European Heart Journal (2009) 30, 44 55 doi:10.1093/eurheartj/ehn514 CLINICAL RESEARCH Interventional cardiology and angiology Routine stent implantation vs. percutaneous transluminal angioplasty in femoropopliteal artery disease: a meta-analysis of randomized controlled trials Christos Kasapis 1, Peter K. Henke 2, Stanley J. Chetcuti 1, Gerald C. Koenig 1, John E. Rectenwald 2, Venkataramu N. Krishnamurthy 3, Paul Michael Grossman 1, and Hitinder S. Gurm 1 * 1 Division of Cardiovascular Medicine, University of Michigan Health System, TC B1 226, 2A394, 1500 E. Medical Center Drive, Ann Arbor, MI 48109-5853, USA; 2 Division of Vascular Surgery, University of Michigan, Ann Arbor, MI, USA; and 3 Division of Radiology, University of Michigan, Ann Arbor, MI, USA Received 14 April 2008; revised 20 September 2008; accepted 23 October 2008; online publish-ahead-of-print 21 November 2008 Aims We performed a meta-analysis of randomized controlled trials comparing routine stenting (ST) with percutaneous transluminal angioplasty (PTA) for symptomatic superficial femoral-popliteal artery (SFPA) disease.... Methods Ten trials were pooled randomizing patients to ST (n ¼ 724 limbs) or PTA with provisional stenting (n ¼ 718 limbs) and results with a follow-up period of 9 24 months. The mean lesion length was similar in the two groups (45.8 mm in the ST group and 43.3 mm in the PTA group). We calculated the summary risk ratios (RRs) for immediate technical failure, restenosis, and target vessel revascularization (TVR) using random-effects models. The immediate technical failure was higher in the PTA group than in the ST group [17.1 vs. 5.9%, respectively, RR ¼ 0.28, 95% confidence interval (CI) ¼ 0.15 0.54, P, 0.001], with 10.3% of the PTA patients undergoing stenting because of suboptimal result. There was a trend for lower restenosis in the ST group (37.6% in ST vs. 45.3% in PTA, RR ¼ 0.85, 95% CI ¼ 0.69 1.06, P ¼ 0.146), but no difference in the need for TVR (20% in ST vs. 20.2% in PTA, RR ¼ 0.98, 95% CI ¼ 0.78 1.23, P ¼ 0.89). In an analysis restricted to nitinol stents, there was a trend towards reduction in TVR (RR ¼ 0.79, 95% CI ¼ 0.59 1.06, P ¼ 0.12).... Conclusion Despite the higher immediate success, routine stenting was not associated with a significant reduction in the rate of restenosis or TVR. Our data do not support use of routine stenting as the primary endovascular treatment for short SFPA lesions. ----------------------------------------------------------------------------------------------------------------------------------------------------------- Keywords Peripheral vascular disease Meta-analysis Balloon Angioplasty Stents Introduction Peripheral arterial disease (PAD) affects a large segment of the adult population, with an age-adjusted prevalence of 4 15%, 1 and increasing up to 29% with the presence of cardiovascular risk factors. 2 The mainstay of treatment for PAD includes risk factor modification, exercise program, and antiplatelet therapy with revascularization reserved for patients with significant disability or tissue loss. 3,4 Given the lower risk of procedural complications, endovascular therapy is generally preferred as the first choice for symptomatic patients with PAD. 5 Unlike most other vascular beds, where stenting is the preferred modality of endovascular revascularization, the optimal therapy for superficial femoral-popliteal artery (SFPA) disease remains unknown. Recently published guidelines from the American College of Cardiology/American Heart Association and the Trans-Atlantic Inter-Society Consensus for the management of PAD recommend PTA as the initial preferred option for endovascular treatment of * Corresponding author. Tel: þ1 734 232 4276, Fax: þ1 734 764 4142, Email: hgurm@med.umich.edu Published on behalf of the European Society of Cardiology. All rights reserved. & The Author 2008. For permissions please email: journals.permissions@oxfordjournals.org.
Routine ST vs. PTA in SFPA disease 45 symptomatic SFPA lesions, reserving provisional stenting for salvage therapy after a suboptimal or failed result from balloon dilation. 3,4 However, the evolution of endovascular therapies has inspired a considerable and ongoing debate over the merits of various therapies. Endovascular stenting avoids the problems of early elastic recoil, residual stenosis, and flow-limiting dissection after balloon angioplasty and can thus be used for the treatment of long and calcified lesions. In contrast, SFPA is subject to longitudinal stretching, external compression, torsion, and flexion, which may lead to stent fractures and eventually to restenosis. Although evolution in stent material and design have overcome some of these limitations, the clinical impact remains unclear. A previous meta-analysis, 6 including only one randomized controlled trial in which stenting was compared with PTA 7 and 18 non-comparative studies published between 1993 and 2000, demonstrated similar long-term patency rates for stent implantation and balloon dilation for symptomatic SFPA disease. Subsequent clinical trials comparing routine stent implantation with percutaneous angioplasty for SFPA disease have produced conflicting results. We performed a comprehensive meta-analysis of randomized controlled trials to systematically evaluate currently available data comparing ST vs. PTA in symptomatic SFPA disease. Methods We performed a computerized search to identify relevant articles from 1960 to October 2007 in the MEDLINE, Embase, ISI Web of Knowledge, Current Contents, International Pharmaceutical Abstracts databases and the Cochrane Central Register of Controlled Trials. We combined expanded medical subject headings and keyword searches for peripheral vascular diseases, femoral artery, popliteal artery, femoropopliteal, stents, endovascular, angioplasty, balloon dilatation and peripheral catheterization. The literature search protocol was verified in consultation with a librarian of University of Michigan. Cross-references from the retrieved articles were also explored. In addition, abstract lists from the 2006 and 2007 scientific meetings of the American Heart Association, the American College of Cardiology, the