Comparison of angioplasty and bypass surgery for critical limb ischaemia in patients with infrapopliteal peripheral artery disease

Transcription

1 Original article Comparison of angioplasty and bypass surgery for critical limb ischaemia in patients with infrapopliteal peripheral artery disease S. D. Patel 1,L.Biasi 1, I. Paraskevopoulos 2, J. Silickas 1,T.Lea 1, A. Diamantopoulos 2, K. Katsanos 2 and H. Zayed 1 Departments of 1 Vascular and Surgery and 2 Interventional Radiology, Guy s and St Thomas NHS Foundation Trust, London, UK Correspondence to: Mr S. D. Patel, Department of Vascular Surgery, Guy s and St Thomas NHS Foundation Trust, St Thomas Hospital, 1st Floor North Wing, Westminster Bridge Road, London SE1 7EH, UK ( drsanjaypatel@hotmail.com) Background: Both infrapopliteal (IP) bypass surgery and percutaneous transluminal angioplasty have been shown to be effective in patients with critical limb ischaemia (CLI). The most appropriate method of revascularization has yet to be established, as no randomized trials have been reported. The aim of this study was to compare the outcomes of patients with similar characteristics treated using either revascularization method. Methods: Consecutive patients undergoing IP bypass and IP angioplasty for CLI (Rutherford 4 6) at a single institution were compared following propensity score matching. The study endpoints were primary, assisted primary and secondary patency, and amputation-free survival at 12 months, calculated by Kaplan Meier analysis. Results: Some 279 limbs in 243 patients were included in the study. The two groups differed significantly with respect to the incidence of diabetes (P = 0 024), estimated glomerular filtration rate (P = 0 006), total lesion length (P < 0 001) and Rutherford classification (P = 0 008). These factors were used to construct the propensity score model, which yielded a matched cohort of 125 legs in each group. Primary patency (54 4 versus 51 4 per cent; P = 0 014), assisted primary patency (77 5 versus 62 7 per cent; P = 0 003), secondary patency (84 4 versus 65 8 per cent; P < 0 001) and amputation-free survival (78 7 versus 74 1 per cent; P = 0 043) were significantly better after bypass than angioplasty. However, limb salvage was similar (90 4 versus 94 2 per cent; P = 0 161), and overall complications (36 0 versus 21 6 per cent; P = 0 041) as well as length of hospital stay (18(4 134) versus 5(0 110); P = 0 001) were worse in the surgical bypass group. Conclusion: There was no difference in limb salvage rates, but patency and amputation-free survival rates were better 1 year after bypass surgery. Paper accepted 7 July 16 Published online 21 September 16 in Wiley Online Library ( DOI: 10.2/bjs Introduction The management of critical limb ischaemia (CLI) in patients with distal infrapopliteal (IP) arterial disease remains a major challenge. There is limited high-quality evidence to support treatment choices in this area, with both bypass surgery and endovascular intervention shown to be effective in preventing limb loss 1 3. IP bypass surgery is associated with amputation-free survival and patency rates approaching per cent at 1 year in experienced centres 4. Overall survival and limb salvage rates vary between centres, ranging from 79 to 90 and 66 to per cent respectively at 1 year 4 6. Data from RCTs are limited to the BASIL ( versus Angioplasty in Severe Ischaemia of the Leg) 7 and PREVENT III 8 trials. The latter included 14 patients with CLI with an IP target vessel in 65 per cent of patients. The primary patency, limb salvage and survival rates at 1 year were 61 0, 88 0 and84 0 per cent respectively 8. The BASIL and PREVENT III trials reported perioperative mortality rates of 5 5and2 7 per cent respectively 7 9, giving a clear insight into the risks associated with bypass surgery in an already elderly and co-morbid group of patients. This, together with advances in balloon and stent technology, has led many centres to support an endovascular-first approach in patients with CLI due to IP 16 BJS Society Ltd BJS 16; 103:

2 1816 S. D. Patel, L. Biasi, I. Paraskevopoulos, J. Silickas, T. Lea, A. Diamantopoulos et al. disease 10,11. Recent systematic reviews and meta-analysis 2,3 of randomized trials have shown endovascular treatment to be technically feasible and relatively safe in this group of patients, with a technical success rate of up to 96 per cent. One-year outcomes for primary patency, secondary patency, limb salvage and overall survival were 66, 74, 88 and 88 per cent respectively 2,3. These results are comparable to those for IP bypass, although any such comparison is undermined by confounding factors and selection bias. RCTs in this area will not report for several years and there is currently a lack of data offering meaningful comparisons between the different treatment modalities. The aim of this study was, therefore, to compare outcomes in consecutive patients undergoing IP bypass surgery with those undergoing IP endovascular therapy for patients with CLI in a single centre, using propensity score analysis to allow valid comparisons to be made. Methods An analysis was undertaken of consecutive patients undergoing IP revascularization for CLI (Rutherford 4 6) 12,13 in a single centre between 10 and 14. A prospectively collected database including patients demographic information, cardiovascular risk factors, angiographic findings, procedural variables and follow-up results (clinical and radiological) was analysed. The revascularization strategy was stratified into surgical bypass or endovascular therapy. Revascularization was defined as IP if the distal anastomosis (in bypass group) or target vessel recanalization (endovascular group) involved the anterior tibial, tibioperoneal trunk, peroneal artery, posterior tibial artery or dorsalis pedis artery, with or without a concomitant inflow procedure. All patients gave informed consent, which included data collection. In accordance with National Health Service Research and Ethics definitions (Institutional Review Board equivalent; this study is not classified as research requiring formal ethics approval. All patients with CLI were considered for revascularization. Diagnostic imaging included duplex ultrasonography as the first imaging modality, followed by CT angiography (CTA) or magnetic resonance angiography as indicated. All patients were discussed in a dedicated multidisciplinary team (MDT) meeting (including a diabetic foot MDT meeting where appropriate) where the best revascularization strategy was agreed upon, after careful consideration of patients co-morbidities, availability of suitable venous conduit, anatomical distribution and extent of the peripheral arterial disease. Within the endovascular group, 17 patients (13 6 per cent) did not have an adequate vein conduit. Procedural details Technical details of the bypass procedure have been published previously 1,4. Preoperative duplex imaging was used to identify a venous conduit where possible, with the great saphenous vein (GSV) as the preferred conduit, followed by the short saphenous vein (SSV) or arm veins (cephalic and basilic veins). grafts were tunnelled anatomically, and used either reversed or non-reversed (with valvulotome) as deemed appropriate, depending on the size match between the vein and the inflow and outflow artery. interventions were undertaken either in a dedicated angiography suite or in hybrid theatre by consultant interventional radiologists or consultant vascular surgeons. The lesion was treated according to the operator s preference, preferentially intraluminally (rather than subintimally) by balloon angioplasty (plain balloon angioplasty, drug-coated balloon); stenting (bare metal stenting, drug-eluting stenting) was considered as a bail-out option in the event of suboptimal results. For both bypass surgery and endovascular therapy, unfractionated heparin ( units/kg) was given intravenously and additional boluses administered to maintain the activated clotting time between 0 and 300 s. Patients received dual antiplatelet therapy immediately after the procedure for 3 6 months, unless contraindicated. Patients already receiving anticoagulation for a different medical condition were discharged on 75 mg aspirin in addition to their anticoagulant therapy. Patients were enrolled in a duplex surveillance programme consisting of imaging before discharge, and 3, 6, 9 and 12 months after the procedure, and yearly thereafter, if no intervention was necessary after bypass surgery; and scans at 6 weeks, 6 and 12 months, and yearly thereafter following endovascular treatment. Study endpoints The primary endpoints were primary patency, assisted primary patency, secondary patency, amputation-free survival and limb salvage, defined according to the published Society for Vascular Surgery reporting standards 13. For the endovascular group, technical success per limb was defined as successful recanalization of at least one tibial artery with straight inline flow to the foot and a residual stenosis of less than 30 per cent on completion angiography. Technical success per target vessel was defined as immediate patency with residual stenosis of less than 30 per cent for each of the treated arteries. Lesion length for the bypass group was defined as the continuous length of artery bypassed. Lesion length for

3 Angioplasty and bypass surgery for critical limb ischaemia in infrapopliteal peripheral artery disease 1817 Table 1 Preprocedural characteristics of the cohort before and after propensity score matching surgery (n = 127) Unmatched cohort therapy (n = 152) P surgery (n = 125) PS-matched cohort therapy (n = 125) P Age (years)* 74(10) 73(13) (10) 73(11) Sex ratio (M : F) 92 : : : : Ischaemic heart disease 31 (24 4) 30 (19 7) (24 8) 25 ( 0) Stoke/transient ischaemic attack (15 7) (13 2) (16 0) 15 (12 0) Diabetes mellitus 62 (48 9) 96 (63 2) (48 8) 69 (55 2) Smoker 52 ( 9) (39 5) (41 6) 51 ( 8) Hypertension 98 (77 2) 108 (71 1) (76 8) 87 (69 6) Hypercholesterolaemia 59 (46 4) (39 5) (46 4) 52 (41 6) egfr (ml per min per 1 73 m 2 )* 74(32) 63(31) (32) 68(31) Rutherford category (46 5) 49 (32 2) 58 (46 4) 43 (34 4) 5 56 (44 1) 71 (46 7) 55 (44 0) 61 (48 8) 6 12 (9 4) 32 (21 1) 12 (9 6) 21 (16 8) Femoropopliteal TASC type A 2 (1 6) 8 (5 3) 2 (1 6) 8 (6 4) B 26 ( 5) 17 (11 2) 24 (19 2) 16 (12 8) C 25 (19 7) 25 (16 4) 25 ( 0) 21 (16 8) D (31 5) 34 (22 4) (32 0) 29 (23 2) Tibial TASC type B 0 (0) 2 (1 3) 0 (0) 2 (1 6) C 45 (35 4) 38 (25 0) 43 (34 4) 30 (24 0) D 82 (64 6) 112 (73 7) 82 (65 6) 93 (74 4) Lesion length (cm)* 32(13) 26(12) < (13) 28(12) Values in parentheses are percentages unless indicated otherwise; *values are mean(s.d.). PS, propensity score; egfr, estimated glomerular filtration rate; TASC, TransAtlantic Inter-Society Consensus. χ 2 test, except independent-samples t test. Table 2 Binary logistic regression analysis showing preoperative factors favouring bypass surgery as the treatment choice Hazard ratio Diabetes 1 (0 97, 2 64) egfr 1 01 (1 00, 1 02) Lesion length 1 03 (1 01, 1 06) Values in parentheses are 95 per cent confidence intervals. egfr, estimated glomerular filtration rate. the endovascular group was defined as the length of treated artery. Where more than one IP artery was treated, the shortest length achieving inline flow to the foot was taken. Measurements were done using CTA reconstructions on three-dimensional workstations (Aquarius intuition Viewer ; Aquarius, TeraRecon, San Matteo, California, USA). Morbidity and mortality data were collected retrospectively, and further subclassified using the Clavien Dindo scale (grade I V) 14. Statistical analysis A propensity score model was constructed using logistic regression analysis. All preoperative factors were compared in the two groups (bypass versus endovascular therapy) and P Table 3 Target vessels in the bypass surgery and endovascular therapy groups surgery (n = 125) therapy (n = 244) Tibioperoneal trunk 25 ( 0) 34 (13 9) Anterior tibial artery 33 (26 4) 92 (37 7) Posterior tibial artery 34 (27 2) 64 (26 2) Peroneal artery 22 (17 6) 54 (22 1) Dorsalis pedis artery 11 (8 8) 0 (0) Values in parentheses are percentages. P = (χ 2 test). factors found to be statistically different (P < 0 050) were then used to construct the model. Continuous variables are expressed as mean(s.d.) for normally distributed data and median (range) for those without a normal distribution, and compared using the independent samples t test and Mann Whitney U test respectively. Categorical variables were compared using the χ 2 test. Primary endpoints were analysed using the Kaplan Meier method, expressed as percentage survival, with 95 per cent confidence interval (c.i.), and compared by means of the log rank test. P < was considered statistically significant. All analyses were carried out using GraphPad Prism 6 (GraphPad Software, San Diego, California, USA) and SPSS version 22 (IBM, Armonk, New York, USA).

