Brachial Artery Ligation with Total Graft Excision is a Safe and Effective Approach to Prosthetic Arteriovenous Graft Infections

Abstract

OBJECTIVE

While autogenous arteriovenous access is preferred, prosthetic arteriovenous grafts (AVG) are still required in a large number of patients. Infection of AVGs occurs frequently and may cause life-threatening bleeding or sepsis. Multiple treatment strategies have been advocated (ranging from graft preservation to excision with complex concomitant reconstructions), indicating a lack of consensus on appropriate management of infected AVGs. We undertook this study to evaluate if, in the setting of anastomotic involvement, brachial artery ligation distal to the origin of the deep brachial artery accompanied by total graft excision (BAL) is safe and effective.

METHODS

All prosthetic arteriovenous graft infections managed by a single surgeon between 1995 and 2006 were reviewed retrospectively. Patients were identified from a computerized vascular registry and data were obtained via patient charts and the electronic medical record.

RESULTS

We identified 45 AVG infections in 43 patients. Twenty-one patients (49%) demonstrated arterial anastomotic involvement and were treated with BAL; these form the cohort for this analysis. Mean patient age was 53.2 (SD 9.5) years. The primary etiologies for ESRD were hypertension (29%), HIV (24%), and diabetes (19%). An upper arm AVG was present in 95% of patients; one (5%) had a forearm AVG. The majority of grafts were PTFE (90%). Follow-up was 100% at 1 month, 86% at 3 months, and 67% at 6 months. No ischemic or septic complications occurred in the 21 patients who underwent BAL.

CONCLUSIONS

BAL is an effective and expeditious method to deal with an infected arm AVG in frequently critically ill patients with densely scarred wounds. In the short term, BAL appears to be well tolerated without resulting ischemic complications. Further study with longer duration of follow-up is necessary to ascertain whether BAL results in definitive cure, or whether patients may ultimately manifest ischemic changes and require additional intervention.

INTRODUCTION

In the United States, 480,000 patients were estimated to have end stage renal disease (ESRD) in 2005, of whom 340,000 were managed with hemodialysis.¹ Although the National Kidney Foundation Dialysis Outcomes Quality Initiative² and the Fistula First Project³ have recommended autogenous arteriovenous access as the preferred hemodialysis method, significant numbers of ESRD patients require prosthetic arteriovenous grafts (AVG) for hemodialysis access.⁴

Complication of AVGs are frequent,⁵ and infection comprises a major source of morbidity and mortality for this patient population. Because an AVG is a foreign body that is repeatedly exposed to the external environment by needle puncture, infectious complications can occur in 3% to 35% of patients.⁶ Potential sequelae from these complications can range from minor soft tissue infections to major life-threatening hemorrhage or sepsis. Numerous treatment strategies have been developed to address infected AVGs, depending on the age of the graft, the extent of infection, the presence of anastomotic involvement, and operator preference. The techniques most commonly described are total graft excision with vein patching of the arterial anastomotic site,⁷ subtotal graft excision with oversewing of a prosthetic remnant,^{8, 9} and partial excision with segmental bypass through an uncontaminated field.^{10, 11}

Because all of these treatment strategies maintain an intact arterial suture line in a potentially infected field, they create a risk for brachial artery bleeding complications. Therefore, a method to manage an infected AVG that also definitively controls the brachial artery is needed. The purpose of the present study is to evaluate the safety and efficacy of brachial artery ligation distal to the origin of the deep brachial artery (BAL) accompanied by total graft excision in the setting of an infected AVG involving the arterial anastomotic site.

PATIENTS AND METHODS

All AVG infections managed by a single vascular surgeon in a University-associated practice between January, 1995 and January, 2006 were identified from a computerized vascular registry and were reviewed retrospectively. Data were abstracted from both patient charts and the electronic medical record. Demographics, AVG configuration and material, timing and type of operative revision, as well as follow-up data were analyzed.

Brachial arterial involvement in AVG infection was diagnosed by ultrasound visualization of fluid around the arterial anastomosis and by clinical findings at the time of surgery. A standardized procedure for BAL was followed for all patients. Adequate preoperative arterial perfusion to the ipsilateral hand was assessed by physical exam; preoperative arteriography or pressure measurements were not routinely performed. An Esmark tourniquet was applied from the hand to the upper arm in order to empty the upper extremity of blood. A pneumatic tourniquet was then placed on the proximal upper arm and inflated to 300 mmHg. An incision was made overlying the graft to the brachial artery anastomosis. The graft was completely detached from the artery and the arteriotomy was oversewn with a continuous running 5-0 Prolene suture line. All ligations occurred distal to the origin of the deep brachial artery. A small incision was then made over the venous anastomosis in order to detach the graft from the outflow vein. The vein was oversewn and the entire graft was removed. In cases where the AVG was densely incorporated to the surrounding tissues, an external stripper was used to detach it. Both wounds were left open, packed, and allowed to heal secondarily. All patients received empiric antibiotic therapy, or antibiotic therapy based on microbiologic data where available, from the time of diagnosis of AVG infection until post operative day 14.

Perioperative mortality was defined as death within 30 days of the procedure. Upper extremity ischemia was assessed by physical examination of motor strength, neurological function, and evaluation for subjective complaints of hand pain either with exercise or while at rest. This research was approved by the Institutional Review Board of the Mount Sinai School of Medicine.

RESULTS

Patient Population

During the study period, 45 AVG infections were treated in 43 patients. Of these, 22 patients with no anastomotic involvement were treated with partial excision and segmental bypass through an uncontaminated field. Two patients with anastomotic involvement were treated with total graft excision and autogenous patching of the arterial anastomotic site. The remaining 21 patients (49%), all of whom demonstrated arterial anastomotic involvement by duplex ultrasound or clinical findings were treated with BAL (Table 1); these 21 patients form the focus of this study There were 6 men and 15 women, with a mean age of 53.2 years (range, 32-67). The primary etiologies for ESRD were hypertension (29%), HIV (24%), and diabetes (19%). Follow-up was 100% at 1 month, 86% at 3 months, and 67% at 6 months. Because a number of the patients in this cohort were referred into the practice from outside practitioners, the total number of patients who underwent AVG placement during the study period was not obtainable. As a result, the infection incidence rate could not be calculated.

Table 1.

Brachial artery ligation and prosthetic arteriovenous graft excision cohort characteristics.

	Number (%)
Patients	21
Age (mean +/− SD)	53.2 +/− 9.5 yrs
Female gender	15 (71.4)
Cause of ESRD
Hypertension	6 (28.6)
HIV	6 (28.6)
Diabetes mellitus	4 (19.1)
Wegner’s granulomatosis	1 (4.8)
Pyelonephritis	1 (4.8)
Unknown	3 (14.3)
Arteriovenous Graft Characteristics
Upper arm (brachial artery-axillary vein)	20 (95.2)
Forearm (brachial artery-cephalic vein)	1 (4.8)
PTFE	19 (90.5)
Bovine Heterograft	2(9.5)

Treatment Characteristics

Of the AVG infections treated with BAL, 20 (95%) were straight upper arm brachial artery to axillary vein AVGs and 1 (5%) was a brachial artery to cephalic vein loop forearm AVG; all of the AVGs were made of PTFE except for 2 (9%) which were bovine heterografts. The mean time elapsed between original access placement and BAL was 21.3 months (range, 1 to 60; SD 16.6). At the time of BAL, all AVGs were documented to be functional and patent according to either color flow duplex exam or auscultation of a bruit. Data relating to the specific bacteriology associated with each AVG infection were not available in this cohort of patients.

All patients required the placement of a new dialysis access after their procedure. Initially, in the acute phase, temporary dialysis catheters were utilized. Once the patients demonstrated normalization of their fever curves and white blood cell counts as well as negative blood cultures for 48 hours, tunneled dialysis catheters were placed. At this time, repeat evaluation for long-term dialysis access options in the contralateral extremity were pursued (i.e. vein mapping and arterial inflow assessment).

Outcomes

All 21 patients (100%) who underwent BAL were seen and evaluated by the senior author immediately after surgery and at 1 month post-procedure. None of the patients complained of pain or coldness in the ipsilateral hand. On physical examination, motor strength and neurological sensation were unchanged from their preoperative examination. All patients had completed antibiotic therapy at 1 month post-procedure; all demonstrated complete resolution of infection as documented by physical exam, absence of fever, and normalization of the white blood cell count.

At 3 months, 18 (86%) of the patients returned for evaluation. At 6 months, 14 (67%) of the patients returned for follow-up. During the entire follow-up period there were no ischemic or septic complications observed in the cohort of patients who underwent BAL. Furthermore, no patients required additional interventions directed at the ipsilateral extremity.

DISCUSSION

Ligation of the brachial artery for the management of infected AVGs has traditionally been reserved for situations in which other therapies either fail or are not possible.^{8, 12-14} To our knowledge, this technique has not been promoted as a first-line therapy due to concern about the potential risk for upper extremity ischemia. In this study, no evidence of ischemia was seen in 21 patients treated with BAL, and all patients experienced complete resolution of their infections without need for additional interventions.

Acute traumatic ligation of the brachial artery has been reported to be accompanied by significant morbidity, including rates of amputation up to 26% in combat settings.¹⁵ As a result, in the civilian trauma setting, direct brachial artery repair has been advocated.^{16, 17} However, in this study, we report no ischemic complications following therapeutic BAL for AVG infections. Three possible explanations for this discrepancy arise. First, it is possible that the historical ischemic complication rates are higher than current rates, and may have improved with contemporary medical and surgical management. Second, outcomes following traumatic BAL may be confounded by concurrent upper extremity injury and therefore may not be generalizable to patients with AVG infections. Third, patients with a functioning AVG may adapt to decreased brachial artery inflow by developing compensatory collateral circulation: 91% of patients with a patent AVG demonstrate reversal of flow in the radial and ulnar arteries but never manifest access-induced ischemia (symptomatic steal syndrome).¹⁸ This adaptive increase in collateral circulation may allow ligation to be better tolerated in patients with chronically decreased brachial artery inflow.

One additional report has examined the effect of BAL for patients with a vascular access in place.¹⁴ Tan and colleagues found that of 30 patients who underwent BAL, three (10%) patients developed ischemia (two were successfully treated with bypass surgery, and one required amputation). Of note, in their series, all three of the patients who experienced ischemic complications underwent ligation of the brachial artery in the upper arm, proximal to the origin of the deep brachial artery. Conversely, as in our series, none of the patients in the Tan report who underwent BAL at the level of the elbow went on to develop any signs or symptoms of ischemia.

There are two primary advantages of BAL for the routine treatment of an infected AVG with anastomotic involvement. First, this strategy obviates the need for potentially technically difficult dissection and placement of an autogenous patch or oversew of a prosthetic remnant in a scarred and infected field. Second, BAL definitively manages the arterial anastomosis, and in doing so, avoids the potentially devastating risks of vein patch rupture,^{13, 19, 20} persistent prosthetic remnant infection^{8, 12} or anastomotic dehiscence²⁰.

A potential disadvantage associated with this treatment strategy is the possibility of limiting options for future ipsilateral AV access. Because many of the patients from this cohort were referred into the practice from outside practitioners and then returned to their primary access surgeon once the acute infection resolved, we lack complete follow-up relating to their next access site. Unfortunately, we are therefore unable to comment definitively on future hemodialysis access options after BAL. Anecdotally, it has been our practice to pursue contralateral extremity options and if none exist, we have performed several successful proximal brachial artery to axillary vein loop AVGs.

This study is subject to the limitations inherent in its retrospective study design, including a potential selection bias as well as the lack of a control group. In addition, assessment of ischemia was based on patient symptom report and physical exam findings, rather than on objective digital pressure measurement. Finally, this study reports on follow-up to six months post-procedure. Although the pathophysiology of BAL suggests that the risk of ischemia will be highest in the short term, data to substantiate this claim are not yet available; the possibility that patients described in this series may develop ischemic symptoms in the future persists. Despite these limitations, this series adds support to previously reported data, suggesting both a low complication risk and a high probability of resolution of infection when BAL is performed distal to the deep brachial artery for the management of AVG infection.

CONCLUSIONS

This study demonstrates that in 21 patients with an infected AVG, BAL resulted in complete resolution of the infection with no evidence of upper extremity ischemia. We therefore propose that BAL with complete removal of all prosthetic material is an effective and safe first-line therapy for patients who present with an AVG infection involving the brachial artery at the level of the elbow.