European Society of Cardiology, and the Transcatheter Cardiovascular Therapeutics were searched. Published review articles, editorials, and internet-based sources of information (www.tctmd.com and www.theheart.org) were also reviewed. Finally, we contacted stent-manufacturing companies and requested information regarding unpublished data from randomized controlled trials comparing ST with PTA in SFPA disease. A study was included if it randomized patients with symptomatic SFPA disease to ST or PTA and provided information on at least 6-month outcomes with regard to primary patency and restenosis rate and/or target vessel revascularization (TVR). Patients were considered symptomatic if there was a history of intermittent claudication or critical limb ischaemia, defined as chronic ischaemic pain at rest or the presence of ischaemic ulcers or gangrene. Information was abstracted using a standardized form that included data on the study population demographic and clinical characteristics [mean age; gender; risk factors for cardiovascular disease hypertension, hyperlipidaemia, smoking, diabetes; history of coronary artery disease; Rutherford stage simplified as intermittent claudication vs. critical limb ischaemia; baseline ankle-brachial index (ABI)], angiographic and interventional data (length of lesion; number of occlusions or stenoses; subjects with poor crural runoff vessels defined as one or no patent runoff vessel; type of stent implanted), and outcomes (immediate technical success/failure; restenosis; TVR; mortality; post-procedural ABI; amputations; vascular complications; early thrombo-embolic complications; bleeding/haematomas). Two readers abstracted the data from each article independently using a standard form. There were only minor differences in the extracted data that were resolved in discussion, and in a few cases, we contacted the authors of the relevant study to resolve discrepancies in the interpretation of the reported data. Each trial was evaluated for the adequacy of allocation concealment, performance of the analysis according to the intention-to-treat principle, and blind assessment of the outcomes of interest. We used the Jadad criteria to assess the quality of the trials included in our meta-analysis, with 3 points considered as high quality. 8 As some trials enrolled by limbs treated and could possibly enrol a patient more than once, correction for within-patient clustering was evaluated as an additional quality metric. All the studies were randomized, and reported the results on an intention-to-treat basis. None of the trials was blinded, given the difficulty in blinding operators to the use of a stent. Definitions and endpoints The endpoints of interest were immediate technical success/failure, rate of TVR, and restenosis in 9 24 months of follow-up. Immediate technical success was defined as residual stenosis,30% without flowlimiting dissection, unless otherwise defined by the individual study investigators. From the studies included in the analysis, six used the cutoff of,30% for immediate technical success, 7,9 14 two studies used,50%, 15,16 one used,50% for PTA and,30% for ST, 17 and one study used,20% 18 (in the latter, there was eventually no immediate technical failure). By this definition, cross-over from PTA to bailout stenting was automatically considered as immediate technical failure of the PTA. Binary restenosis was defined as a reduction in the luminal diameter of more than 50% on follow-up conventional angiography or restenosis more than 50%, as determined by follow-up duplex ultrasound peak velocity ratio, except from one study that used the cutoff of 70% of angiographic restenosis. 14 TVR was defined as repeat revascularization of the same SFPA, proximal or distal to or involving the index lesion, or surgical bypass of the SFPA. Statistical analysis From each trial, results were organized into a two-by-two table to permit calculation of effect sizes for PTA in comparison with ST with regard to each outcome. Data on the results were collected on an intention-to-treat basis. Patients who were assigned to a particular therapy but received no therapy or crossed over to the alternate strategy were considered to belong to the group randomized to the original therapy. When the outcome did not occur in either group, we were unable to calculate effect sizes due to the empty cells, and data were excluded from that particular trial. We used fixed-effects and random-effects models to produce across-study summary risk ratios (RRs) with 95% confidence intervals (CIs). As there was significant heterogeneity for some of the endpoints, the random-effects models are preferentially reported. Cochran s Q-test was used to assess heterogeneity and P-values,0.1 were considered significant for heterogeneity. To assess the effect of individual studies on the summary estimate of effect, we did an influence analysis, in which the pooled estimates were recalculated omitting one study at a time. Publication bias was assessed by the funnel plot and by calculating the classic fail safe N and Orwins N. 19,20 An exploratory metaregression was performed to assess the impact of lesion length on possible benefit of stenting. All analyses were performed using
46 C. Kasapis et al. Figure 1 Flow diagram depicting the selection of studies included in the meta-analysis. Comprehensive Meta-Analysis software, version 2.0 (Biostat, Englewood, NJ, USA). Results A total of 656 citations published between 1960 and October 2007 were screened. Our search identified 13 trials that randomized patients with symptomatic SFPA disease to primary PTA vs. ST (Figure 1). Of these, 10 had been published in peer-reviewed journals, 7,9 14,17,18 two had been published only as an FDA approval letter available at the FDA website, 15,16 and one has been released in the 2007 TCT meeting with information on limited endpoints. 21 Two studies published in German journals 22,23 and one study in English literature 14 involved the same authors and overlapping enrolment periods. One of the authors, involved in all three studies, was contacted and reported potential overlap in the studied populations and suggested use of the English study 14 in the meta-analysis. In addition, we excluded the RELISIENT trial, which was released in the 2007 TCT meeting, because its endpoint definitions varied dramatically from the prior trials. 21 We attempted to obtain outcome data congruent with prior definitions from the study investigators but were unsuccessful. A total of 1343 patients and 1442 limbs from 10 randomized controlled trials constituted our final study population. Five of the trials were of high quality by the Jadad criteria (score 3). 8 Two studies reported the results per limb treated with a small number of patients undergoing PVI in both limbs, but no adjustment was made for possible interdependencies within patients. 14,15 The Jadad scores for each study, the baseline demographic characteristics of the patient population, and the indications for the procedure and the pre-procedural ABI are listed in Table 1. Table 2 summarizes the angiographic and interventional characteristics, including the type of stent, the mean length of the lesion, the percentage of lesions with occlusion vs. stenosis, and the state of runoff vessels (poor or good). Three trials used nitinol stents, 11,12,15,17 four used Palmaz stainless steel stents, 7,13,14,18 one used tantalum stents, 10 and two used nitinol-covered stent grafts. 9,16 There were no statistical differences in the baseline demographic, angiographic, and interventional characteristics, with the exception of a number of occlusions treated with stenting. Patients in all studies received aspirin with dose ranging from
Table 1 Demographic and clinical characteristics Study (year) Krankenberg (2007) Schillinger (2006) Viabahn (2005) Becquemin (2003) Saxon (2003) No. of patients (limbs) 244 104 197 227 28 Study quality a (Jadad score) High (3) Low (3)... Low (2) High (3)... Low (2)... ST PTA ST PTA ST PTA ST PTA ST PTA n ¼ limbs 123 121 51 53 97 100 115 112 15 13 Age 67 + 9 66 + 10 65 + 10 68 + 10 67.2 + 9.7 66.9 + 9.5 66.5 + 11.1 66.4 + 11.7 68.1 72.6 Male sex 77 (62.6%) 91 (75.2%) 30 (59%) 25 (47%) 80 (82.5%) 70 (70%) 76 (66%) 66 (59%) N/A N/A HTN 102 (82.9%) 100 (82.6%) 48 (94%) 47 (89%) N/A N/A 61 (53%) 57 (51%) 8 (53.3%) 12 (92.3%) HL 74 (60.2%) 74 (61.2%) 47 (92%) 46 (87%) N/A N/A 43 (38%) 48 (44%) 9 (60%) 9 (69.2%) DM 44 (35.8%) 37 (30.6%) 22 (43%) 17 (32%) 36 (37.1%) 34 (34%) 11 (10%) 16 (14%) 4 (26.7%) 5 (38.5%) Smoking 84 (68.3%) 88 (72.7%) 27 (53%) 19 (36%) 45 (46.4%) 51 (51%) 79 (59%) 69 (60%) 7 (46.7%) 4 (30.8%) CAD 52 (42.3%) 38 (31.4%) 34 (67%) 40 (75%) 23 (23.7%) 30 (30%) 30 (26%) 29 (26%) 5 (33.3%) 8 (61.5%) Claudication 115 (96.6%) 109 (95.6%) 45 (88%) 46 (87%) N/A N/A 91 (79%) 89 (79%) N/A N/A Critical ischaemia 3 (2.5%) 4 (3.5%) 6 (12%) 7 (13%) N/A N/A 24 (20.9%) 23 (20.5%) N/A N/A ABI 0.68 + 0.16 0.72 + 0.15 0.57 + 0.19 0.54 + 0.20 0.74 + 0.17 0.67 + 0.18 0.74 + 0.20 0.72 + 0.22 0.71 (0.65 0.77) 0.71 (0.66 0.76) Study (year) Intracoil (2001) Cejna (2001) Grimm (2001) Zdanowski (1999) Vroegindeweij (1997) No. of patients (limbs) 266 (352) 141 (154) 53 32 51 Study quality a (Jadad score) Low (2) High (3)... High (3) Low (2)... High (3)... ST PTA ST PTA ST PTA ST PTA ST PTA n ¼ limbs 177 175 77 77 30 23 15 17 24 27 Age 66.8 + 10.6 68.1 + 10.2 68.6 (39.2 87) 65.5 (39.2 83) 70.5 + 9.8 68.1 + 8.4 72 (62 80) 71 (41 86) 65 (46 78) 64 (41 82) Male sex 87 (64.4%) 83 (63.4%) 49 (63.6%) 46 (59.8%) 22 (73.3%) 10 (43.5%) 10 (66.7%) 4 (23.5%) 17 (70.8%) 19 (70.3%) HTN N/A N/A 31 (36.4%) 36 (44.2%) N/A N/A 4 (26.7%) 4 (23.5%) 3 (12.5%) 6 (22.2%) HL N/A N/A 32 (35%) 37 (46.8%) N/A N/A N/A N/A 9 (37.5%) 7 (25.9%) DM 49 (36.2%) 49 (37.4%) 32 (39%) 31 (40.2%) N/A N/A 5 (33%) 5 (29.4%) 3 (12.5%) 3 (11.1%) Smoking 104 (77%) 104 (79.3%) 49 (62.4%) 43 (61%) N/A N/A 5 (33.3%) 6 (35.3%) 14 (58.3%) 18 (66.7%) CAD 48 (35.5%) 37 (28.2%) N/A N/A N/A N/A N/A N/A 6 (25%) 9 (33.3%) Claudication N/A N/A 50 (64.9%) 58 (75.3%) 30 (100%) 23 (100%) 2 (13.3%) 3 (17.6%) 27 (100%) 24 (100%) Critical ischaemia N/A N/A 27 (35%) 19 (24.6%) 0 0 13 (86.7%) 14 (82.4%) 0 0 ABI N/A N/A 0.63 + 0.20 0.62 + 0.22 0.47 + 0.36 0.62 + 0.3 0.48 (0.13 0.79) 0.42 (0.49 0.65) N/A N/A ST, primary stenting; PTA, percutaneous transluminal angioplasty; HTN, hypertension; HL, hyperlipidaemia; DM, diabetes mellitus; CAD, coronary artery disease; ABI, ankle-brachial index; N/A, not available. a The Jadad score was used to evaluate the quality of the trials (3 points was considered high quality). 8 Routine ST vs. PTA in SFPA disease 47
48 Table 2 Angiographic and interventional characteristics Study (year) Krankenberg (2007) Schillinger (2006) Viabahn (2005) Becquemin (2003) Saxon (2003) Type of stent Nitinol (Bard Nitinol (Dynalink or Viabahn stent graft Palmaz Hemobahn stent graft Luminexx 3) Absolute) Stent diameter (mm) 5 7... 6... N/A... Various 6 7... ST PTA ST PTA ST PTA ST PTA ST PTA n ¼ limbs 123 121 51 53 97 100 115 112 15 13 Length of lesion (mm) 45.2 + 27.9 44.5 + 28 101 + 75 92 + 64 73 + 36 67 + 37 25.4 + 18 25.1 + 17.8 74.1 (56.3 91.9) 63.2 (44.4 82.0) Occlusion no. of patients 45 (36.6%) 30 (24.8%) 19 (37.2%) 17 (32.1%) 20 (20.6%) 29 (29%) 29 (25.2%) 21 (18.8%) 3 (20%) 0 Poor runoff 49 (39.8%) 44 (36.4%) 7 (13.7%) 12 (22.6%) N/A N/A N/A N/A N/A 3 (11.1%) Good runoff 74 (60.2%) 77 (63.6%) 44 (86.3%) 41 (77.4%) N/A N/A N/A N/A N/A N/A Study (year) Intracoil (2001) Cejna (2001) Grimm (2001) Zdanowski (1999) Vroegindeweij (1997) Type of stent Nitinol (Intracoil) Palmaz Palmaz Tantalum (Strecker) Palmaz Stent diameter (mm) N/A... 4 6... 5 6... 6 N/A... ST PTA ST PTA ST PTA ST PTA ST PTA n ¼ limbs 177 175 77 77 30 23 15 17 24 27 Length of lesion (mm) 35.6 + 30 32.6 + 29.6 25.6 + 1.6 22.3 + 1.4 28 + 15 31 + 25 74 (35 140) 71 (20 200) 0 20 in 16 patients and 20 50 in eight patients Occlusion no. of patients 40 (22.6%) 31 (17.7%) 35 (45.4%) 25 (32.5%) 13 (43.3%) 3 (13%) 15 (100%) 17 (100%) 4 (16.7%) 5 (18.5%) Poor runoff N/A N/A N/A 21 (27.3%) 15 (19.5%) 0 0 7 (46.7%) 5 (29.4%) 2 (8.3%) Good runoff N/A N/A 56 (72.7%) 62 (80.5%) 30 (100%) 23 (100%) 8 (53.3%) 12 (70.6%) 22 (91.7%) 24 (88.9%) 0 20 in 18 patients and 20 50 in nine patients ST, primary stenting; PTA, percutaneous transluminal angioplasty; N/A, not available. C. Kasapis et al.
Routine ST vs. PTA in SFPA disease 49 80 to 325 mg daily indefinitely. In addition, patients received clopidogrel 75 mg daily for 1 3 months in three studies 9,11,12,17 and ticlopidine 250 mg twice daily for 1 month in one study. 15 The raw endpoint rates for each trial with relative follow-up periods and patient attrition are listed in Table 3. There was significant heterogeneity with regard to immediate technical failure and restenosis rates and no heterogeneity with regard to the endpoint of TVR. Thus, the random-effects models are discussed preferentially, although fixed-effects models provided similar estimates (Table 4). From a total of 1442 limbs/lesions included in the analysis, 724 were assigned to ST and 718 to PTA. The follow-up period ranged between 9 and 24 months. The mean lesion length was similar in the two groups (45.8 mm in the ST group and 43.3 mm in the PTA group). The immediate technical failure rate was significantly higher in the PTA group when compared with the ST group (17.1 vs. 5.9%, respectively, RR ¼ 0.28, 95% CI ¼ 0.15 0.54, P, 0.001) (Figure 2). In two studies, the outcome of immediate failure did not occur in both treatment arms, and data were excluded from the analysis as we were unable to calculate effect sizes. 10,18 From the PTA group, 10.3% of the patients crossed over to secondary stenting, mainly because of suboptimal PTA result defined as flow-limiting dissection or residual stenosis of.30 50%. There was a trend for lower rate of restenosis during the 9 24 month follow-up period in the ST group when compared with the PTA group, i.e. 37.6% in the ST vs. 45.3% in the PTA group (RR ¼ 0.85, 95% CI ¼ 0.69 1.06, P ¼ 0.146) (Figure 3). Four studies used angiographic follow-up, 10,13 15 four studies duplex ultrasound follow-up, 7,9,16,17 and two studies used both 11,12,18 to determine restenosis. The rate of TVR within 9 24 months was similar between the two groups (20% in the ST vs. 20.2% in the PTA group, RR ¼ 0.98, 95% CI ¼ 0.78 1.23, P ¼ 0.89) (Figure 4). One study did not report TVR rates and was not included in the analysis. 7 With regard to other outcomes, mortality (4.35% in the ST vs. 4.6% in the PTA group) and amputation rate (1.76% in the ST vs. 1.54% in the PTA group) were similar in the two treatment arms during the follow-up period. Overall, vascular complications (6.75% in the ST vs. 4.76% in the PTA group), early thrombo-embolic events (3.67% in the ST vs. 2.37% in the PTA group), and bleeding complications (2.25% in the ST vs. 1.4% in the PTA group) were lower in both groups. Only two studies reported information on stent fractures, with frequencies of 2 and 12% at 12 months, respectively. 11,17 Interestingly, in the study with the highest frequency of stent fractures, the authors noted that the binary restenosis rate in patients with stent fracture was not statistically different from those without stent fracture. 17 To assess the effect of individual studies on the summary estimate, we performed a sensitivity analysis, in which the pooled estimates for the endpoint of TVR were recalculated, omitting one study at a time, but this did not alter the results (data not shown). We also performed a subgroup analysis based on different stent types, i.e. nitinol, Palmaz, and stent grafts. The use of nitinol stents was associated with a significant higher immediate technical success (RR ¼ 1.19, CI ¼ 1.03 1.39, P ¼ 0.02) and a nonsignificant trend for lower TVR (RR ¼ 0.79, CI ¼ 0.59 1.06, P ¼ 0.12), compared with angioplasty, whereas there was no significant difference in the rate of restenosis (Figure 5). Conversely, the older Palmaz stents resulted in a trend for higher TVR, although not significant (RR ¼ 1.46, CI ¼ 0.99 2.16, P ¼ 0.056). The effect of the use of stent grafts on TVR was neutral. The exclusion of the stent grafts from the analysis produced similar results, with significantly higher immediate technical failure rate in the PTA group compared with the ST group (RR ¼ 0.24, 95% CI ¼ 0.10 0.57, P ¼ 0.001) and no significant differences in the rates of restenosis (RR ¼ 0.93, 95% CI ¼ 0.79 1.10, P ¼ 0.36) and TVR (RR ¼ 1.02, 95% CI ¼ 0.69 1.52, P ¼ 0.92). Furthermore, to assess the effect of the lesion length on TVR, an exploratory meta-regression was performed and suggested that stenting may be beneficial in longer lesions (b ¼ 20.009, 95% CI 20.017 to 20.001, P ¼ 0.026). Publication bias A funnel plot to assess publication bias was symmetric, suggesting lack of publication bias (Figure 6). As the combined effect estimate for TVR was not statistically significant, the classic fail safe N (which addresses the concern that the observed significance may be spurious) was not relevant. The Orwin fail safe N, which differs from the classic fail safe N in that the mean RR in the new missing studies can be a value other than nil and that the criterion value is an effect size rather than a P-value, was therefore calculated. 19,20 The Orwin fail safe N for our meta-analysis was 9, suggesting that nine studies with a mean RR of 0.5 will be required to achieve a pooled RR for TVR under 0.7. Discussion The key finding of our analysis is that there is no significant difference in the rate of TVR between PTA with provisional stenting and routine stenting for symptomatic patients with short SFPA lesions, although there is a trend for lower restenosis and a significant higher immediate technical success in favour of routine stenting. The results of our meta-analysis corroborate and extend current guidelines supporting PTA with provisional stenting as the primary endovascular treatment for symptomatic SFPA disease 3,4 and are in agreement with the previous meta-analysis that yielded similar long-term patency rates for PTA and stent implantation. 6 TVR, arguably, represents a more robust endpoint than restenosis by itself, as it is a decision driven by both the clinical status and by the angiographic or Doppler evidence of restenosis. On the basis of the similar TVR rates in the two treatment arms, our analysis suggests that balloon angioplasty with provisional stenting is equivalent to routine stenting. In addition, the use of provisional stenting as a bailout in the case of suboptimal PTA result, which was noted in 10.3% in our pooled data, can almost offset the 11.2% difference in immediate technical success rate in favour of routine stenting. Thus, PTA with provisional stenting remains a reasonable and simple first option in the endovascular treatment of SFPA. This contention becomes even more justified, if the additional cost of stenting over PTA is taken into account. A previous costeffective analysis of a trial that randomized patients with SFPA disease to treatment with a self-expanding nitinol stent vs. PTA
50 Table 3 Primary and secondary outcomes Study (Year) Krankenberg (2007) Schillinger (2006) Viabahn (2005) Becquemin (2003) Saxon (2003) Follow-up (months) 12 24 12 12 24............. ST PTA ST PTA ST PTA ST PTA ST PTA Cross over 0/123 13/121 (11%) 0/51 17/53 (32%) 0/97 0/100 0/115 15/112 (13%) 0/15 1/13 (7.7%) Lost to clinical follow-up 9 (7.3%) 6 (4.9%) 5 (9.8%) 1 (1.9%) N/A N/A 0 0 0 0 Immediate technical success 117/123 (95%) 96/121 (79%) 50/51 (98%) 36/53 (67.9%) 91/97 (94%) 84/100 (84%) 111/115 (96.5%) 95/112 (84.8%) 15/15 (100%) 12/13 (92%) Restenosis 32/101 (31.7%) 39/101 (38.6%) 21/46 (45.7%) 36/52 (69.2%) 34/97 (35.1%) 60/100 (60%) 26/75 (34%) 21/65 (32%) 2/15 (13.3%) 10/13 (69.2%) TVR 17/114 (14.9%) 21/115 (18.3%) 17/46 (37%) 28/52 (53.8%) 19/97 (19.6%) 21/100 (21%) 14/115 (12.2%) 9/112 (8%) 2/15 (13.3%) 2/13 (15.4%) Mortality 4/123 (3.2%) 1/121 (0.8%) 1/51 (2%) 0/53 2/97 (2%) 3/100 (3%) 8/115 (7%) 14/112 (12.5%) 0/15 0/13 ABI post-procedure 0.89 + 0.16 0.87 + 0.15 0.88 + 0.18 0.78 + 0.17 0.93 0.86 0.96 + 0.16 0.96 + 0.17 N/A N/A Amputation 2/123 (1.8%) 0/121 0/51 1/53 (1.9%) 1/97 (1%) 2/100 (2%) 6/112 (5.4%) 2/115 (1.7%) 0/15 0/13 Vascular Complications 6/123 (4.9%) 4/121 (3.3%) 1/51 (2%) 1/53 (1.9%) 2/97 (2%) 1/100 (1%) 11/115 (9.6%) 9/112 (8%) 3/15 (20%) 0/13 Early thrombo-embolic complications 0/123 1/121 (0.8%) 1/51 (2%) 1/53 (1.9%) 4/97 (4.1%) 1/100 (1%) 7/115 (6.1%) 4/112 (3.6%) 3/15 (20%) 0/13 Bleeding/haematomas 3/123 (2.4%) 0/121 N/A N/A 1/97 (1%) 1/100 (1%) 4/115 (3.5%) 1/112 (0.9%) 0/15 0/13 Study (Year) Intracoil (2001) Cejna (2001) Grimm (2001) Zdanowski (1999) Vroegindeweij (1997) Follow-up (months) 9 24 24 12 12............. ST PTA ST PTA ST PTA ST PTA ST PTA Cross over N/A N/A 0/77 10/77 (13%) 0/30 0/23 0/15 0/17 4/24 (8%) 0/27 Lost to clinical follow-up 31 (17.5%) 34 (19.4%) N/A N/A 0 0 0 0 0 0 Immediate technical success 152/177 (85.9%) 143/175 (81.7%) 76/77 (98.7%) 65/77 (84.4%) 30/30 (100%) 23/23 (100%) 15/15 (100%) 17/17 (100%) 24/24 (100%) 24/27 (88.9%) Restenosis 40/97 (41.2%) 31/92 (33.7%) 26/56 (46.4%) 26/55 (47.2%) 8/30 (26.7%) 5/23 (21.7%) 10/12 (83.3%) 8/8 (100%) 9/24 (37.5%) 7/27 (25.9%) TVR 24/146 (16.4%) 25/141 (17.7%) 28/77 (36.4%) 16/77 (20.8%) 8/30 (26.7%) 7/23 (30.4%) 2/15 (13.3%) 2/17 (11.8%) N/A N/A Mortality 1/135 (0.7%) 4/131 (3.1%) 12/77 (15.6%) 7/77 (9.1%) 0/30 0/23 N/A N/A N/A N/A ABI post-procedure N/A N/A 0.99 + 0.18 0.97 + 0.20 0.91 + 0.19 0.85 + 0.2 N/A N/A 0.78 + 0.18 0.81 + 0.18 Amputation 0/135 1/131 (0.8%) 2/77 (2.6%) 4/77 (5.2%) 0/30 0/23 0/15 0/17 N/A N/A Vascular Complications 5/135 (3.7%) 6/131 (4.6%) 7/77 (9.1%) 6/77 (7.8%) N/A N/A 1/15 (6.7%) 3/17 (17.6%) 8/24 (33.3%) 1/27 (3.7%) Early thrombo-embolic complications 1/135 (0.7%) 4/131 (3.1%) 7/77 (9.1%) 4/77 (5.2%) 0/30 0/23 1/15 (6.7%) 0/17 1/24 (4.2%) 1/27 (3.7%) Bleeding/haematomas 1/135 (0.7%) 1/131 (0.8%) 3/77 (3.9%) 2/77 (2.6%) N/A N/A 1/15 (6.7%) 3/17 (17.6%) N/A N/A ST, primary stenting; PTA, percutaneous transluminal angioplasty; TVR, target vessel revascularization; N/A, not available. C. Kasapis et al.
Routine ST vs. PTA in SFPA disease 51 Table 4 Summary risk ratios and heterogeneity for the primary endpoints by fixed- and random-effects models Endpoint Number of studies Fixed-effects model. Random-effects model P-value for heterogeneity Risk Ratio 95% CI Risk ratio 95% CI... Immediate failure 8 0.43 0.30 0.60 0.28 0.15 0.54 0.027 Restenosis 10 0.83 0.73 0.95 0.85 0.69 1.06 0.019 TVR 9 0.98 0.79 1.21 0.98 0.78 1.23 0.366 TVR, target vessel revascularization; CI, confidence interval. Figure 2 The Forest plot of risk ratios of immediate technical failure using random-effects model. Sizes of data markers are proportional to the weight of each study in the meta-analysis. Horizontal bars, 95% confidence interval. Figure 3 The Forest plot of risk ratios of restenosis using random-effects model. Sizes of data markers are proportional to the weight of each study in the meta-analysis. Horizontal bars, 95% confidence interval. demonstrated that the use of routine stenting increased the procedure duration, equipment costs, and physician services, resulting in initial hospital costs of $3500/patient higher for patients randomized to the nitinol stent compared with PTA. The authors of that study concluded that a strategy of routine stent implantation for patients with SFPA disease is not optimal on economic grounds and that PTA with provisional stenting should be preferred. 24 These findings are likely still extant, as there has been no substantial decrease in equipment cost. Another important consideration with the use of stents in SFPA disease is the occurrence of stent fractures that have been associated with in-stent restenosis. As mentioned earlier, information on stent fractures was reported in only two studies in our meta-analysis, with frequencies of 2 and 12% at 12 months,
52 C. Kasapis et al. Figure 4 The Forest plot of risk ratios of target vessel revascularization (TVR) using random-effects model. Sizes of data markers are proportional to the weight of each study in the meta-analysis. Horizontal bars, 95% confidence interval. One study was excluded from the analysis, as there were no data on TVR. Figure 5 Subgroup analysis for the use of nitinol stents. The Forest plots of risk ratios of immediate technical success (A), restenosis (B), and TVR (C) using random-effects model. Sizes of data markers are proportional to the weight of each study in the meta-analysis. Horizontal bars, 95% confidence interval.
Routine ST vs. PTA in SFPA disease 53 Figure 6 Symmetric funnel plot of standard error by log risk ratio, suggesting lack of publication bias. respectively. 11,17 Although in the study with the highest frequency of stent fractures, the binary restenosis rate was similar in fractured and non-fractured stents; 17 an earlier trial investigating the occurrence and the clinical impact of stent fractures after femoropopliteal stenting with nitinol stents suggested a considerable risk of stent fractures (24.5%), especially following long segment femoral artery stenting. Stent fractures in this study were associated with a higher in-stent restenosis and re-occlusion rate after a mean follow-up of 11 months. 25 In contrast, our meta-regression analysis suggested that stenting may be beneficial in longer lesions. Although this finding must be considered as hypothesis-generating, the possibility of alpha error cannot be excluded, given the relatively small number of studies included in the meta-regression. Based on the discrepancy of these results and the improvement of stent designs over the years, more data are required to elucidate whether factors such as the length of the stented segment and the stent design and stent surface are more likely to affect stent patency. Our subgroup analysis suggested that use of nitinol stents was associated with a higher immediate technical success rate and a non-significant trend for lower TVR over PTA, whereas the use of the older generation Palmaz stents resulted in a non-significant trend for higher TVR over PTA. This finding supports the concept that the improvement in nitinol stent design, including improvement in radial strength, the ability to recover from being crushed, and reduced foreshortening, has led to better anatomical and clinical outcomes than the older stainless steel stents. This finding is in agreement with previous studies showing superior patency outcomes with the new generation nitinol stents compared with stainless steel stents. 26,27 However, this non-significant trend for lower TVR with nitinol stents over PTA would still not validate universal routine stenting as the primary endovascular treatment for SFPA disease. Future adequately powered randomized controlled trials would be required to provide more evidence to answer this important question. Based on the assumption that nitinol stents decrease TVR over PTA to the same degree as in our analysis, (RR ¼ 0.79, 95% CI ¼ 0.59 1.06, P ¼ 0.129), a trial comparing nitinol stents with PTA would need to enrol 2400 patients to have adequate statistical power to demonstrate potential advantage of nitinol stents over PTA. Such a trial would be difficult, although possible, to be performed, given cost and logistics involved. Furthermore, given the equivalent efficacy and the lower cost of PTA as the primary endovascular treatment of SFPA disease, stents have to demonstrate significantly enhanced patency rates before they can be considered as the preferred tool for endovascular therapy. Moreover, in view of the consistently favourable results of PTA demonstrating equivalent outcomes even when compared with bypass surgery, 28 we need to move away from historical control studies as the evidence base for the evaluation of therapeutic strategies and focus on randomized controlled trials. Currently, there are four ongoing randomized trials in the recruitment phase comparing different types of stents with PTA, as well as one study comparing the efficacy of Viabahn endoprosthesis with nitinol stents in the treatment of SFPA disease (Table 5). Although the target population to be enrolled in these studies ranges from 120 to 480 patients, far below the above-estimated sample size, the completion of these studies will provide further insight and evidence on this continuing debate for the optimal endovascular treatment of symptomatic SFPA disease. Indeed, in order for stents to emerge as the preferred endovascular device, major design evolution that would dramatically reduce TVR is required. Clinicians and regulatory authorities also need to be vigilant against commercial sponsors introducing arbitrary endpoint definitions to suggest possibility of benefit with their device. Such a variation in endpoint was noted in the recently presented RESILIENT trial, in which the primary endpoint of TVR included the occurrence of bailout stenting, which was advised for major flow-limiting dissection or residual stenosis.30%. 21 This deviation from the previously used definitions made it inevitable that the stenting arm would appear superior, although there is no clear evidence that there is any clinical superiority of the stenting system that was evaluated. Limitations Our analysis encompasses trials that have wide variation in endovascular technology, and their relevance to contemporary practice may be limited. With the advances in stent design and the newer generation of different types of nitinol stents, one should be careful in interpreting the results of pooled studies spanning a 10-year period. We tried to adjust for this heterogeneity over time by performing a sensitivity and subgroup analysis based on the stent type to account for the advancement in the stent design. Furthermore, there was statistical heterogeneity for some of the outcome measures; however, there was no evidence of heterogeneity with regard to the outcome of TVR. In addition, all the studies included in our analysis were not blinded, given the impossibility of blinding the operator to the use of stent vs. PTA, and two of the studies have not been published in peer-reviewed journals with the data obtained from the FDA approval letter. However, there is no reason to doubt the accuracy of the data that were submitted to the FDA. Furthermore, our meta-analysis is inherent to the major limitations of all meta-analyses, which include publication bias (although tested non-significant in our study) and the difficulties in comparing the results because of the different study populations, study designs and reporting methods, and the absence of individual patient data, which prohibit adjustment for confounding factors. 29 Finally, almost half of the trials were considered to be of moderate-to-poor quality and
54 Table 5 Ongoing randomized controlled trials of balloon angioplasty and/or stents in symptomatic SFA disease Study SUPER UK SIT-UP DURAVEST ZILVER PTX GORE SIROCCO SFA SUPER SL PACCOCATH-FEM I Name VIABAHN Status Recruiting Recruiting Recruiting Recruiting Recruiting Active, not recruiting Active, not Active, not recruiting recruiting Phase IV IV IV I IV II IV I-II NCT ID 00232843 00309595 00289055 00120406 00228384 00232869 00235131 00472472 Treatment arm I Treatment arm II Type of study Cordis SMART nitinol stent Balloon angioplasty multicenter Cordis SMART nitinol stent Balloon angioplasty multicentre Cordis SMART nitinol stent ZILVER PTX drug-eluting stent VIABAHN endoprosthesis Sirolimus-coated Cordis SMART nitinol stent Cordis SMART nitinol stent Balloon angioplasty Balloon angioplasty Bare nitinol stent SMART bare metal stent Bard Luminexx stent multicentre multicentre multicentre double-blinded, multicentre multicentre Paclitaxel-coated balloon angioplasty Plain balloon angioplasty double-blinded, multicentre Size (n) 150 120 120 480 150 90 200 79 Inclusion Symptomatic SFA disease (5 14.5 cm) Symptomatic SFA disease (5 22 cm) Symptomatic SFA disease (5 14.5 cm) Symptomatic SFA disease (,7 cmin phase I and,14 cm in phase II) Symptomatic SFA disease (.8 cm) Symptomatic SFA disease Symptomatic SFA disease (5 22 cm) Symptomatic SFA disease Primary endpoint Primary patency by duplex ultrasound at 1 year Primary patency by duplex ultrasound at 1 year Primary patency by duplex ultrasound at 1 year Primary patency by duplex ultrasound at 1 year Primary patency at 3 years In-stent stenosis via quantitative angiography at 6 months Primary patency by duplex ultrasound at 1 year Start date March 2005 December 2005 November 2005 March 2005 September 2005 February 2001 May 2005 April 2004 Completion August 2008 April 2008 December 2008 N/A December 2010 December 2008 March 2008 June 2007 date Angiographic lumen loss at 6 months NCT ID, ClinicalTrials.gov Identifier; SFA, superficial femoral artery; N/A, not available. There are also five ongoing non-randomized trials evaluating safety and efficacy of different types of stents (NCT ID#00180505, n ¼ 120; NCT ID#00475566, n ¼ 100; NCT ID#00530712, n ¼ 287; NCT ID#00496041, n ¼ 200; NCT ID#00542646, n ¼ 45). Data are obtained from http://www.clinicaltrials.gov. C. Kasapis et al.
Routine ST vs. PTA in SFPA disease 55 there was no adjustment for within-patient clustering. However, these limitations are, in general, more likely to lead to underestimation of standard errors and spuriously low P-values and should not impact the null results of our meta-analysis. Thus, our results invoke the need for a large adequately powered, high quality trial to test the utility of primary stenting in SFA disease. Conclusion Currently available data suggest no significant difference in the rate of TVR between PTA with provisional stenting and routine stenting for symptomatic patients with short SFPA lesions, although there is a trend for lower restenosis and a significant higher immediate technical success rate in favour of routine stenting as well as a nonsignificant trend towards lower TVR with the newer generation nitinol stents. However, several limitations prevent wider generalizability of the results, and there is no definitive evidence favouring one strategy over another. Given the absence of data to the contrary and lower equipment cost, PTA with provisional stenting should remain the preferred endovascular therapy for patients with SFPA disease. Conflict of interest: none declared. References 1. Selvin E, Erlinger TP. Prevalence of and risk factors for peripheral arterial disease in the United States: results from the National Health and Nutrition Examination Survey, 1999 2000. Circulation 2004;110:738 743. 2. Hirsch AT, Criqui MH, Treat-Jacobson D, Regensteiner JG, Creager MA, Olin JW, Krook SH, Hunninghake DB, Comerota AJ, Walsh ME, McDermott MM, Hiatt WR. Peripheral arterial disease detection, awareness, and treatment in primary care. JAMA 2001;286:1317 1324. 3. Norgren L, Hiatt WR, Dormandy JA, Nehler MR, Harris KA, Fowkes FG, Bell K, Caporusso J, Durand-Zaleski I, Komori K, Lammer J, Liapis C, Novo S, Razavi M, Robbs J, Schaper N, Shigematsu H, Sapoval M, White C, White J. Inter-Society Consensus for the Management of Peripheral Arterial Disease (TASC II). Eur J Vasc Endovasc Surg 2007;33(Suppl. 1):S1 S75. 