FormalPara Key Points

Patients with PAD, including those with symptoms of intermittent claudication and more-severe critical limb ischaemia, benefit from the combination of revascularisation and supervised exercise therapy. Combined therapy leads to significant improvements in walking capacity over revascularisation or exercise therapy alone.

Improvements in walking capacity with combined therapy are achieved with no further improvement in leg blood flow compared with revascularisation alone. The mechanisms that underpin the benefits of the combined therapy approach remain to be determined.

Combining supervised exercise training as adjunct to revascularisation in patients with PAD is a promising treatment strategy. There is a need to define optimal and specific exercise training programs for PAD patients who have undergone revascularisation, as well as strategies to improve compliance with exercise programs.

1 Introduction

Peripheral arterial disease (PAD) is an atherosclerotic disease that leads to stenosis and occlusion of the arteries, most commonly affecting the lower extremities. PAD affects more than 200 million adults globally [1], and is associated with exercise intolerance, leading to impaired physical function [2–5] and quality of life [6–8]. Hospitalisation rates in patients with PAD are high and PAD is associated with high rates of cardiovascular events and revascularisation procedures [9].

Treatment for PAD aims to reduce cardiovascular risk, reduce leg symptoms and improve exercise tolerance [10, 11]. The recommended initial treatment for PAD comprises atherosclerotic risk factor modification with or without pharmacological therapy, and supervised exercise training [10, 11]. For patients who are unresponsive to exercise and pharmacological therapy, those with severe functional disability, or those with limb-threatening ischaemia, lower limb revascularisation, including endovascular and open surgical procedures, is indicated to improve limb circulation and for limb salvage [10, 11].

Revascularisation procedures and supervised exercise training have both independently been shown to improve functional outcomes and quality of life in patients with symptomatic PAD [12, 13]. Revascularisation procedures restore blood flow to the lower limbs and thereby improve muscle oxygen delivery and exercise capacity [14]. Conversely, exercise training leads to improvements in exercise capacity, usually without any significant gains in limb blood flow [15]. This raises the possibility that combining both therapies would result in greater improvements in exercise capacity compared to either revascularisation or exercise training alone. This combined treatment approach where exercise therapy is delivered as an adjunct to surgical procedures is commonly recommended for patients undergoing cardiac, thoracic or orthopaedic surgery in order to promote recovery and secondary prevention [16]. Given that the potential utility of this treatment strategy has been acknowledged for patients with PAD [11, 17], a better understanding of the benefits of combining revascularisation and exercise therapy would facilitate this approach being adopted into treatment guidelines and standard care for patients with more severe PAD. Therefore, we aimed to systematically review the available evidence of the effect of lower limb revascularisation combined with supervised exercise training in patients with PAD, on walking capacity, limb haemodynamics and quality of life.

2 Methods

This review was conducted according to the guidelines of the 2009 Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement [18].

2.1 Search Strategy

The literature search was performed, with no language restrictions, using the electronic databases Cochrane Central Register of Controlled Trials (CENTRAL) (1996 to present), PubMed (1996 to present), Embase (1947 to present), Scopus (1966 to present), CINAHL (1980 to present), SPORTDiscus (1871 to present) and Web of Science (1900 to present). Databases were searched from the earliest records up to January 2016 (last search 04/01/2016). Bibliographies of all eligible primary studies were manually searched for additional studies. Search terms and strategies for each database are presented in Electronic Supplementary Material Tables S1–S7. This review was registered in the international database of prospectively registered systematic reviews in health and social care (PROSPERO; CRD42015023983), publically available at http://www.crd.york.ac.uk/PROSPERO/.

2.2 Study Eligibility Criteria

Study populations consisted of atherosclerotic PAD patients who were considered eligible for both revascularisation and supervised exercise training therapy. This included the presence of target lesions suitable for revascularisation, intermittent claudication (IC) symptoms that were not improved with standard medical care and the absence of co-morbidities for which exercise is contraindicated. Only parallel-group randomised controlled trials including combined therapy of lower limb revascularisation plus supervised exercise training, and a comparison group of either revascularisation alone, supervised exercise training alone, standard medical therapy or no therapy (control) were eligible for inclusion in the present review. There was no restriction based on the stage of the disease and disease severity, the type of lower limb revascularisation procedure (surgical or endovascular), the site of the primary lesion, or the type of supervised exercise training performed.

The primary outcome measures were pain-free walking distance and maximal walking distance measured by a treadmill test. Studies using either incremental (graded) or constant-load treadmill walking tests were included, but the outcomes were considered separately based on test type. The secondary outcome measures were 6-min walking distance, ankle-brachial index (ABI), resting calf arterial blood flow (venous occlusion plethysmography) and quality of life (physical and mental health summary scores of the 36-item Short-Form Health Survey [SF-36]; 0 [severe limitation] to 100 [no limitation]). Trials were excluded if group allocation was not randomised, if they did not report primary outcome data or if group data were not presented separately. Non-peer-reviewed publications (e.g. conference abstracts) were excluded.

2.3 Study Selection and Data Extraction

One reviewer (ALM) conducted the literature search and the results from all databases were downloaded into bibliographic software (EndNote X6; Thompson Reuters, New York, USA). After eliminating duplicates, all papers identified by the search strategy were screened by title and abstract. Full-text evaluation was performed to confirm the suitability of the selected trials for inclusion in the review, and the bibliographies of all included studies were manually searched for additional studies. Papers where abstracts were published in languages other than English were evaluated by experienced researchers who were native-speakers of the particular language with the use of a standardised review template designed for this study. Quality assessment of all eligible papers was undertaken separately and in duplicate by two reviewers (ALM and RMR) using the Physiotherapy Evidence Database Scale for randomised controlled trials (PEDro) [19]. One reviewer (ALM) extracted the data from included studies and a second reviewer (CDA) checked the accuracy of extracted data. Disagreements between the reviewers were resolved by consensus or by a third reviewer when necessary.

Data extracted from each included trial consisted of: (1) characteristics of trial participants; (2) characteristics of the revascularisation procedure; (3) characteristics of the supervised exercise training intervention; (4) protocols for assessment of primary and secondary outcomes; and (5) between-group comparative data pre (baseline) and immediately post-intervention.

2.4 Data Analysis

Trials were divided into two groups: (1) combined therapy versus revascularisation alone, which assessed the added effect of exercise over revascularisation, and (2) combined therapy versus supervised exercise training alone, which assessed the added effect of revascularisation over exercise therapy. There were no other comparative treatment combinations identified in the literature. Due to the heterogeneity of protocols for the assessment of primary outcomes, a systematic review was conducted without meta-analysis. Where there were missing outcome data, or when data were not provided as mean and standard deviation (SD), corresponding authors were contacted. Where duplicate data were suspected, authors were contacted to clarify whether data were published across multiple papers. If partial data sets had been published previously, the more complete data sets were extracted for analysis.

Mean differences and between-group effect sizes (ES) were calculated for each trial using the pre-post intervention changes in mean and SD for each comparison group. Between-group ES, adjusted using Hedges’ bias-correction for small sample sizes, and 95 % confidence intervals (CIs) were calculated for each outcome measure where applicable as: between-group ES = (∆treatment − ∆control)/pooled baseline SD; where ∆ indicates change. In some trials, the minimum outcome data required for ES calculation were not reported. Despite an attempt to request missing data, few of the data could be obtained. Therefore, where findings were presented as median, range, interquartile range or standard error, data were converted to derive means and SD prior to the calculation of the ES, as per Hozo [20]. ES were interpreted according to Cohen’s classification of ‘trivial’ (<0.20), ‘small’ (≥0.20 to <0.50), ‘moderate’ (≥0.50 to <0.80) and ‘large’ (≥0.80) [21]. ES and 95 % CIs were graphed as forest plots and the between-group ES for each study was deemed to have no statistical significance when the CI crossed the vertical midline (i.e. included zero) [22]. Values are presented as mean and SD and significance was set at p < 0.05.

3 Results

Results of the literature search are presented in Fig. 1. The search identified a total of 9701 citations. Fifteen studies were retrieved for full-text examination; however, eight trials did not meet inclusion criteria and were subsequently excluded [13, 23–29]. The reasons for the exclusion of studies are described in Electronic Supplementary Material Table S8. One study was divided into two trials as patients were stratified according to the anatomical site of revascularisation (femoropopliteal or aortoiliac) prior to randomisation [30]. Therefore, seven studies and a total of eight trials were included in the systematic review [30–36].

Fig. 1
figure 1

Study flow diagram. CENTRAL Cochrane Central Register of Controlled Trials

Full size image

3.1 Quality Assessment

Overall quality of the included trials was modest, with an average of six of the ten quality criteria being present (mean ± SD 6 ± 2, range 3–8/10), as shown in Table 1. The most prevalent study limitation was related to the blinding of participants and research personnel (>75 %), which is somewhat expected considering the nature and design of these studies. Four trials (50 %) [30, 31, 33] presented insufficient information about allocation concealment. Six trials (75 %) [30, 33–36] reported an intention-to-treat analysis or stated that all participants received treatment or control conditions as originally allocated. Three trials (38 %) [30–32] measured key outcomes in less than 85 % of the participants initially allocated to a treatment group; therefore, a greater than 15 % drop-out rate may have disguised the true treatment effect of the intervention in these trials. Eligibility criteria (100 %), random allocation (100 %), baseline similarity (88 %), between-group statistics (100 %) and point measures and measures of variability (e.g. mean and SD) (100 %) were sufficiently reported by most studies. The consistency of outcomes listed in the clinical trials registries of included trials (when available) [30, 32, 34–36] were compared with their respective published results. For these studies, the registered outcome measures and time frames for outcome assessment were consistent with published outcome results.

Table 1 Assessment of methodological quality using the Physiotherapy Evidence Database (PEDro) rating scale for randomised controlled trialsa
Full size table

3.2 Study Characteristics

A summary of the study and patient characteristics is presented in Table 2. The total number of patients screened for inclusion across all trials was 3791 (range 48 to 1401); of these, 726 were enrolled (n = 253 underwent revascularisation therapy alone; n = 145 underwent supervised exercise training alone and n = 328 underwent combined therapy). Included patients presented with a mean age of 66 ± 3 years, and a resting ABI of 0.66 ± 0.05. History of IC was an inclusion criterion for all trials, and two trials (25 %) [31, 35] also included patients with lower limb ischaemic rest pain. Four trials (50 %) [30, 32, 34] stated that included patients had stable IC that had persisted for more than 3 months of standard medical care prior to study entry. Co-morbidities associated with PAD, including diabetes (22 ± 17 %), hypertension (62 ± 13 %), coronary artery disease (31 ± 10 %), as well as history of previous ischaemic events, were reported in most trials.

Table 2 Study and patient characteristics
Full size table

A summary of the intervention approaches used in the trials is presented in Table 3. Revascularisation was performed to treat femoropopliteal and/or aortoiliac artery stenoses or occlusions by endovascular (percutaneous transluminal angioplasty) or open surgical (e.g. bypass graft) procedures. