Revision vascular surgery

Introduction

Revision of vascular reconstructions is frequently required beyond the first 6 weeks because of progressive atherosclerosis, graft occlusion, aneurysm formation or infection. Up to 40% of infrainguinal bypass surgery grafts require reintervention within 5 years.¹ The same is true for endovascular intervention, where it has been shown that up to 30% of patients need a reintervention within 5 years after endovascular aortic aneurysm reconstruction (EVAR).^2–6 Revision surgery requires experience and judgment; open surgery is technically more difficult because of fibrosis and the loss of easily definable tissue planes, which necessitates careful sharp dissection to gain arterial control. The anatomy may be altered by previous operations or interventions, and operating times, blood loss, infection rates and operative risk are therefore increased. Endovascular revision requires careful assessment with regard to when and how to intervene, and especially when to convert to open surgery.

Graft occlusion

Graft thrombosis usually presents acutely but is occasionally foreshown by increasing ischaemic symptoms, ranging from mild claudication to critical ischaemia; asymptomatic graft occlusion can also occur. Occasionally, simultaneous distal embolisation causes digital ischaemia (blue toe syndrome or gangrene). Graft thrombosis with loss of run-off due to distal embolisation is associated with a high risk of limb loss. Whereas graft thrombosis in the first 6 weeks is generally due to technical error or poor run-off, most late occlusions result from intimal hyperplasia within the bypass, progressive inflow or run-off disease (see Chapter 4). Graft stenoses are usually asymptomatic and occur in 20–30% of infrainguinal vein grafts, mostly in the first year.⁷^,⁸ Stenoses of greater than 70% (velocity > 3 m/s or velocity ratio > 3.0) compromise flow and often occlude if left untreated.⁹

Factors influencing graft occlusion

Local factors

These are essentially the quality of the inflow, the run-off and the conduit itself (see Chapter 4).

Patency is better for suprainguinal than infrainguinal grafts and graft occlusion is more frequent in femorotibial than in femoropopliteal bypasses.¹⁰ Infrainguinal bypass patency of autologous vein is better than Dacron, polytetrafluoroethylene (PTFE) or human umbilical vein.¹¹ Dacron and PTFE used in femoropopliteal bypass grafting did not differ at 5 years.¹² However, a recently published Scandinavian multicentre randomised trial revealed that a heparin-bonded PTFE graft significantly reduced the overall risk of primary graft failure by 37%. Risk reduction was 50% in femoropopliteal bypass cases and in cases with critical ischaemia.¹³ Registry data acknowledged these results.¹⁴

In terms of technique, the results of reversed and in situ vein grafts are equivalent.¹⁵^,¹⁶ In a meta-analysis on the long-term primary and secondary patency and foot preservation after popliteal-to-distal bypass grafts, there was a superiority trend favouring reversed vein grafts.¹⁷

Arm veins are similar to the long saphenous vein provided that angioscopically detected defects are corrected.¹⁸ Preoperative ultrasound mapping is important to identify the best autogenous conduit available.¹⁹ A recently published retrospective study of infrainguinal bypasses for critical limb ischaemia (CLI) using arm vein conduits or prosthetic grafts revealed that arm vein conduits, even when spliced, are superior to prosthetic grafts in terms of midterm assisted primary patency, secondary patency, and leg salvage in infrapopliteal bypasses for CLI.²⁰

A recently published Cochrane Review demonstrated a clear primary patency benefit for autologous vein when compared to synthetic materials for above-knee bypasses. In the long term (5 years) Dacron confers a small primary patency benefit over PTFE for above-knee bypass. PTFE with a vein cuff improved primary patency when compared to PTFE alone for below-knee bypasses.²¹

The risks of ischaemic complications and the need for emergency limb revascularisation are greater with occlusion of prosthetic grafts compared to venous grafts. This is because the thrombus in the prosthetic graft extends into the outflow artery. Vein cuffs at the distal anastomosis reduce the risk of outflow impairment after thrombosis of PTFE grafts (Fig. 7.1).²²

Figure 7.1 Femoropopliteal bypass with venous cuff (Miller cuff) at the distal anastomosis. (a) Perioperative angiography. (b) Preservation of outflow vessels after thrombosis at 3 years.

While most studies demonstrate a clear advantage for vein, the Dutch BOA Study found no difference between venous and prosthetic graft material in the risk of amputation after infrainguinal bypass occlusion.²³

The quality and number of run-off vessels appear to potentially identify patients at high risk for a poor initial outcome from infrainguinal bypass reconstructions.²⁴^,²⁵ Bypasses for gangrene are at higher risk of occlusion than those performed for other indications, probably related to the poorer run-off associated with worsening ischaemia.²⁶

General factors

Continued smoking after lower limb bypass surgery results in an at least threefold increased risk of graft failure.²⁷ Diabetes and renal failure compromise patient survival but not graft patency.^27–29

Raised fibrinogen, hyperlipidaemia, thrombophilias (e.g. protein C, protein S or antithrombin III deficiency, antiphospholipid antibodies) and increased platelet aggregation favour graft thrombosis.³⁰ The role of factor V Leiden mutation is debatable.³¹^,³²

Black race and female gender are risk factors for adverse outcomes after vein bypass surgery for limb salvage, with graft failure and limb loss being more common events in black patients, and black women being a particularly high-risk group.³³ Hormone replacement therapy potentiates this increased risk.³⁴^,³⁵

Prevention of graft thrombosis

Antiplatelet therapy with aspirin had a slight beneficial effect on the patency of peripheral bypass grafts but seemed to have a lesser effect on venous graft patency compared with artificial grafts.³⁶ The Dutch BOA trial revealed that aspirin is more effective for infrainguinal prosthetic grafts, whereas warfarin is better for vein grafts.³⁷ There is, however, no indication for routine warfarinisation.

In the CASPAR trial the combination of clopidogrel plus aspirin did not improve limb or systemic outcomes in the overall population of peripheral arterial disease (PAD) patients requiring below-knee bypass grafting. A subgroup analysis suggested that clopidogrel plus aspirin has benefit in patients with prosthetic grafts without significantly increasing major bleeding risk.³⁸

Surveillance for suprainguinal bypass grafts does not seem to be cost-effective, offering relatively small clinical benefit. Occlusion is far more frequent after infrainguinal bypass grafting and since most of the occlusions occur within 2 years after the implantation, patients should be carefully followed during this time period. This includes medical history, clinical examination and Doppler pressure measurements. Additional duplex scanning is not effective after prosthetic bypass.

Some published reports tend to argue in favour of duplex surveillance on the basis of patency alone.³⁹ In a recent large randomised trial, intensive surveillance with duplex scanning did not show any additional benefit in terms of limb salvage rates for patients undergoing femoropopliteal or femorocrural vein bypass graft operations.⁴⁰

Management of graft stenosis (the failing graft)

Although there is some discussion about asymptomatic stenosis, it is clear that symptomatic graft stenosis should be treated by either angioplasty or surgical revision. Open surgical revision of infrainguinal vein grafts provides an increased freedom from further reinterventions or major amputation; however, early success rates for endovascular procedures are similar, particularly for non-occluded grafts. With time, endovascular revisions require an increasing number of reinterventions and manifested higher rates of failure.⁴¹ Surgical revision is probably more durable in the longer term, but an endovascular approach is actually preferred because of the acceptable short-term patency and a low rate of complications, particularly for late graft stenosis (> 3 months), short (< 2 cm) and single stenoses.⁴²^,⁴³

Short graft stenosis, whether midgraft or anastomotic, usually responds well to angioplasty with high inflation pressures (up to 2020 kPa; Fig. 7.2). Cutting balloons have been advocated, but seem to offer only small benefit at the expense of an elevated complication rate.^44–46 Stents (preferably self-expandable bare nitinol) should be used only selectively.⁴⁷

Figure 7.2 Vein graft stenosis near the below-knee popliteal anastomosis (a), successfully treated by balloon angioplasty (b).

A randomised clinical study demonstrated that implantation of a polymer-free, paclitaxel-coated nitinol stent in patients with moderate-length lesions of the superficial femoral artery and proximal popliteal artery was safe and associated with a superior 12-month patency compared with both percutaneous transluminal angioplasty (PTA) and bare metal stent placement.⁴⁸

Longer graft stenoses are best treated by open surgery and may be bypassed using the contralateral long saphenous or superficial femoral vein.¹⁸ Tibial or distal popliteal anastomotic stenoses, resistant to angioplasty, are best treated by a jump graft to a fresh run-off vessel in order to avoid the scar tissue or adherent tibial veins (Fig. 7.3).

Figure 7.3 Vein patch angioplasty (a), bypass of long vein graft stenosis (b) and jump graft around stenosed distal anastomosis of a femoropopliteal bypass graft to the popliteal artery (c).

The underlying cause of axillofemoral graft occlusion is most frequently an anastomotic or inflow vessel stenosis.⁵⁵ Thrombectomy can be performed from the groin, but most often an incision along the graft is needed. The underlying cause needs to be addressed. A new graft in unscarred tissues offers the best solution in cases of long-standing thrombosis or when thrombectomy is insufficient. For an occluded femorofemoral graft the same principles apply.

Infrainguinal graft thrombosis

In cases of chronic occlusion, the decision regarding revascularisation should be taken in the light of the clinical history, examination and results of investigations (Doppler, duplex scan, angiography).

In patients with acute ischaemia, an immediate decision should be taken regarding thrombolysis or open surgery. Prosthetic grafts can usually be thrombectomised by a groin approach if the occlusion is less than 1 week old. Anastomotic stenosis must be corrected, and inflow and outflow vessels need to be checked by perioperative angiography. If necessary, outflow vessels can be selectively approached by an infragenicular approach, and intraoperative thrombolysis may be an adjunct in selected cases (see Chapter 8).⁵⁶ Venous grafts are usually more difficult to thrombectomise. Here it may be wise to place a new graft, as is the case in prosthetic grafts where thrombectomy is insufficient. If at all possible, a vein conduit should be used. Following thrombectomy for acute ischaemia, four-compartment fasciotomy should be considered to relieve pressure and improve distal perfusion, especially if there is any calf swelling or tenderness preoperatively.⁵⁷^,⁵⁸

Graft infection

Graft infection is relatively uncommon (1–5%) but has a high amputation and mortality risk.⁵⁹ In a recent multicentre audit of 55 graft infections, 31% died, 33% underwent amputation and only 45% left the hospital alive without amputation.⁶⁰ Treatment has therefore to focus on patient survival, eradication of infection, and revascularisation by a method that is durable and does not predispose to further infection.

Causes

Graft infection is thought to occur most commonly by inoculation of bacteria from the patient’s skin at the time of surgery (e.g. skin commensals) or by direct spread in the perioperative period, often secondary to wound breakdown.⁶¹ Surgical site infection (SSI) after open surgery for lower extremity revascularisation is a serious complication that is associated with a more than twofold increased risk of early graft loss and re-operation.⁶²

The risk of infection is higher in patients with gangrene, the elderly, obese and those undergoing re-operation during the same hospital admission. Preoperative shaving, open surgical drainage for more than 3 days, operations lasting over 4 hours, emergency surgery, redo surgery, female gender, diabetes, steroids, renal failure, recent angiography and wound haematoma are further risk factors.⁶²^,⁶³ Blood-borne bacteria from intravenous lines or systemic infections may also cause graft inoculation and sepsis.

Venous grafts are more resistant to infection than prostheses but direct bacterial erosion can occur, especially when exposed in an open wound.

Prevention

Patients should be admitted as near to surgery as possible and isolated from patients with known infections, especially methicillin-resistant Staphylococcus aureus (MRSA). Many hospitals are adopting a policy of screening elective surgical patients for MRSA colonisation before admission and eradicating any infection before proceeding with the surgery.

Strict aseptic technique and laminar air-flow theatres minimise infection rates. Iodine-impregnated adhesive drapes help isolate the operative field.

Prophylactic antibiotics (cephalosporin or co-amoxiclav) reduce wound and graft infection. There is no evidence for using more than three perioperative doses. Some surgeons add a single dose of gentamicin or vancomycin as cover for MRSA.

Prophylactic systemic antibiotics reduce the risk of wound infection and early graft infection. Antibiotic prophylaxis for more than 24 hours appears to be of no added benefit. There was no evidence that prophylactic rifampicin bonding to Dacron grafts reduced graft infection at either 1 month or 2 years.⁶⁴ The same is true for silver-coated grafts.⁶⁵^,⁶⁶ There was no evidence of a beneficial or detrimental effect on rates of wound infection with suction groin-wound drainage⁶⁴ or of any benefit from a preoperative bathing or shower regimen with antiseptic agents.⁶⁷

There is also little evidence regarding the need for prophylactic antibiotics before other surgical or dental procedures in the presence of prosthetic grafts.

Presentation

Wound infections after vascular surgery are classified according to the depth of tissue involvement: type 1 involves the skin only, type 2 the subcutaneous tissue and type 3 the graft itself.⁶⁸ Prosthetic graft infection can present at any time from days to years after surgery with pyrexia, systemic sepsis, local abscesses and sinuses, graft exposure, thrombosis or anastomotic haemorrhage. On rare occasions septic emboli can be the first sign (Fig. 7.5). Septic erosion of exposed vein grafts can occur at any point.

Figure 7.5 Septic emboli at the foot as the presenting sign of aortofemoral graft infection.

Infrarenal aortic grafts can erode the third or fourth part of the duodenum causing an aortoduodenal fistula, which may present with one or two sentinel gastrointestinal bleeds before the inevitable catastrophic haemorrhage. Occasionally, an aortic graft can also erode any other part of the bowel, including the appendix, and also the ureter. Such aortoenteric erosion will lead to localised peritonitis with retroperitoneal or groin abscesses (Fig. 7.6a, b) and ultimately catastrophic bleeding. The mortality rate of aortoenteric complications is high (> 50%), with recurrent infection or aortic stump blowout in over 25%.⁶⁹

Figure 7.6 Aortoenteric fistula leading to retroperitoneal and groin abscesses. (a) Aspect of the groin. (b) CT scan.

Bacteriology

Most graft infections are due to skin organisms.⁷⁰ Staphylococcus epidermidis is the least virulent, producing a biofilm or an infected seroma after months or years. It is difficult to culture, requiring homogenisation of explanted graft material to dislodge adherent bacteria. Staphylococcus aureus is more virulent and usually presents earlier. MRSA infections have a particularly high morbidity and mortality and the incidence unfortunately is increasing. Apart from staphylococci, there are a wide variety of organisms that can cause graft infection. Gram-negative species include Escherichia coli and Pseudomonas aeruginosa, the latter being recognised by a high tendency of anastomotic disruption and bleeding. It may be difficult or even impossible to culture the causative organism, particularly after prolonged periods of antibiotic therapy.

Diagnosis

In most cases there is a perigraft collection or sinus. Aspiration of frank pus or turbid fluid from around the graft and the subsequent culture of a causative organism are diagnostic. Computed tomography (CT), magnetic resonance imaging (MRI) or ultrasound usually demonstrate perigraft fluid and inflammation but can underestimate the extent of infection, especially if a sinus is present. A sinogram may be useful in very selective cases, though other modes of imaging will usually suffice. Aortic graft infection is often difficult to diagnose, requiring a high index of suspicion and aggressive investigation. Leucocyte count, erythrocyte sedimentation rate and C-reactive protein are often raised. Persistence of perigraft fluid or perigraft soft-tissue attenuation beyond 3 months or perigraft gas (Fig. 7.7) beyond 4–7 weeks should be presumed to be infection. Perigraft fluid can be aspirated using an image-guided technique and characterised for pathogens, though this does pose a further infection risk.⁷¹ One in four anastomotic aneurysms results from graft infection, but this is not always evident from CT. Where doubt exists, indium-labelled leucocyte or positron emission tomography (PET) scanning is occasionally helpful. In aortoenteric fistula, the graft may be seen eroding the duodenum at endoscopy (Fig. 7.8). Definitive confirmation of an infection is sometimes only made at the operation by the presence of pus around the graft and the absence of tissue incorporation (Fig. 7.9).

Figure 7.7 Perigraft gas bubbles in between duodenum and aortic graft (see arrow).

Figure 7.8 Aortic graft, eroding the duodenum as seen during endoscopy.

Figure 7.9 Intraoperative situs with absence of tissue incorporation of the graft and discoloration by duodenal secretion.

Management

General principles

Once infection is confirmed, semi-urgent treatment is required to pre-empt catastrophic haemorrhage, graft thrombosis or systemic sepsis. An infected prosthesis acts as a foreign body, rendering bacteria inaccessible to antibiotics. Conservative measures (including prolonged antibiotic therapy, drainage and irrigation of abscesses, muscle flaps) may be helpful and can buy time, but they are rarely curative.

The mainstay of vascular graft infection management is as follows. First, excision of the graft, as this as a foreign body that may potentiate the infection. Second, wide and complete debridement of devitalised and infected tissue to provide a clean wound in which healing can occur. Third, establishing a vascular flow to the distal bed. Fourth, intensive and prolonged treatment with antibiotics, in order to reduce the risk for sepsis and secondary graft infection.⁷²

Complete graft excision may be mandatory. Partial graft resection commonly requires later replacement.⁷³ Although contrary to conventional concepts, partial or complete graft preservation combined with aggressive drainage and groin wound debridement is an acceptable option for treatment of infection involving an entire aortic graft in selected patients with prohibitive risks for total graft excision.⁷⁴^,⁷⁵

Simple graft excision without revascularisation can heal infection but usually results in major amputation, even in patients operated on initially for claudication. Revascularisation is traditionally performed by extra-anatomical bypass (axillofemoral graft for infrarenal aortic graft infection, obturator bypass for infection at the groin, lateral bypass for infection of a femoropopliteal graft).⁵⁹