4. Hirsch AT, Haskal ZJ, Hertzer NR, Bakal CW, Creager MA, Halperin JL, Hiratzka LF, Murphy WR, Olin JW, Puschett JB, Rosenfield KA, Sacks D, Stanley JC, Taylor LM Jr, White CJ, White J, White RA, Antman EM, Smith SC Jr, Adams CD, Anderson JL, Faxon DP, Fuster V, Gibbons RJ, Hunt SA, Jacobs AK, Nishimura R, Ornato JP, Page RL, Riegel B. ACC/AHA 2005 Practice Guidelines for the management of patients with peripheral arterial disease (lower extremity, renal, mesenteric, and abdominal aortic): a collaborative report from the American Association for Vascular Surgery/Society for Vascular Surgery, Society for Cardiovascular Angiography and Interventions, Society for Vascular Medicine and Biology, Society of Interventional Radiology, and the ACC/AHA Task Force on Practice Guidelines (Writing Committee to Develop Guidelines for the Management of Patients With Peripheral Arterial Disease): endorsed by the American Association of Cardiovascular and Pulmonary Rehabilitation; National Heart, Lung, and Blood Institute; Society for Vascular Nursing; Trans- Atlantic Inter-Society Consensus; and Vascular Disease Foundation. Circulation 2006;113:e463 e654. 5. Hunink MG, Wong JB, Donaldson MC, Meyerovitz MF, de Vries J, Harrington DP. Revascularization for femoropopliteal disease. A decision and cost-effectiveness analysis. JAMA 1995;274:165 171. 6. Muradin GS, Bosch JL, Stijnen T, Hunink MG. Balloon dilation and stent implantation for treatment of femoropopliteal arterial disease: meta-analysis. Radiology 2001;221:137 145. 7. Vroegindeweij D, Vos LD, Tielbeek AV, Buth J, van de Bosch HC. Balloon angioplasty combined with primary stenting versus balloon angioplasty alone in femoropopliteal obstructions: a comparative randomized study. Cardiovasc Intervent Radiol 1997;20:420 425. 8. Jadad AR, Moore RA, Carroll D, Jenkinson C, Reynolds DJ, Gavaghan DJ, McQuay HJ. Assessing the quality of reports of randomized clinical trials: is blinding necessary? Control Clin Trials 1996;17:1 12. 9. Saxon RR, Coffman JM, Gooding JM, Natuzzi E, Ponec DJ. Long-term results of eptfe stent-graft versus angioplasty in the femoropopliteal artery: single center experience from a randomized trial. J Vasc Interv Radiol 2003;14: 303 311. 10. Zdanowski Z, Albrechtsson U, Lundin A, Jonung T, Ribbe E, Thorne J, Norgren L. Percutaneous transluminal angioplasty with or without stenting for femoropopliteal occlusions? A randomized controlled study. Int Angiol 1999;18:251 255. 11. Schillinger M, Sabeti S, Loewe C, Dick P, Amighi J, Mlekusch W, Schlager O, Cejna M, Lammer J, Minar E. Balloon angioplasty versus implantation of nitinol stents in the superficial femoral artery. N Engl J Med 2006;354:1879 1888. 12. Schillinger M, Sabeti S, Dick P, Amighi J, Mlekusch W, Schlager O, Loewe C, Cejna M, Lammer J, Minar E. Sustained benefit at 2 years of primary femoropopliteal stenting compared with balloon angioplasty with optional stenting. Circulation 2007;115:2745 2749. 13. Becquemin JP, Favre JP, Marzelle J, Nemoz C, Corsin C, Leizorovicz A. Systematic versus selective stent placement after superficial femoral artery balloon angioplasty: a multicenter prospective randomized study. J Vasc Surg 2003;37:487 494. 14. Cejna M, Thurnher S, Illiasch H, Horvath W, Waldenberger P, Hornik K, Lammer J. PTA versus Palmaz stent placement in femoropopliteal artery obstructions: a multicenter prospective randomized study. J Vasc Interv Radiol 2001;12: 23 31. 15. US Food and Drug Administration. Center for Devices and Radiological Health. Intracoil self-expanding peripheral stent P000033 (2001). http://www.fda.gov/ cdrh/pdf/p000033.html (November 2007). 16. US Food and Drug Administration, Center for Devices and Radiological Health. Gore Viabahn Endoprosthesis P040037 (2005). http://www.fda.gov/cdrh/pdf4/ p040037.html (November 2007). 17. Krankenberg H, Schluter M, Steinkamp HJ, Burgelin K, Scheinert D, Schulte KL, Minar E, Peeters P, Bosiers M, Tepe G, Reimers B, Mahler F, Tubler T, Zeller T. Nitinol stent implantation versus percutaneous transluminal angioplasty in superficial femoral artery lesions up to 10 cm in length: the femoral artery stenting trial (FAST). Circulation 2007;116:285 292. 18. Grimm J, Muller-Hulsbeck S, Jahnke T, Hilbert C, Brossmann J, Heller M. Randomized study to compare PTA alone versus PTA with Palmaz stent placement for femoropopliteal lesions. J Vasc Interv Radiol 2001;12:935 942. 19. Orwin R. A fail safe N for effect size in meta-analysis. J Educ Stat 1983;8: 157 159. 20. Rosenberg MS. The file-drawer problem revisited: a general weighted method for calculating fail safe numbers in meta-analysis. Evol Int J Org Evol 2005;59:464 468. 21. Katzen BT, Laird J; RESILIENT investigators. The RESILIENT trial. 12-month analysis. In: TCT 2007 (Transcatheter Cardiovascular Therapeutics). Washington, DC, 20 25 October 2007. http://www.cardiosource.com/clinicaltrials/ trial.asp?trialid51615 (5 November 2008). 22. Cejna M, Schoder M, Lammer J. [PTA vs. stent in femoro-popliteal obstruction]. Radiologe 1999;39:144 150. 23. Grenacher L, Saam T, Geier A, Muller-Hulsbeck S, Cejna M, Kauffmann GW, Richter GM. [PTA versus Palmaz stent placement in femoropopliteal artery stenoses: results of a multicenter prospective randomized study (REFSA)]. Rofo 2004;176:1302 1310. 24. Greenberg D, Rosenfield K, Garcia LA, Berezin RH, Lavelle T, Fogleman S, Cohen DJ. In-hospital costs of self-expanding nitinol stent implantation versus balloon angioplasty in the femoropopliteal artery (the VascuCoil Trial). J Vasc Interv Radiol 2004;15:1065 1069. 25. Scheinert D, Scheinert S, Sax J, Piorkowski C, Braunlich S, Ulrich M, Biamino G, Schmidt A. Prevalence and clinical impact of stent fractures after femoropopliteal stenting. J Am Coll Cardiol 2005;45:312 315. 26. Hayerizadeh R, Zeller T, Krankenberg H, Scheinert D, Rastan A, Geier S, Braunlich S, Wegscheider K, Biamino G. Long-term outcome of superficial femoral artery stenting using nitinol stents compared with stainless steel stents. A multicenter study. Am J Cardiol 2003;92(suppl):S157. 27. Sabeti S, Schillinger M, Amighi J, Sherif C, Mlekusch W, Ahmadi R, Minar E. Primary patency of femoropopliteal arteries treated with nitinol versus stainless steel self-expanding stents: propensity score-adjusted analysis. Radiology 2004; 232:516 521. 28. Adam DJ, Beard JD, Cleveland T, Bell J, Bradbury AW, Forbes JF, Fowkes FG, Gillepsie I, Ruckley CV, Raab G, Storkey H. Bypass versus angioplasty in severe ischaemia of the leg (BASIL): multicentre, randomised controlled trial. Lancet 2005;366:1925 1934. 29. L Abbe KA, Detsky AS, O Rourke K. Meta-analysis in clinical research. Ann Intern Med 1987;107:224 233.