4 1818 S. D. Patel, L. Biasi, I. Paraskevopoulos, J. Silickas, T. Lea, A. Diamantopoulos et al. Primary patency (~) surgery therapy Assisted primary patency (~) a Primary patency b Assisted primary patency Secondary patency (~) 244 c Secondary patency Fig. 1 Kaplan Meier analysis of patency after bypass surgery versus endovascular therapy: a primary patency, b assisted primary patency and c secondary patency. Estimates are shown with 95 per cent confidence intervals. a P = 0 014, b P = 0 003, c P < (log rank test) Results During the study interval, a total of 279 limbs in 243 patients underwent lower limb IP revascularization for CLI (127 bypass surgery, 152 endovascular therapy). Propensity score-matched cohorts were created and yielded a matched cohort of 125 patients in each group (Table 1). The original bypass and endovascular intervention groups differed significantly with respect to the incidence of diabetes (48 9 versus 63 2 per cent respectively; P = 0 024), estimated glomerular filtration rate (egfr) (mean(s.d.) 74(32) versus 63(31) ml per min per 1 73 m 2 ; P = 0 006) and total lesion length (mean 32(13) versus 26(12) cm; P < 0 001). Rutherford classification was also significantly different in the two groups (P = 0 008), with more tissue loss in the endovascular group. Although femoropoliteal TransAtlantic Inter-Society Consensus (TASC) classification did not differ significantly between the groups (P = 0 138), the percentage of patients with no femoropopliteal disease was significantly higher in the endovascular group (26 8 versus 44 7 per cent; P = 0 002). Binary logistic regression analysis was used to identify independent factors influencing the treatment decision between surgical bypass and endovascular intervention (Table 2). A lower egfr, shorter lesion length and the presence of diabetes were found to predict the selection of endovascular treatment as the method of revascularization. These factors, in addition to femoropopliteal TASC classification and Rutherford category, were used to construct the propensity score model. These factors were no longer statistically different following propensity scoring matching (Table 1), with the exception of the percentage of

5 Angioplasty and bypass surgery for critical limb ischaemia in infrapopliteal peripheral artery disease 1819 Limb salvage (~) surgery therapy Amputation-free survival (~) a Limb salvage b Amputation-free survival Fig. 2 Kaplan Meier analysis of limb salvage after bypass surgery versus endovascular therapy: a limb salvage and b amputation-free survival. Estimates are shown with 95 per cent confidence intervals. a P = 0 161, b P = (log rank test) patients with no femoropopliteal disease, which remained significantly higher in the endovascular group (27 2 versus 8 per cent for bypass versus endovascular therapy; P = 0 028). Outcomes in propensity matched cohort Indications for treatment in the matched cohort were Rutherford category 4 (46 4 versus 34 4 per cent in bypass and endovascular groups respectively), category 5 (44 0 versus 48 8 per cent) and category 6 (9 6 versus 16 8 per cent) CLI. The proximal anastomosis involved the common femoral (34 4 per cent), above-knee popliteal (44 0 per cent), below-knee popliteal ( 8 per cent) or adjacent tibial (1 6 per cent) artery. Venous conduit was the GSV (90 4 per cent), SSV (1 6 per cent), arm vein (4 8 per cent) or composite vein (3 2 per cent). The vein was used either reversed (48 0 per cent) or non-reversed (52 0 per cent), and had a mean preoperative diameter of 3 5(1 5) mm. There were 244 IP endovascular target vessels (Table 3). A single vessel was targeted in 42 patients (33 6 per cent), with multiple vessels targeted in the remaining 83 (66 4 per cent). Treatment modality included plain balloon angioplasty (91 8 per cent), drug-coated balloon (7 8 per cent), bare metal stenting (2 0 per cent) and drug-eluting stenting (34 0 per cent). At completion angiography, the technical success rate per target vessel was 86 9 per cent, and the technical success rate per limb was 92 8 per cent. Comparing outcomes of surgical bypass and endovascular therapy after 1 year, the primary patency rate was 54 4 (95 per cent c.i to 62 7) versus 51 4 (42 3 to 62 1) per cent (P = 0 014), assisted primary patency 77 5 (69 8 to 86 4) versus 62 7 (54 3 to74 1) per cent (P = 0 003) and secondary patency 84 4 (75 6 to92 8) versus 65 8 (57 3 to 79 4) (P < 0 001), all significantly better after bypass surgery (Fig. 1). Rates of freedom from reintervention/target lesion revascularization at 1 year were similar in the two groups: 64 4 (53 7to74 2) versus 71 2( 4to78 9) per cent (P = 0 354). Over the study interval there were 15 major limb amputations in the bypass group and six in the endovascular group; however, follow-up was significantly longer in the bypass group (mean (14) months versus 10(9) months in bypass group; P = 0 001). Kaplan Meier analysis demonstrated that limb salvage was not significantly different between the groups at 1 year: 90 4 (83 8 to95 3) per cent after bypass versus 94 2(85 3to97 1) per cent after endovascular treatment (P = 0 161) (Fig. 2a). There was one death (0 8 per cent) in the bypass group and two (1 6 per cent) in the EV group in the first 30 days. At 1 year, overall survival (87 7 ( 9 to 93 4) versus 79 1 (68 6 to86 4) per cent; P = 0 004) and amputation-free survival (78 7 (71 4 to 85 3) versus 74 1 (63 7 to 82 8) per cent; P = 0 043) were significantly better in the bypass surgery group (Fig. 2b). The most common perioperative/procedural complications in the surgical bypass group were wound infection (8 8 per cent), pneumonia (8 8 per cent), arrhythmia requiring medical intervention (8 8 per cent), acute kidney injury (6 4 per cent), graft thrombosis requiring thrombectomy (4 0 per cent) and urinary tract infection (4 0 per cent). In the endovascular group they were acute kidney injury (4 0 per cent), arterial rupture (3 2 per

6 18 S. D. Patel, L. Biasi, I. Paraskevopoulos, J. Silickas, T. Lea, A. Diamantopoulos et al. Table 4 Morbidity and mortality stratified according to the Clavien Dindo classification Complication grade surgery (n = 125) therapy (n = 125) P* I 15 (12 0) 7 (5 6) II 16 (12 8) 5 (4 0) III 11 (8 8) 12 (9 6) IV 1 (0 8) 0 (0) V 1 (0 8) 2 (1 6) Values in parentheses are percentages. *χ 2 test; grade I and II (minor complications), grade III and IV (major complications). cent), arterial dissection requiring an unplanned stent (3 2 per cent), pseudoaneurysm (2 4 per cent) and groin bleeding requiring surgical intervention (2 4 per cent). Overall complication rates were higher in the bypass than in the endovascular treatment group (36 0 versus 21 6 per cent; P = 0 041). Comparison of morbidity according to Clavien Dindo grade showed that minor complications (grade I and II) were significantly more common in the surgical bypass group (P = 0 028; relative risk (RR) 2 08, 95 per cent c.i to 3 91) (Table 4), whereas there was no difference in the incidence of major complications (grade III IV) (P = 0 562; RR 1 12, 95 per cent c.i to 2 29). The median duration of hospital stay was significantly longer in the bypass surgery group than in the endovascular group: 18 (range 4 134) versus 5 (0 110) days (P = 0 001). Discussion In propensity score-matched groups of patients with CLI due to IP disease, outcomes at 1 year were significantly better after bypass surgery than endovascular therapy in terms of patency, amputation-free survival and overall survival. However, limb salvage rates were similar and the incidence of all (but not major) complications and length of stay was higher in the bypass surgery group. A recent update by the TASC steering committee 10 and the Task Force on the Diagnosis and Treatment of Peripheral Artery Diseases of the European Society of Cardiology 15 recommended an endovascular-first approach in patients with CLI and IP disease. There are currently two RCTs (BEST-CLI (Best versus best Surgical Therapy in patients with Critical Limb Ischaemia) 16 and BASIL-2 17 ) that will report on the outcomes of patients with CLI and IP disease. Until