Does the timing of reoperation influence the risk of graft infection?
Article Outline
Objective
This study compared the incidence and characteristics of graft infection in patients who underwent early vs late revisional surgery of lower extremity arterial bypass grafts.
Methods
Between 1992 and July 2005, 500 revisional procedures were performed on 198 lower extremity bypass grafts. Patients whose revisions were performed <30 days after the primary bypass were in the early revision (ER) group (n = 99), and those done >30 days after bypass were in the late revision (LR) group (n = 99). Infection was defined as cellulitis with graft exposure or purulence in continuity with a graft that required antibiotics and operation for infection control. Mean follow-up was 60 months (range, 2 to 60 months). Groups were compared using Student’s t test.
Results
The ER group included 66 autogenous and 33 prosthetic grafts. The LR group consisted of 53 autogenous and 46 prosthetic grafts. Of the 500 revisional procedures performed, 17 graft infections occurred (3.4%). Twelve (70.6%) were prosthetic grafts and five (29.4%) were autogenous grafts (P = .004). Defining the infection rate per graft rather than per revisional procedure, the ER group had a significantly higher graft infection rate at 11% (11/99) compared with 6.1% in the LR group (6/99; P = .012). The risk of infection for prosthetic grafts was significantly higher within the ER group at 27.3% (9/33) compared with autogenous grafts at 3.1% (2/66; P = .0001). Infection developed in three vein grafts and three prosthetic grafts in the LR group (P = NS). For prosthetic graft revisions only, infection risk was 27.3% (9/33) in the ER group and 6.5% (3/46) in the LR group (P = .005). The most common cultured pathogen was methicillin resistant Staphylococcus aureus (ER, 6/11 vs LR, 3/6; P = NS. Within the ER group, the prevalence of Pseudomonas aeruginosa was significantly higher at 27.3% (3/11) compared with 0% (0/6) in the LR group (P = .04).
Conclusions
Early revision of lower extremity arterial bypass grafts has a significantly higher risk of graft infection compared with revision >1 month after surgery. Infection will develop in approximately 25% (9/33) of prosthetic grafts that are reoperated on early. If feasible, reoperation should be delayed >1 month for prosthetic grafts needing revision. Endovascular or extra-anatomic interventions should be considered if early revision is mandated in this group.
Infrainguinal bypass is a well-established therapeutic option in patients with disabling claudication or threatened limb loss. Graft failure remains a significant problem, however, with 5% to 20% of grafts failing in the first 30 days, and 20% to 50% failing >30 days.1 Postoperative duplex graft surveillance is critical to identify a failing graft, and reintervention of failing grafts, or those that have failed, allows long-term salvage of many of these grafts.
Graft infection is a major complication associated with reoperative surgery on lower extremity bypasses.2, 3 These are uncommon but can be a devastating complication with high amputation and mortality rates.4, 5, 6 When a graft occludes or a surveillance study identifies a graft-threatening lesion, the timing of surgery, risk of graft failure, severity of ischemia, and risk of graft infection secondary to revision surgery must be weighed. Although reoperation is a well-established risk for infection,7, 8, 9 the magnitude of the risk and the influence on timing of reoperative surgery has not been established.
We have analyzed the type of bacteria associated with early and late graft infections,2, 3 and this study compared the incidence and characteristics of graft infection after revisional surgery of lower extremity arterial bypass grafts. We questioned whether there is a difference in the incidence of graft infection for early vs later graft revision. In addition, we were interested in whether the type of conduit used (autogenous or prosthetic) affected the incidence of graft infection.10, 11
Methods
We identified patients who underwent revisional procedures on lower extremity bypass grafts between 1992 and 2005 by the vascular surgery service at Pennsylvania Hospital in Philadelphia. Patients were classified as early revision (ER) if the revisional procedures was performed <30 days from the original operation and late revision (LR) if it was performed >30 days after the original operation.
All procedures were performed in an endovascular operating suite. The key outcome analyzed was graft infection, which was defined as cellulitis with graft exposure or purulence in continuity with a graft requiring parenteral antibiotic therapy and surgical operation for infection control. Patient demographics, clinical factors, and outcome were recorded from office and hospital charts.
