not possible due to contamination via endogenous or exogenous routes. In this paper we analyse a prospective w o u n d infection study of 603 consecutive patients who underwent arterial reconstructive surgery. The results are compared with findings in the literature and have provided the information required to evaluate preventive measures and surgical techniques in our department.

During the period October 1983 to March 1987, 603 patients (194 women, 409 men) undergoing arterial surgical procedures were studied. Five hundred and ten (84.6%) were elective and 93 (15.4%) emergency procedures. The indications for surgery were acute thrombo-embolic pathology, arterio-sclerotic occlusive disease and aneurysm. Most of the operation s were performed by the two vascular surgeons in our department with the patients under general or local anaesthesia. All patients were operated on in conventional operating theatres and cared for in general surgical wards. Tincture of iodine was used, as a preoperative antiseptic skin preparation except in cases of allergy when a red chlorhexidine gluconate solution was used. Cloth drapes were standard and exposed skin was completely covered with adhesive plastic drapes (steridrape). Routine antibiotic prophylaxis of l g cefazoline was given I h preoperatively and for surgery lasting longer than 4 h a further I g of cefazolin was administered. In those cases where an emergency laparotomy was performed, 500 mg metronidazol was given as a supplement and in elective cases where intra-abdominal surgery was required patients

Please address all correspondence to: F. J. G. van Himbeeck, Department of Surgery, CanisiusWilhelminaHospital, St Annastraat 289, Postbus 9015, 6500GS Nijmegen,The Netherlands. 0950-821X/92/050494+05 $08.00/0© 1992Grune & Stratton Ltd.

Wound Infection

received a regime of systemic antibiotics to decontaminate the digestive tract. Reversed autogenous vein grafts, harvested through interrupted as opposed to continuous incisions were used for peripheral arterial reconstruction where possible. When operating in the groin, a vertical incision was preferred with a lateral approach to minimise lymphatic interruption. During aorto-iliac bypass surgery the median abdominal incision was combined if necessary with a horizontal incision at the groin site and less frequently with a suprainguinal oblique incision for a retro-peritoneal approach. Wound infection was defined as a purulent discharge or leakage from the w o u n d edges which showed a positive culture for pathogenic microorganisms with signs of inflammation. Whenever these signs were absent, and the culture was negative, this secretion was listed as a seroma. A haematoma was defined as a localised mass of extravasated blood within the wound. Swabs were taken from purulent discharge or secretions; and drains were cultured aerobically and anaerobically according to standard bacteriological procedures. The definition used for wound infection is equivalent to Szilagyi and associates' grade II, 1 in which the subcutaneous region is involved. As to whether the infection had extended to the arterial conduit (grade III infection) was judged by the responsible surgeon. Registration and examination of the wounds postoperatively was carried out by an Infection Control Nurse. Once patients were discharged, they were routinely followed up as outpatients by the two attending vascular surgeons in collaboration with the Infection Control Nurse. The follow-up was concluded by December 1989 and the data reviewed and analysed by means Of the chi-squared and Wilcoxon tests.

Results

Bypass procedures were performed in 395 patients: 154 (25.5%) had an aortic interposition and bifurcation prostheses and 241 (40%) had ilio-femoral and infra-inguinal grafts. In 158 cases of reconstruction, autogenous vein was used, compared with 216 cases in which synthetic Dacron was the material of choice. At the beginning of this study 21 patients (3.5%) received umbilical vein grafts as implants. Thrombo-endarterectomies, embolectomies and patch grafts were performed in 208 patients (34.5%; Table 1).

495

Table 1. Type of implant and incidence of wound infection

No. of operations Incidenceof (%) infectiorl(%) Syntheticprostheses (Dacron)

216 (36)

11 (5)

Autogenous vein grafts

158 (26)

13 (8.2)

21 (3,5)

1 (4.8)

Venous allografts(Dardik)

Endarterectomies, embolectomy, patchgrafts 208 (34.5) Total

603 (100)

6 (2.9) 31 (5.1)

Postoperative wound infection occurred in 31 (5.1%) of the 603 patients. Of these, 17 (out of 194, 8.8%) were female and 14 (out of 409, 3.4%) males. Females significantly outnumbered the males (p 0.01) and the men with infection were older than the men without (mean age infection group 70.6 years vs. 64.2 years for the non-infected group; p < 0.01). This correlation was absent in the female patient group (mean age of the infection group was 68.6 years vs. 68.0 years). Staphylococcus epidermidis was isolated from 25 drains and six wounds, usually in low counts. These were not included in the total of w o u n d infections as there were no signs of infection in these patients. No graft infections occurred. Wound haematoma was seen in 23 cases (3.8%) and seroma in 26 cases (4.3%). The overall incidence of w o u n d complication after arterial reconstruction was 13% (Table 2) and none of these wound complications extended to the graft. No deaths were attribuTable 2. Type of wound complications

No. of patients

%

Seroma

26

4.3

Haematoma

23

3.8

Wound infection(grade II)

31

5.14

Graft infection(grade III)

0

0

table to w o u n d complications and the operative (30 days) mortality rate was 1.7% (10). The cause of death was cardio-pulmonary failure (5 cases), pulmonary embolus (two cases), ruptured thoracic aortic aneurysm (1 case), impaired neurological function (1 case) and disseminated cancer of the gastrointestinal tract (1 case). Most of the operations (70.5%) were performed by the two attending surgeons, one of whom operated on 259 patients who had 11 wound infections (4.2%) compared with 166 patients operated upon by Eur J Vasc Surg Vol 6, September1992

496

F . J . G . van Himbeeck et aL

the other, with 10 wound infections (6%) (N.S.). The rest of the team, mostly final year residents operated upon 158 patients, resulting in 10 w o u n d infections (5.6%). The site of the w o u n d infection was predominantly the groin (Table 3). Of the 452 incisions at this

Table 4. Culture results (mixed infections occurred in 25 patients with positive cultures, resulting in a total of 33 microorganisms)

Gram-positive

Gram-negative 17

Escherichiacoli

4

Enterococcus

2

Proteus mirabilis

3

Streptococcus agalactiae

1

Morganella morganii

1

Enterobacter

3

Shigella

1

Pseudomonas

1

Staphylococcus

Table 3. Occurrence of infection by site of incision

Site of i n c i s i o n

Infections/number of incisions

%

Abdominal median

0/170

0

Suprainguinal oblique

2/54

3.7

Groin

27/45

26

Leg

1/208

0.5

Others (cubitalfossa)

1/28

3.6

site 27 (6%) became infected as compared with 414 incisions at the remaining sites, where four (0.8%) infections occurred (p < 0.01). Also seroma and haematoma occurred mainly in the groin. There is no significant relation between the type of graft or type of operation and infection (Table 1). Four factors were analysed as potential predictors of surgical w o u n d infection: diabetes mellitus, hypertension, obesity and smoking and none of these correlated with a greater risk of infection. Four patients out of 37 with a remote distal infection and 27 patients out of 566 without distal infection were complicated by a w o u n d infection (p K 0.10). The microorganisms cultured from the remote infection site and from the wound were identical in only two of these four patients. In one case S. aureus, and in the other case Streptococcus agalactiae, was found at both sites. Cultures were sterile in six of the 31 w o u n d infections. On the other 25 cases 33 different microorganisms were cultured; in four mixed Gram-positive and Gram-negative infections were found. Gram-positive cocci, mainly S. aureus, accounted for 61% and Gramnegative rods for 39% (Table 4). In most cases the wound infections were treated by local wound irrigation and where necessary appropriate antibiotics for the organism(s) cultured. All wound infections resolved without further surgi~ cal intervention. The follow-up was carried out until December 1989, resulting in a mean follow-up time of 45 months (range 21-75 months). Ninety patients (14.9%) were lost to follow-up. From the 90 patients, 50 were randomly selected and follow-up information was Eur J VascSurg Vol 6, September1992

Total

20

13

There were six sterile cultures.

obtained. Six patients had died without obvious signs of infection reported.

Discussion

The number of w o u n d infections amounted to 5.1%, this being the upper limit as stated by the C.D.C. for clean wounds. 8 The overall incidence of w o u n d complications following arterial reconstruction including seroma and haematoma was 13%. In an extensive review of 62 939 surgical wounds at Foothills Hospital, Cruse and Foord 9 noted that the infection rate for clean wounds was 1.5%. The subset of clean vascular wounds demonstrated an incidence of 3% and was the highest infection rate of the various departments. Holm demonstrated that different w o u n d infection rates can be found in the same patient group according to the definition used and because of different definitions it is difficult to compare the rate of postoperative w o u n d infections between various clinics. 10 Our chosen method of definition agreed with the objective method for w o u n d infections defined by Szilagyi et al. In our study definition was carried out by one full-time Infection Control Nurse, resulting in a consistent application of criteria. The true incidence of graft infection is difficult to establish due to its low incidence and because manifestation of this type of infection can occur many years after implantation. The reported incidence of prosthetic graft infections ranges from 0.9 to 6.0%. 1"3-6"11'12 In series where autogenous vein grafts were used, the incidence of grafts infections was even lower, such as the 0.4% reported by Szilagyi. 1 In our series at the time of the conclusion of this study in December 1989 the incidence of graft

Wound Infection

infection is zero. This is unexpected considering the number of patients (603) involved and the length of the follow-up period (mean 45 months). Low rates of wound-infection depend on a number of measures: preoperative skin preparation, antibiotic prophylaxis, and surgical technique are considered the most important. The skin preparation consisted of shaving, the application of tincture of iodine and adhesive plastic drapes. Cruse and Foord 9 have shown that shaving the operation site increases the infection rate of clean wounds. Other authors have stressed the importance of shaving immediately prior to surgery to prevent bacterial growth in the razor nicks. 13 The latter method was used in our study. From a retrospective study reported by Kaiser et al. 14 w e know that an iodine-based agent for skin antisepsis is preferred to a hexachlorophene ethanol agent and achieves better results. Many authors have found that the use of adhesive plastic drapes did not reduce the infection rate of clean wounds but we find the continuing use of these justified to avoid skin contamination of the graft. The routine use of preoperative antibiotics is supported by the work of Kaiser et al. 14 In a prospective study they demonstrated a highly significant difference in the infection rates among 237 placebo recipients (6.8%) compared with 225 cefazolin recipients (0.9%) (p < 0.001). The two infections occurring in the cefazoline group were grade II infections. It was striking that all four graft infections (grade III) occurred in the placebo group. In operative procedures carrying a risk of infection by anaerobic organisms from the gastrointestinal tract the prophylaxis mentioned above was supplemented with metronidazole and gastrointestinal decontaminants. Of great importance in the prevention of infection is a careful surgical technique together with gentle handling of tissues and attention to w o u n d closure. A poor technique leads to further contamination of the wound, lengthening of operation time, wound haematoma, leaking of lymph, and devitalised tissue. In this study the frequency of complications for each operating surgeon was not statistically different. In thirty-one patients with grade II infections nine different organisms were isolated; and in six cases sterile cultures were found. As would be expected, 3'4'6"11"15'18 the leading pathogen was S. aureus, found in 17 patients, but increasing numbers of Gram-negative organisms are also being reported 4,18 and occurred in 39% of our patients. Staphylococcus epidermidis was often isolated, as reported before, especially from incisions in the groin, without

497

causing infection, but it is a potential threat for grafts. We analysed a number of risk factors with regard to the class II infections. The groin was the predominant region for w o u n d infection which has also been reported in other series. 1"3"5-7'15-17 What makes the groin so vulnerable to infection? It is difficult to prepare and drape this area, and to prevent the moisture and consistent motion that occurs postoperatively. 3 One must also consider the proximity of the groin to the perineal area and the rich lymphatic supply of the region. Lymphatic bacterial spread from a remote distal infection (infected ulcer or toe gangrene) was not the cause of infection in this study or in others, s' 17 Only four patients out of 37 with a distal ulcer developed an inguinal w o u n d infection. We could not establish a relationship with transfemoral arteriography even though Landreneau et al. have described a raised incidence of postoperative w o u n d complications in the groin in relation to this procedure.16 Although Bunt found no correlation between preoperative angiography and the development of either arterial or nodal positives cultures at the inguinal incision, he agreed with the thesis that any disruption of local tissue mechanics might be expected to result in impaired w o u n d healing and possibly an increased infection rate. 17 Other risk factors were: female sex, possibly related with the close anatomical relationship between the groin, anus and urethra; and age (but only in males). In this study no link was found between urinary tract infection and w o u n d infection. In the literature a higher incidence of w o u n d and graft infection was seen, when synthetic grafts were used in comparison with vein grafts,l,2,18 this was not confirmed by our results. In agreement with earlier studies, 1'2'16 obesity and diabetes were found to have no correlation with infection in this study and hypertension and smoking were similarly unconnected with infection. In this series 603 patients who underwent arterial surgery, an acceptable number of w o u n d complications was found. The main localisation of complications was the groin, and unexpectedly, none of these complications have led to graft infection. Prophylactic antibiotics, preoperative skin preparation and a good surgical technique may have contributed to these results. As for the graft infections, we must emphasise that a longer follow-up period is needed before we can claim a zero graft infection rate as some infections particularly those from aortic grafts can present many years after the operation.

Eur J Vasc Surg Vol 6, September1992

498

F . J . G . van Himbeeck et aL

References •1 SZILAGYIDE, SMITHRF, ELLIOTTJP, VRANDEClCMP. Infection in arterial reconstruction with synthetic grafts. Ann Surg 1972; 176: 321-333. 2 JOHNSONJA, COGBILLTH, STRUTTPJ, GRUNDERSENAL. Wound complication after infrainguinal bypass: classification predisposing factors, and management. Arch Surg 1988; 123: 859862. 3 EDWARDSWJ JR, MARTIN RS, JENKINS JM, EDWARDSWH SR, MULHERINJU Primary graft infections. J Vasc Surg 1978; 6: 235-

239. 4 BUNTTJ. Synthetic vascular graft infections. I. Graft infections. Surgery 1983; 93: 733-746. 5 LORENTZENJE, NIELSENON, ARENDRUPH, et al. Vascular graft infection: an analysis of sixty-two graft infections in 2411 consecutively planted synthetic vascular grafts. Surgery 1985; 98: 81-86. 6 GOLDSTONE J, MOORE WS. Infection in vascular prostheses: clinical manifestations and surgical management. Am J Surg 1974; 128: 225-233. 7 JAMIESONGG, DEWEESE JA, ROB CG. Infected arterial grafts. Ann Surg 1975; 181: 850-852• 8 C.D.C. GUIDELINES FOR PREVENTION OF SURGICALWOUND INTECTIONS. Infect Cont 1986; 7: 193-200. 9 CRUSE PJE, FOORD Rd~. A 10-year prospective study of 62,939 wounds. Surg Clin N Am 1980; 60: 27-40.

Eur J Vasc Surg Vol 6, September 1992

10 HOLM J. Wound and graft infection. Clinical aspects and prophylaxis. Acta Chirurg Scand 1985; 529 (suppl): 87-89. 11 HOFFERTPW, GENSLERS, HAIMOVICIH. Infection complicating arterial grafts. Arch Surg 1965; 90: 427-435. 12 WILLWERTH BM, WALDHAUSENJA. Infection of arterial prosthesis. Surg Gynecol Obstet 1974; 139: 446-452. 13 SEROPIAN R, REYNOLDS BM. Wound infections after preoperative depilatory versus razor preparation. Am J Surg 1971; 121: 251-254. 14 KAISERAB, CLAYSONKR, MULHERINJL, et al. Antibiotic prophylaxis in vascular surgery. Ann Surg 1987; 188: 283-289. 15 LIEKWEG WG, GREENFIELDLJ. Vascular prosthetic infections: collected experience and results of treatment. Surgery 1977; 81: 335-342. 16 LANDRENEAUMD, RAJu S. Infections after elective bypass surgery for lower limb ischemia: the influence of preoperative transcutaneous arteriography. Surgery 1981; 90: 956-61. 17 BUNT TJ. Sources of Staphylococcus epidermidis at the inguinal incision during peropheral revascularization. Am Surg 1986; 52: 472-473. 18 BOUHOUSTOSJ, CHAVATZASD, MARTIN P, MORRIS T. Infected synthetic arterial grafts. Br J Surg 1974; 61: 108-111. 19 KIKTAMJ, GOODSONSF, BISHARARA, et al. Mortality and limb loss with infected infrainguinal bypass grafts. J Vasc Surg 1987; 5: 566-571.

Accepted 4 March 1992