these trials publish their findings, statistical methods such as regression and propensity score models offer the best way to draw meaningful conclusions that help clinicians to make treatment decisions. In this study, the decision on the best revascularization strategy was made in an MDT meeting with surgeons and interventional radiologists, and aided by information from specialist geriatricians who helped in the assessment and optimization of elderly patients with CLI. In general, fitness for open surgical intervention, availability of a vein conduit, longer lesion length and good run-off are factors favouring open intervention. Conversely, shorter lesion length, higher operative risk and lack of a venous conduit favour endovascular intervention. This highlights the inherent confounding present when comparing treatment groups or studies outside of an RCT. Propensity score matching has been used with success as a statistical method to compare two treatments in retrospective studies 18,19, and offers a method of minimizing selection bias and confounding factors. Other statistical methods are available that allow for this correction, such as regression and case control matching. However, propensity score analysis was chosen because, unlike other methods that focus on outcome as the endpoint, propensity scoring analysis builds a model around the treatment option. There are many ways in which variables can be selected for inclusion in the propensity score model, which in turn can affect the validity of the model. Some authors,21 advocate including all variables related to the outcome as well as exposure. In the present study, a stepwise variable selection algorithm was used to develop a good predictive model. The model included diabetes, egfr and total lesion length. These variables were major confounders at baseline and were found on logistic regression analysis to be independent predictors determining whether a patient was more likely to have bypass surgery or endovascular treatment. The Harrell s C statistic of 0 78 for this model was within the range of creating reliable propensity score-matched pairs (over 0 7) for bypass and endovascular treatment, indicating that a statistical analysis between the groups was valid. Finally, the model also corrected for the baseline confounding factors following propensity score matching. The results of this study demonstrated that successful restoration of straight-line flow down to the foot (technical success per limb) was achieved in 92 8 per cent of patients in the endovascular group. A contemporary meta-analysis 3 reported a technical success rate of 92 3 per cent following primary IP angioplasty. Furthermore, a meta-analysis of RCTs 2 reported primary patency, secondary patency and amputation-free survival rates at 1 year similartothosehere(51 4 per cent in the present study versus per cent, 65 8 versus per cent and 74 1 versus per cent respectively). Similarly, following IP bypass the primary and secondary patency rates, and amputation-free survival at 1 year were similar

7 Angioplasty and bypass surgery for critical limb ischaemia in infrapopliteal peripheral artery disease 1821 in the present study to those in other large contemporary series (54 4 versus per cent, 84 4 versus per cent and 78 7 versus 83 per cent) 4,22. The higher patency rate in the surgical bypass group did not lead to a higher rate of limb salvage or improved freedom from reintervention. A recent systematic review 23 of IP revascularization also found better patency, although similar limb salvage rates, in a comparison of infrapopliteal bypass surgery with angioplasty. This is perhaps a reflection of the fact that achieving limb salvage is dependent on a number of factors other than a technically successful revascularization, such as the degree and pattern of tissue loss, patient co-morbidities, and variation in the pathogenicity and treatment of any superimposed wound infection. Life expectancy remains poor for patients with CLI as demonstrated by the mortality rate in this study of 16 0 per cent at 1 year, which is in keeping with the results of a meta-analysis of prospective studies 24 that revealed overall mortality rates of 17 5 percent at 1year and35 per cent at 3 years, regardless of the treatment strategy. In addition the survival rate at 1 year in the present study was significantly higher in the bypass group than in the endovascular group. The mode of revascularization was selected based on MDT discussions rather than a randomized process, and there is a natural tendency to select fitter patients for bypass surgery; this may be reflected in the better long-term survival in this group. Given that limb salvage was the same in both groups, this may explain in part why amputation-free survival was better in the bypass group. Based on the Clavien Dindo classification, the incidence of minor complications (grade I and II) was significantly higher after surgical bypass, whereas the incidence of major complications (grade III and IV) and perioperative mortality was similar in the two groups. A direct comparison of complications following lower limb bypass or endovascular intervention (US Medicare population) 25 found that, although mortality rates were higher following bypass (0 8 versus 0 5 per cent), the overall risk of adverse complications was significantly higher in the endovascular group (8 5 versus 7 7 per cent). These numbers are broadly in agreement with the incidence of complications in the present study in both groups. This highlights that both treatment modalities have associated risks, which should be considered carefully in the decision-making process. The main limitation of this study is the potential for selection bias and confounding. Propensity score matching attempts to reduce the effect of bias and confounding factors by producing two equal cohorts based on preoperative variables. Despite matching for co-morbidities, long-term survival was significantly shorter in the endovascular group. This may reflect a tendency to offer an endovascular approach to high-risk patients who naturally have a shorter life expectancy. The obvious mismatch that remained after propensity score matching was the higher percentage of patients with femoropopliteal disease in the bypass cohort. However, this would bias in favour of poorer results in this group, whereas surgical bypass was shown to have better long-term outcomes than endovascular treatment. The slight differences in duplex ultrasound surveillance intervals may have led to differences in the detection and treatment of restenotic lesions between the two treatment groups. However, it should be noted that overall there was no difference in the freedom from reintervention/target lesion revascularization between surgical bypass and endovascular treatment. Disclosure The authors declare no conflict of interest. References 1 Patel SD, Zymvragoudakis V, Sheehan L, Lea T, Padayachee S, Donati T et al. The efficacy of salvage interventions on threatened distal bypass grafts. JVascSurg 16; 63: Wu R, Yao C, Wang S, Xu X, Wang M, Li Z et al. Percutaneous transluminal angioplasty versus primary stenting in infrapopliteal arterial disease: a meta-analysis of randomized trials. JVascSurg14; 59: Yang X, Lu X, Ye K, Li X, Qin J, Jiang M. Systematic review and meta-analysis of balloon angioplasty versus primary stenting in the infrapopliteal disease. Vasc Surg 14; 48: Slim H, Tiwari A, Ahmed A, Ritter JC, Zayed H, Rashid H. Distal versus ultradistal bypass grafts: amputation-free survival and patency rates in patients with critical leg ischaemia. Eur J Vasc Endovasc Surg 11; 42: Albers M, Romiti M, Pereira CA, Antonini M, Wulkan M. Meta-analysis of allograft bypass grafting to infrapopliteal arteries. Eur J Vasc Endovasc Surg 04; 28: Albers M, Romiti M, Brochado-Neto FC, Pereira CA. Meta-analysis of alternate autologous vein bypass grafts to infrapopliteal arteries. JVascSurg05; 42: Adam DJ, Beard JD, Cleveland T, Bell J, Bradbury AW, Forbes JF et al. versus angioplasty in severe ischaemia of the leg (BASIL): multicentre, randomised controlled trial. Lancet 05; 366: Conte MS, Bandyk DF, Clowes AW, Moneta GL, Seely L, Lorenz TJ et al. Results of PREVENT III: a multicenter, randomized trial of edifoligide for the prevention of vein graft failure in lower extremity bypass surgery. JVascSurg 06; 43: Bradbury AW, Adam DJ, Bell J, Forbes JF, Fowkes FG, Gillespie I et al. versus Angioplasty in Severe