Clinical factors and outcome measures included vascular conduit (autogenous or prosthetic), indications for primary and revisional procedures, preoperative work-up, infection site, clinical infection criteria, microbiology, limb loss, and mortality. Indications for initial operation included ulcer or gangrene, rest pain, or claudication. Indications for revisional procedures identified were thrombosed graft, failing graft, anastomotic bleeding, and hematoma evacuation. Graft infections were evaluated with intravenous contrast computed tomography (CT) scans, and arteriography was used liberally.
Data were analyzed in accordance with suggested standards on reporting of lower extremity ischemia procedures.12 Mean follow-up time for the cohort was 60 months. Groups were compared using Student’s t test and analysis of variance valuation. P < .05 was defined as statistically significant.
Results
Between 1992 and 2005, 500 surgical revisions were performed on 198 lower extremity bypass grafts, of which 99 had a revision <30 days from the primary graft construction (ER), and 99 were identified in the LR group (revision >30 days). The ER group included 66 autogenous and 33 prosthetic grafts, and the LR group consisted of 53 autogenous and 46 prosthetic grafts (P = .059).
Preoperative demographics and risk factors are summarized in Table I. Men comprised about 50% of the population, and 75% of patients were white. There were no significant differences in age, gender, ethnicity, the presence of hypertension, diabetes, hypercholesterolemia, and chronic obstructive pulmonary disease, and primary surgical indication between the ER and LR groups or infected grafts and those that did not develop infection.
Table 1. Patient demographics
| Total population | Early revision | Late revision | P | |
|---|---|---|---|---|
| Age | 69.33 ±0.79 | 68.67±1.14 | 69.58±1.15 | .289 |
| Sex (%) male | 51.90±0.03 | 53.53±0.05 | 49.49±0.05 | .285 |
| Race (%) white | 75.00±0.02 | 76.76±0.04 | 75.51±0.04 | .419 |
| Hypertension | 77.14±0.02 | 76.76±0.04 | 77.77±0.04 | .433 |
| Diabetes mellitus | 40.00±0.03 | 46.46±0.05 | 36.36±0.04 | .075 |
| Hypercholesterolemia | 37.80±0.03 | 36.36±0.05 | 36.84±0.05 | .945 |
| COPD | 16.12±0.02 | 16.16±0.03 | 16.16±0.37 | .9999 |
Graft infections occurred in 17 (3.4%) of the 500 revisional procedures performed, of which 12 (70.6%) were in prosthetic grafts and five (29.4%) were in autogenous grafts (P = .004). No deaths occurred in the 17 patients who developed infection.
Defining the infection rate over the lifetime of the grafts that required revision, rather than per revisional procedure, the overall graft infection rate was 8.6% (17/198). The ER group had a significantly higher graft infection rate at 11.1% (11/99) compared with 6.1% ((6/99) in the LR group (P = .012). Nine of the 11 graft infections that occurred within the ER group were in prosthetic grafts, and two were in autogenous grafts (9/33, 27.3% vs 2/66, 3.1%; P = .0001). Six graft infections occurred in the LR group: three each in prosthetic grafts and in vein grafts (P = NS).
Considering prosthetic graft revisions only, infection risk was 27.3% (9/33) in the ER group and 6.5% (3/46) in the LR group (P = .005). For autogenous grafts, infection risk was 3.1% (2/64) vs 6.0% (3/50) for early vs late revisions (P = NS).
The amputation rate among patients who developed graft infection was 29.4% (5/17). All amputations that were required occurred shortly after the onset of infection. There was no statistical difference in limb salvage between the ER (72.7%) and LR groups (66.6%).
The operative infection site and surgical indications are listed in Table II. The groin incision was more likely than other sites to be involved with infection, with 80% of graft infections involving the groin. The most common primary indication for surgery for grafts that ultimately became infected was ulcer or gangrene (8/17). The most common indication for revision among all infected grafts was a thrombosed graft (9/17).
Table II. Operative sites and indications
| Early revision n = 11 (%) | Late revision n = 6 (%) | |
|---|---|---|
| Infection site | ||
| Groin | 10(90.9) | 5(83.3) |
| Other | 1(9.1) | 1(16.7) |
| Primary operative indication | ||
| Ulcer/gangrene | 4(36.4) | 4(66.6) |
| Rest pain | 5(45.5) | 1(16.6) |
| Ulcer/rest pain | 2(18.1) | 1(16.6) |
| Reoperative indication | ||
| Thrombosed graft | 7(54.5) | 2(33.3) |
| Failing graft | 0 | 4(66.6) |
| Anastomotic bleed | 2(18.8) | 0 |
| Hematoma evacuation | 2(18.8) | 0 |
| Clinical infection criteria | ||
| Exposed graft | 3(27.3) | 2(33.3) |
| Purulent drainage | 5(45.4) | 3(50.0) |
| Disrupted anastomosis | 3(27.3) | 1(16.7) |
We further analyzed the number of revisions performed on the graft after the initial surgery and before the diagnosis of infection. Within the ER group, 54.5% (6/11) of the infections occurred after only one intervention, and 45.5% (5/11) presented after multiple revisional operations. In the LR group, 83.3% (5/6) of the infections occurred after only one operative procedure. These numbers were compared with the rest of the cohort, of which 120 of 181 grafts underwent only one intervention and 61 of 181 underwent two or four interventions during the follow-up period. The average interval between initial revision and subsequent infection was 183.2 days (range, 37 to 592 days) in the ER group and 301.4 days (range, 29 to 828 days) days in the LR group (P = .112).
Infection in the ER group presented as three exposed grafts with drainage, five perigraft abscesses or draining sinuses, and three disrupted anastomoses. In the LR group, there were two exposed grafts, three purulent drainage sites, and one disrupted anastomosis.
The most common cultured pathogen was methicillin resistant Staphylococcus aureus (MRSA) (ER, 6/11 vs LR, 3/6; P = NS). Other bacteria identified included methicillin sensitive S aureus, Pseudomonas aeruginosa, Alcaligenes xylosoxidans, Serratia marcescens, and vancomycin-resistant Enterococcus (Table III). The prevalence of grafts infected with P aeruginosa was significantly higher in the ER group at 27.3% (3/11) compared with the LR group at 0% (0/6; P = .04).
Table III. Microbiology
| Organism | Early revision (n) | Late revision (n) |
|---|---|---|
| MR Staphylococcus aureus | 6 | 3 |
| MS Staphylococcus aureus | 3 | 1 |
| Pseudomonas aeruginosa | 3 | 0 |
| Alcaligenes xylosoxidans | 1 | 1 |
| Serratia marcescens | 1 | 0 |
| VR Enterococcus | 0 | 1 |
| Polymicrobial | 1 | 0 |
Discussion
Despite advances in surgical technique, infrainguinal graft failure remains a significant problem. Early graft occlusion, defined as <30 days after surgery, is often attributed to technical problems and occurs in 5% to 20% of cases. Late graft failure (>30 days) occurs in 20% to 50% of cases.13 Although in our series the differences in the demographics and risk factors were not significant, the ER and LR groups are not strictly comparable, and confounding risk factors may influence the infection rate.
Infection is a known complication of any revisional surgery. In this series, the overall graft infection rate was 8.5%, which is somewhat high but within the 2.5% to 12% range reported by others.14, 15, 16, 17 This may reflect the fact that all grafts underwent at least one surgical revision to be included in this series. Two thirds of the graft infections occurred in ER group, and >80% of the infections involved prosthetic grafts.
Most vascular surgeons try to avoid using prosthetic grafts in the setting of foot ulcers and gangrene. We also avoid prosthetic grafts in this setting, if possible, but there are times when no suitable conduit is available, and prosthetic grafts have been used after appropriate institution of antibiotics and local measures for infection. There was no statistically significant increased risk of infection within these subgroups, but our sample size was relatively small.
As might be expected, the ER group included a significant number of acute graft thromboses that prompted emergent or urgent surgical intervention. The presence of thrombosis at the time of revision appears to increase the risk of graft infection. This may reflect the fact that thrombus, wound seroma, lymphatic interruption, and hematoma provide a rich medium for bacterial growth and are relatively poorly penetrated by parenteral antibiotics. Others have shown that the presence of a thrombosed prosthetic graft near a patent prosthetic bypass may increase the risk of infection of the functioning graft.18
When the 17 graft infections were reviewed, we found that four grafts had predisposing factors that may have led to the graft infection. Two of the grafts developed hematoma after the original operation that required reoperation, one developed a lower extremity cellulitis requiring intravenous antibiotics after a revisional procedure, and one graft developed a lymphocele requiring drainage.
All patients with infrainguinal bypass are placed in a duplex ultrasound surveillance protocol in our institution. In 62 of the cohort of 198 duplexes, evidence of graft stenosis was the indication for the primary revisional procedure. In the LR group, most of the operations were performed for duplex ultrasound evidence of a failing graft. The LR grafts tended to be well incorporated, and revision occurred in a more elective situation, which may account for the lower infection rate.
Infections that developed in the LR group occurred after a longer interval from the time of revision compared with the ER group, 301 vs 183 days, respectively. Although this did not achieve statistical significance, the trend may also reflect that well-incorporated and patent grafts are relatively resistant to infection. In the ER grafts, a lack of resistance to bacterial inoculation may lead to a more rapid clinical evidence of graft infection. Of note, the interval between revision and infection in both groups was unexpectedly long. This adds some doubt that the revision was the key factor in the development of the infection rather than the original operation or a remote issue.
Eighty percent of the grafts in the ER group that developed infection were prosthetic. Autogenous grafts can become infected but clearly are more resistant to this complication than prosthetic grafts.
Most of the graft infection in both groups occurred after a single revision, but the remainder had two to four revisions. We were not able to break down the multiple-revision groups into small subgroups to analyze for statistical significance because the sample sets are too small.
Regardless of the timing of revision, MRSA was the most common infecting pathogen, with a higher prevalence in the ER group. MRSA has emerged as the leading cause of postoperative infection in vascular surgery,19, 20 and MRSA graft infection is associated with higher morbidity and amputation rates compared with other pathogens, excepting Pseudomonas species.2, 20, 21, 22 Graft infections with MRSA tend to present earlier than infections with other organisms occurring from direct seeding from colonized skin at the time of operation.23 Unexpectedly, we had no documented cases of S epidermidis. Groin incisions were the site of 88% of all the infections. Direct contamination of grafts from groin organisms is a common cause of graft infection.8
The definition and description of early wound complications can be inconsistent, making it difficult to collect this data retrospectively. Indeed, this is why we had a fairly rigorous definition for graft infection, which was cellulitis with graft exposure or purulence in continuity with a graft that required parenteral antibiotic therapy and surgical operation for infection control and requiring operative control. One patient (1/17) had a noninfectious wound complication of a lymphocele that required drainage.
Lymphatic spread of organisms from a pedal infection site has also been implicated in early prosthetic graft infection, with disrupted inguinal nodal basins providing access.24 Cultures were not routinely done for foot ulcers, so we are unable to correlate with cultures taken at time of graft infection with the presenting foot ulcer.
All patients undergoing a revisional procedure should be treated with antibiotics appropriate for MRSA and gram-negative flora.20, 21, 25, 26 In addition to appropriate spectrum, the surgeon must ensure proper administration time (1 hour before incision) and repeated dosing to maintain effective serum and tissue levels throughout the operation. The optimal duration of antibiotic coverage remains unclear, however. We generally maintain prophylactic therapy for 48 hours for early revisional procedures on prosthetic grafts, although this is not based on any supporting data.
In our series, the highest risk for infection is early revision of a prosthetic graft. The sobering finding that more than one quarter (9/33) of prosthetic grafts that underwent early revision ultimately developed infection mandates reappraisal of our approach in these patients. Ideally, early graft exploration should be avoided or delayed if limb ischemia is tolerated. Unfortunately, delaying intervention may not be an acceptable strategy with acute graft occlusion and severe ischemia or with postoperative bleeding, which was the case in all of our early revisional procedures that developed infection.
Because 80% of graft infections occurred at a groin wound, avoidance of reopening the groin seems prudent if feasible. Extra-anatomic approaches to graft replacement should be considered and is one of our favored strategies, especially using the external iliac artery as a virgin inflow source approached retroperitoneally through a small transverse suprainguinal incision and tunneling a new graft laterally around the previous groin incision to a more distal site on the pre-existing graft.
More recently, endovascular interventions, such as mechanical thrombectomy or subintimal angioplasty of chronic occlusions, have been used in preference to reopening fresh surgical wounds. If groin wounds must be explored early, meticulous wound closure technique is important. Postoperative avoidance of skin moisture, maceration, and fungal overgrowth in the groin area are likewise critical.
Conclusions
Early revision of lower extremity arterial bypass grafts carries a significantly higher risk of graft infection compared with revision >1 month after surgery. Infection will ultimately develop in as many as one quarter of prosthetic grafts reoperated on early, with MRSA being the most common pathogen. If feasible, reoperation for prosthetic grafts that need revision should be delayed >1 month. When reoperation early is mandated, strategic planning is essential, as is appropriate antibiotic use and sterile technique. Endovascular or extra-anatomic interventions should be considered if early revision is mandated in this group, and groin incisions should be avoided if possible.
Author contributions
References
- Infrainguinal vein bypass grafts revision: factors affecting long term outcome. J Vasc Surg. 2004;40:916–923
- . Differences in outcome and management of early versus late extracavitary arterial graft infections. J Vasc Surg. 1995;22:680–685
- . Are gram- negative bacteria a contraindication to selective preservation of infected prosthetic arterial grafts?. J Vasc Surg. 1992;16:337–346
- . Mortality and limb loss with infected infrainguinal bypass grafts. J Vasc Surg. 1987;5:566–571
- . A modified classification and approach to the management of infections involving peripheral arterial prosthetic grafts. J Vasc Surg. 1988;8:147–153
- . Surgical management of infrainguinal arterial prosthetic graft infections: review of a thirty-five-year experience. J Vasc Surg. 1995;21:782–790discussion 790-1
- . The influence of groin sepsis on extraanatomic bypass patency in patients with prosthetic graft infection. Ann Vasc Surg. 1992;6:80–84
- Vascular graft infection: an analysis of sixty-two graft infections in 2411 consecutively implanted synthetic vascular grafts. Surgery. 1985;98:81–86
- . Infection in vascular prostheses (Clinical manifestations and surgical management). Am J Surg. 1974;128:225–233
- . Management of infected lower extremity autologous vein grafts by selective graft preservation. Am J Surg. 1992;164:291–294
- . Selective preservation of infected prosthetic arterial grafts: analysis of a 20-year experience with 120 infected grafts. Ann Surg. 1994;220:461–471
- Suggested standards for reports dealing with lower extremity ischemia. J Vasc Surg. 1986;4:80–94
- . Preferred strategies for secondary infrainguinal bypass: lessons learned from 300 consecutive reoperations. J Vasc Surg. 1995;21:282–293discussion 293-5
- Long-term outcome after early infrainguinal graft failure. J Vasc Surg. 1997;26:425–437
- . Local infections after above-knee prosthetic femoropopliteal bypass for intermittent claudication. Surg Infect. 2004;5:1–9
- . Arterial and prosthetic graft infection. Ann Vasc Surg. 1992;7:485–491
- . Synthetic vascular graft infections (I. Graft infections). Surgery. 1983;9:733–746
- . The thrombosed prosthetic graft is a risk for infection of an adjacent graft. Am J Surg. 1996;172:175–177
- . Methicillin-resistant Staphylococcus aureus infections in vascular surgery: increasing prevalence. Surg Infect. 2004;5:180–187
- . Recommendations for initial antibiotic treatment of peripheral arterial graft infections. Am J Surg. 1995;170:123–125
- Improved management of infrainguinal bypass graft infection with methicillin-resistant Staphylococcus aureus. Br J Surg. 1999;86:1433–1436
- . Methicillin-resistant Staphylococcus aureus infection in vascular surgical patients. Ann R Coll Surg Engl. 2001;83:158–163
- . Mechanism of late prosthetic vascular graft infection. Cardiovasc Surg. 1997;5:486–489
- . The role of the lymphatic system in acute arterial prosthetic graft infections. J Vasc Surg. 1985;2:92–98
- . Infection of vascular prostheses. Aust N Z J Surg. 1991;61:432–435
- Systemic and local antibiotic prophylaxis in the prevention of prosthetic vascular graft infection: an experimental study. Eur J Vasc Endovasc Surg. 2002;23:127–133
Competition of interest: none.CME article
PII: S0741-5214(06)01632-6
doi:10.1016/j.jvs.2006.09.007
© 2007 The Society for Vascular Surgery. Published by Elsevier Inc. All rights reserved.