Antibiotic Prophylaxis in Vascular Disease Management

Skin Flora

The normal skin flora is the most important source of bacteria. Accordingly, preoperative skin preparation influences subsequent infection rates. Kaiser and coworkers noted a higher rate of infection with a hexachlorophene-ethanol preparation compared with povidone-iodine. Close and colleagues reported that hexachlorophene is more effective alone than when used in combination with ethanol. In a prospective study, Cruse demonstrated that preoperative hexachlorophene showering can be effective in reducing wound infection rates, and overzealous shaving may actually increase the risk of infection. Wooster and colleagues demonstrated that vascular grafts routinely become contaminated with skin organisms intraoperatively and suggested that careful attention to aseptic technique can significantly reduce this occurrence.

Groin incisions appear to have special significance in the development of vascular graft infections. Grafts involving an inguinal wound have a higher incidence of infection than do those that avoid this region. Jamieson and colleagues reported that the presence of a groin incision increased the risk of graft infection 3.5-fold; the presence of a groin complication such as a seroma or hematoma increased the risk of infection ninefold over patients without groin complications. Up to 33% of groin incisions with hematomas may develop infections. Lorentzen and associates reported that the highest incidence of infection was in patients who underwent aortobifemoral grafting for abdominal aortic aneurysms (5.9%), whereas there were no infections in 425 patients who underwent aortoiliac bypass for aneurysms (213) and occlusive atherosclerosis (212).

Gastrointestinal Flora

The gastrointestinal tract is a potential source of contamination during aortic reconstruction. Cultures of intestinal bag fluid have been reported by some investigators to yield enteric bacteria and skin organisms such as coagulase-negative staphylococci. In a report of 109 bowel bag cultures from abdominal aortic reconstructions, Scobie and colleagues found positive cultures in 14% of patients. S. epidermidis was the single most common organism isolated ( n = 11), whereas enteric flora were cultured in 12.

The impact of concomitant gastrointestinal surgery in the development of vascular graft infection is unclear. In separate series, DeBakey and colleagues, Stoll, and Hardy and colleagues reported a total of 670 patients who underwent aortic graft placement and simultaneous gastrointestinal procedures, with no episodes of graft infection. These authors concluded that such coincident procedures can be undertaken safely. Other investigators, however, described the development of graft infection in patients undergoing simultaneous appendectomy, cholecystectomy and gastrostomy, and anterior resection.

Arterial Colonization

The native arterial tree may harbor bacteria. The presence of pathogenic bacteria, particularly coagulase-negative staphylococci, in vascular tissues not previously operated on has been widely documented ( Table 11.3 ). Lalka and colleagues postulated that transient bacteremias resulting from breaks in the skin or mucous membranes may lead to arterial colonization. Bacterial contamination of vascular prostheses may therefore be inevitable in some cases. It is not yet clear, however, to what extent the presence of positive arterial wall cultures influences the likelihood of subsequent graft infection.

The 1977 report by Ernst and associates of abdominal aortic aneurysmal wall cultures was one of the first to highlight the presence of pathogenic organisms in the native aorta. The overall incidence of positive cultures was 15%, and cultures were more likely to be positive when atherosclerotic disease was more advanced. Asymptomatic aneurysms were less likely to be culture positive (9%) than were symptomatic (13%) or ruptured aneurysms (35%). S. epidermidis was the most frequently isolated organism. The late graft sepsis rate was 10% in the culture-positive group versus 2% in the culture-negative group. In a similar report, Buckels and coauthors described an 8% (22 of 275) incidence of positive cultures from aortic aneurysm contents. The incidence of graft sepsis was 32% (7 of 22) in patients with positive cultures, compared with 2.4% (6 of 253) in the culture-negative group.

Similar data suggest that lower extremity arteries can also become infected. In 1984, Macbeth and colleagues reported on cultures of arterial wall specimens from 88 clean, elective, lower extremity revascularization procedures. Control cultures were taken from adjacent adipose or lymphatic tissue. Although all control cultures were negative, arterial wall cultures were positive in 43% of cases (38 of 88). Of these, 71% (27 of 38) grew S. epidermidis . The authors described three graft infections in 335 cases (0.9% infection rate), all of which had positive arterial wall cultures. Also included in this report was a retrospective review of 22 cases of graft infection for which arterial and graft culture data were available. Of the patients with positive arterial cultures, 57% (8 of 14) had suture line disruption, whereas there were no disruptions in the culture-negative group. Durham and colleagues reported a series of 102 patients undergoing vascular reconstruction with a 74% (75 of 102) incidence of positive arterial wall cultures. S. epidermidis accounted for 56% of the cultured organisms. Six infections (3.5%) occurred over 18 months; all these patients had prior positive arterial cultures. No patients with negative arterial cultures developed graft infection. The greatest risk for graft infection appeared to be in patients with positive arterial wall cultures undergoing reoperation.

Hematogenous and Lymphatogenous Seeding

Hematogenous seeding of vascular prostheses is another potential source of graft infection. Anecdotal reports implicate urinary tract infection, abdominal sepsis, and other infections in the development of vascular graft infections. Laboratory models demonstrate that bacteremia reliably produces prosthetic graft infections.

Other Local and Systemic Factors

Open wounds on the distal lower extremities can be a source of contaminating bacteria. Hoffert and colleagues noted that 75% of patients with graft infections (9 of 12) had open, infected lesions on the distal lower extremity at the time of graft implantation. Liekweg and Greenfield reported that 33% of inguinal infections (20 of 60) occurred proximal to open foot infections. Bunt and Mohr described the presence of bacteria cultured from a distally infected extremity in the inguinal lymph nodes of two patients undergoing lower extremity revascularization; both patients developed graft infection.

Prior vascular surgery has been implicated as a risk factor for vascular graft infection. Dense scar tissue, increased bleeding, and lymphatic leak may all contribute to this phenomenon. Goldstone and Moore noted that 45% of patients (12 of 27) with graft infections had undergone one or more revisions of the original graft before developing an infection in the same region. In 8 of the 12 patients, the infection was in the groin. In the series by Edwards and coworkers, 9 of 18 patients (50%) had undergone a previous vascular surgery at the site of the graft infection. Similarly, a report from Reilly and colleagues described a history of multiple previous vascular procedures at the site of graft infection in 40% of cases. Johnson and associates found that prior vascular procedures were not a significant risk factor for graft infection; however, only 12 of 135 patients in this series had prior operations at the site of infection.

The immunologic status of patients with vascular disease may also have an impact on the development of graft infection. Systemic disease, malnutrition, and medical debility may suppress the host response to invading microorganisms. Kwaan and colleagues reported on 12 patients with advanced, fulminating graft infections, all of whom had critical deficiencies in immune status as determined by serum albumin, hemoglobin, immunoglobulin, and lymphocyte assays and by response to standard skin test antigens. Eight of 12 patients who received total parenteral nutrition had significant enhancement of immune response and accelerated recovery from the graft infection. Of the four patients who did not receive nutritional support, two had a prolonged convalescence, and two subsequently died from complications of graft infection.

Early Experience

Until the mid-1970s, the use of antibiotics in vascular reconstruction with synthetic materials was largely based on personal preference. It is notable that in the series of Szilagyi and colleagues, the graft infection rate among 2145 cases in which prophylactic antibiotics were not administered was 1.5%. Fry and Lindenauer reported an incidence of 1.34% in 890 cases in which no antibiotics were used. These infection rates were comparable with, and often lower than, those reported in series in which prophylactic antibiotics were used. Noting the preponderance of S. aureus in vascular graft infections, particularly in cases involving an inguinal incision, Szilagyi and colleagues suggested a clinical trial of an antibiotic directed at this organism in reconstructions that required an inguinal anastomosis.

In 1974, Goldstone and Moore published a review of the San Francisco Veterans Administration Hospital experience with vascular prosthetic infection. This series of 566 aortofemoropopliteal reconstructions was divided into two time periods: 1959 to 1965, when antibiotics were administered only postoperatively; and 1966 to 1973, when prophylaxis included preoperative, intraoperative, and postoperative antibiotics. The incidence of graft infection in the former group was 4.1% (9 of 222), compared with 1.5% (5 of 344) in the latter. Although the investigators conceded that greater experience and skill might have contributed to the lower incidence of infection, they maintained that the major factor responsible was the more appropriate use of antibiotics in the second group of patients. The following year, Perdue published a similar retrospective review that suggested that the institution of routine antibiotic prophylaxis reduced the incidence of wound infections and other nosocomial infections in patients undergoing major arterial reconstructive procedures.

Prospective Trials

The first large, prospective, randomized, blinded clinical study of antibiotic prophylaxis in vascular reconstructive surgery was published by Kaiser and colleagues in 1978. In that series, 462 patients undergoing aortofemoropopliteal reconstruction were randomized to receive either 1 g of cefazolin or a saline placebo. There were no graft infections among 225 patients who received cefazolin, compared with 4 of 237 placebo recipients (1.7%). When superficial skin infections and subcutaneous skin infections were considered in the analysis (Szilagyi classes I and II), the overall infection rates were 0.9% in the cefazolin group and 6.8% in the placebo group. Given no adverse drug reactions and no noted cefazolin resistance, the authors strongly recommended a short course of cefazolin prophylaxis in patients undergoing arterial reconstructive surgery.

The benefit of a short course of systemic cephalosporin prophylaxis in vascular reconstructive surgery was subsequently confirmed in a number of other prospective, randomized trials. In 1983, Salzmann reported a trial of cefuroxime (a second-generation agent) and later cefotaxime (a third-generation agent) versus placebo in 300 patients undergoing aortofemoropopliteal reconstruction. The prophylaxis regimen was changed from cefuroxime to cefotaxime midway through the study because the latter was found to be more effective in vitro against the most common graft infection pathogens at the author's institution. Graft infection rates were 2.4% for the placebo group and 0.8% for the prophylaxis group. The incidence of wound infection was 15.1% in the placebo group and 3.0% in the prophylaxis group. No differences in infection rate were noted between the two antibiotics, and the author concluded that either agent could be used effectively in the prophylaxis of postoperative infection.

Addressing the question of duration of treatment for antibiotic prophylaxis, Hasselgren and colleagues compared 1- and 3-day courses of cefuroxime versus placebo in lower extremity arterial reconstruction. There was only one graft infection in this small cohort of 110 patients, and it occurred in the placebo group. The wound infection rate was 16.7% for patients receiving placebo, compared with 3.8% in the 1-day and 4.3% in the 3-day prophylaxis groups. The investigators recommended that prophylactic antibiotic therapy be limited to a short-term course.

Robbs and associates reported a trial of cloxacillin plus gentamicin versus cefotaxime in infrainguinal arterial reconstruction. This group had adopted a 48-hour course of cloxacillin plus gentamicin as their routine prophylaxis as a result of the predominance of S. aureus and gram-negative infections at their institution. Length of follow-up ranged from 6 to 20 months. The wound infection and graft infection rates for patients receiving cloxacillin plus gentamicin were 5.4% (7 of 129 wounds) and 1.5% (1 of 63 grafts), respectively. The rates for patients receiving cefotaxime were 6.3% (8 of 127 wounds) and 3.3% (2 of 61 grafts). The differences were not statistically significant. The authors concluded that the multiagent 2-day regimen conferred no advantage over the shorter, single-agent regimen.

Local Therapies

In 1980, Pitt and colleagues reported the results of a controlled study of cephradine prophylaxis in vascular procedures involving groin incisions in which topical, systemic, and topical plus systemic administration were compared. Of 205 patients, 52 had prosthetic grafts placed, whereas the remainder received vein grafts. Infection rates were equivalent in these two groups. Wound infection rates were 0% for those receiving topical administration alone and systemic administration alone, 5.9% for patients receiving both, and 24.5% for controls. No distinction was made between graft (Szilagyi class III) and isolated wound (Szilagyi classes I and II) infections. Minimum follow-up was 4 weeks, but the mean length of follow-up was not indicated. Patients in whom synthetic graft material was used did not experience a higher incidence of wound infection. The authors concluded that topical and systemic prophylaxes were equally efficacious and that combined prophylaxis was unnecessary. The follow-up interval in this study, however, was not long enough to make conclusive statements.

Mohammed and colleagues compared the effect of prophylactic intraoperative wound irrigation with vancomycin (along with systemic antibiotics) to use of systemic antibiotics alone in a retrospective analysis of 454 patients undergoing aortofemoral or infrainguinal vascular surgery. They noted that use of vancomycin irrigation resulted in reduced incidence of superficial infections, but did not reduce inguinal wound dehiscence or deep infections.

Use of a collagen implant impregnated with gentamicin sulfate for prophylaxis was studied retrospectively in 60 nondiabetic and nonobese patients undergoing infrainguinal prosthetic bypass procedures by Costa Almeida and colleagues. The results were promising in that the incidence of surgical site infection and the number of in-hospital days were reduced by the application of this implant when compared to the control group (0% vs. 20% for surgical site infection and mean of 5.66 days vs. 8.10 days for in-hospital days). In another review of current literature, the potential role of a gentamicin-containing collagen implant for prophylaxis was discussed. This approach may have a beneficial effect in prevention of infection in patients who are at a high risk for infection or who are undergoing high-risk procedures. Additional therapies described include the use of antibiotic-impregnated polymethyl methacrylate (PMMA; bone cement) beads with daptomycin, with or without the addition of tobramycin powder.

Pretreatment of vascular grafts with antibiotics and fibrin sealant has also been investigated in in vitro experiments. Use of antibiotics with fibrin allows delayed release of antibiotic and a potentially longer period of resistance to infection. In animal models using a vascular prosthesis, daptomycin has been seen to prevent staphylococcal biofilm formation. In the presence of rifampicin, the minimum inhibitory concentration (MIC) and the minimum bactericidal concentration of daptomycin are lower. Some authors have cautioned against relying solely on systemic antibiotic prophylaxis for vascular grafts because of emergence of resistant microbial strains. Consideration for antimicrobial silver grafts has been highlighted in this regard.

Hemodialysis Access

For hemodialysis access, the risk of infectious complications increases from arteriovenous fistulae (AVF) to arteriovenous grafts (AVGs) and central dialysis catheters. In patients with AVGs, most infections arise from inoculation due to skin flora or seeding from distant sites. Infections due to Staphylococcus have been noted to occur more commonly in carriers than noncarriers. With regard to reduction of AVG infections, two prophylactic measures may have some merit for consideration including reduction of nasal carriage of the pathogen in carriers and use of cryopreserved human femoral vein as an alternative to prosthetic graft.

Bennion and colleagues examined the utility of antibiotic prophylaxis in patients with chronic renal insufficiency undergoing placement of a prosthetic arteriovenous shunt for hemodialysis. Patients were randomized to receive cefamandole or placebo just before placement of a polytetrafluoroethylene (PTFE) graft, followed by two subsequent doses. The wound infection rate for the cefamandole group was 10.5% (2 of 19), with one graft (Szilagyi class III) infection. The wound infection rate in the placebo group was 42.1% (8 of 19), with three graft infections. This high rate of infection is not uncommon in renal failure patients, and the study emphasized the importance of perioperative antibiotic prophylaxis.

A significant number of patients undergo minimally invasive percutaneous interventions aimed at maintenance of their dialysis access. Salman and colleagues recommended antibiotic prophylaxis in two particular situations in such patients: (1) peritoneal dialysis (PD) catheter placement and (2) insertion of an accidentally extruded tunneled hemodialysis catheter (TDC). In their experience, patients who developed clinical infection after percutaneous interventions had other comorbidities such as diabetes or advanced HIV. For prophylaxis, they used cefazolin for extruded TDC patients and vancomycin for PD catheter insertion. Overall, the rates of infection were low with no patients developing infection after PD catheter placement, one after angioplasty (0.04%), and one after tunneled catheter placement (0.3%).

Tunneled central venous catheters can contribute to significant morbidity and mortality in patients with end-stage renal disease (ESRD) due to catheter-related infections. The question of consideration of prophylactic antibiotics in these patients is therefore an important one. In a prospective study, 60 patients were randomized between two groups. The group that received prophylactic antibiotics (1 gram cefazolin) prior to catheter insertion had a lower incidence of catheter loss due to infection ( P < .05), catheter exit-site infections ( P < .05), and tunnel infections ( P < .05). In oncology patients, a Cochrane review in 2013 did not find any benefit to the administration of routine prophylactic antibiotics to prevent Gram-positive catheter-related infections for long-term central venous catheters.

Major Limb Amputation

McIntosh and colleagues conducted a review of literature to study the role of antibiotic prophylaxis in patients undergoing major limb amputations. They concluded that use of prophylactic antibiotics reduced rates of stump infection and even reamputation. The agent used for prophylaxis did not appear to affect outcomes. Sadat and colleagues evaluated outcomes after major lower limb amputation in relation to the duration of antibiotic prophylaxis received. Their results showed lower wound infection, hospital length of stay, and revision rate for the amputation in patients who received 5 days of antibiotic prophylaxis.

Arterial Closure Devices

Very low infection rates have been reported in the literature for patients undergoing angiography and angioplasty followed by deployment of arterial closure devices. A meta-analysis reviewed the incidence of infection in patients undergoing endovascular aortic repair involving use of vascular closure devices at access sites. Jaffan and colleagues reported a meta-analysis of closure device–related infections. They found no statistical difference in the rate of infection between patients who received antibiotic prophylaxis and those who did not. Based on these findings, the use of prophylactic antibiotics is not recommended. Conditions that increase the risk of infectious complications after arterial closure device use include obesity, diabetes, and use of a closure device within the last 6 months.

Angiography, Inferior Vena Cava Filter

In general, aseptic technique appears to be more important than antibiotic prophylaxis in angiographic procedures. Factors that increase the risk of infection in such procedures include repeated puncture or repeated catheterization of a sheath that is already in place or repeated intervention within 7 days if an endovascular stent is involved. Routine use of antibiotic prophylaxis before placement of an inferior vena cava (IVC) filter is currently not recommended. A fresh venous access site instead of a pre-existing conduit such as a central venous catheter is, however, recommended for the placement of IVC filters.

Lower Extremity Superficial Venous Insufficiency Treatment

There is currently insufficient evidence in the literature to support the routine use of prophylactic antibiotics in procedures for lower extremity superficial venous insufficiency treatment such as varicose vein phlebectomy, endovascular thermal ablation, and sclerotherapy. Again, more emphasis is placed on use of aseptic technique.

Autologous Vein Graft Reconstruction of Lower Extremity

A recent retrospective study from the Netherlands specifically included patients undergoing lower extremity arterial bypass using vein grafts between 2004 and 2012. One group received single-dose antibiotic prophylaxis, whereas the second group did not receive prophylaxis. The incidence of surgical site infections between the two groups was greater than 20% and not statistically significant.

Comparisons of Antibiotic Regimens

Because it has become evident that a short course of a cephalosporin antibiotic is the ideal prophylaxis for vascular reconstructive procedures, several studies have focused on whether the most widely used cephalosporin, cefazolin, is the best choice. A large number of graft infections, particularly in abdominal grafts, are due to gram-negative rods. A theoretical disadvantage of first-generation cephalosporins such as cefazolin is that they are more vulnerable to gram-negative beta-lactamase than are second- and third-generation agents. Gram-negative activity is thus limited to E. coli, Proteus species, and Klebsiella species, and many hospital-acquired strains of these organisms are cefazolin resistant. It has also been demonstrated that other cephalosporins, such as the second-generation agent cefamandole, have greater in vitro activity against coagulase-negative staphylococci, which have been found to colonize the native arterial wall in a large number of patients. It is clear from previous studies by Salzmann, Hasselgren and colleagues, and Robbs and colleagues that second- and third-generation cephalosporins can be used effectively in vascular surgery prophylaxis.

In 1989, Lalka and colleagues examined this issue in a prospective study of arterial wall microbiology and antibiotic penetration. Forty-seven patients undergoing aortofemoropopliteal reconstruction were randomized to receive perioperative cefazolin or cefamandole, 1 g every 6 hours for nine doses. Serial samples of serum, subcutaneous fat, thrombus, atheroma, and arterial wall were obtained for culture and assay of drug levels by high-performance liquid chromatography. Serum and tissue levels of cefazolin were significantly higher than those of cefamandole at almost all time points. Positive arterial wall cultures were obtained in 41.4% of patients, and 68.8% of bacterial isolates were coagulase-negative staphylococci (half of these were slime producers). At times, the arterial wall concentration of both antibiotics fell below the geometric mean minimal inhibitory concentration for all organisms combined, but this occurred significantly more often with cefamandole. The investigators concluded that both antibiotics needed to be administered in larger doses (cefazolin, 1.5 g every 4 hours; cefamandole, 2 g every 3 hours) and that the antibiotics were essentially equal in efficacy if administered appropriately. This study corroborated the findings of Mutch and colleagues, who noted that serum antibiotic levels did not correlate well with aortic tissue concentrations of bioactive antibiotic, and it suggested that arterial tissue levels rather than serum levels should be the standard for comparison of antibiotic efficacy.

Edwards and colleagues reported a prospective trial of cefazolin versus the more β-lactamase–stable second-generation cephalosporin cefuroxime in patients undergoing aortic and peripheral vascular reconstruction. Prior studies had suggested that some failures of cefazolin prophylaxis were due to this agent's susceptibility to staphylococcal β-lactamase and that other cephalosporins might provide better protection in cardiac surgery. Antibiotics were administered just before surgery, redosed intraoperatively, and continued every 6 hours postoperatively for 24 hours. Dosage and administration schedules were based on a prior pharmacokinetic study. The infection rate in the cefazolin group was 1% (3 of 287), versus 2.6% (7 of 272) in the cefuroxime group. This difference was not statistically significant. Cefuroxime exhibited lower trough concentrations than did cefazolin, and the length of the operative procedure was found to be a risk factor for infection only in the cefuroxime group. The investigators concluded that despite its lower resistance to β-lactamase, cefazolin provides better perioperative prophylaxis because of its greater antistaphylococcal potency and superior pharmacokinetic profile. Data from this and other studies suggest that intraoperative redosing of cefazolin should be more frequent in prolonged procedures than in routine therapeutic administration—that is, every 4 hours rather than every 6 hours.

A study done in Sweden evaluated an antibiotic regimen for prophylaxis in patients undergoing vascular surgery involving groin incisions ( n = 219). One group received cloxacillin for antimicrobial prophylaxis, whereas the second group received trimethoprim/sulfamethoxazole (TMP-SMX). This study failed to show a difference in the incidence of surgical site infections between the two groups.