8 1822 S. D. Patel, L. Biasi, I. Paraskevopoulos, J. Silickas, T. Lea, A. Diamantopoulos et al. Ischaemia of the Leg (BASIL) trial: an intention-to-treat analysis of amputation-free and overall survival in patients randomized to a bypass surgery-first or a balloon angioplasty-first revascularization strategy. JVascSurg10; 51(Suppl): 5S 17S. 10 Jaff MR, White CJ, Hiatt WR, Fowkes GR, Dormandy J, Razavi M et al. An update on methods for revascularization and expansion of the TASC lesion classification to include below-the-knee arteries: a supplement to the inter-society consensus for the management of peripheral arterial disease (TASC II). Vasc Med 15; 8: Rooke TW, Hirsch AT, Misra S, Sidawy AN, Beckman JA, Findeiss L et al. Management of patients with peripheral artery disease (compilation of 05 and 11 ACCF/AHA Guideline Recommendations): a report of the American College of Cardiology Foundation/American Heart Association Task Force on Practice Guidelines. J Am Coll Cardiol 13; 61: Norgren L, Hiatt WR, Dormandy JA, Nehler MR, Harris KA, Fowkes FG. Inter-Society consensus for the management of peripheral arterial disease (TASC II). JVasc Surg 07; 45(Suppl S): S5 S Rutherford RB, Baker JD, Ernst C, Johnston KW, Porter JM, Ahn S et al. Recommended standards for reports dealing with lower extremity ischemia: revised version. JVascSurg 1997; 26: Clavien PA, Barkun J, de Oliveira ML, Vauthey JN, Dindo D, Schulick RD et al. The Clavien Dindo classification of surgical complications: five-year experience. Ann Surg 09; 250: European Stroke Organisation, Tendera M, Aboyans V, Bartelink ML, Baumgartner I, Clément D et al.; ESC Committee for Practice Guidelines. ESC Guidelines on the diagnosis and treatment of peripheral artery diseases: document covering atherosclerotic disease of extracranial carotid and vertebral, mesenteric, renal, upper and lower extremity arteries: the Task Force on the Diagnosis and Treatment of Peripheral Artery Diseases of the European Society of Cardiology (ESC). Eur Heart J 11; 32: Farber A, Rosenfield K, Menard M. The BEST-CLI trial: a multidisciplinary effort to assess which therapy is best for patients with critical limb ischemia. Tech Vasc Interv Radiol 14; 17: University of Birmingham Clinical Trial Unit. BASIL-2 Trial. trials/bctu/trials/portfolio-v/basil-2/index.aspx [accessed 30 January 16]. 18 Huang Y, Gloviczki P, Oderich GS, Duncan AA, Kalra M, Fleming MD et al. Outcome after open and endovascular repairs of abdominal aortic aneurysms in matched cohorts using propensity score modeling. JVascSurg15; 62: Schermerhorn ML, O Malley AJ, Jhaveri A, Cotterill P, Pomposelli F, Landon BE. vs. open repair of abdominal aortic aneurysms in the Medicare population. N Engl J Med 08; 358: Brookhart MA, Schneeweiss S, Rothman KJ, Glynn RJ, Avorn J, Sturmer T. Variable selection for propensity score models. Am J Epidemiol 06; 163: Rubin DB, Thomas N. Matching using estimated propensity scores: relating theory to practice. Biometrics 1996; 52: Santo VJ, Dargon P, Azarbal AF, Liem TK, Mitchell EL, Landry GJ et al. Lower extremity autologous vein bypass for critical limb ischemia is not adversely affected by prior endovascular procedure. JVascSurg14; : Schamp KB, Meerwaldt R, Reijnen MM, Geelkerken RH, Zeebregts CJ. The ongoing battle between infrapopliteal angioplasty and bypass surgery for critical limb ischemia. Ann Vasc Surg 12; 26: Rollins KE, Jackson D, Coughlin PA. Meta-analysis of contemporary short- and long-term mortality rates in patients diagnosed with critical leg ischaemia. Br J Surg 13; : Vogel TR, Dombrovskiy VY, Haser PB, Graham AM. Evaluating preventable adverse safety events after elective lower extremity procedures. JVascSurg11; 54: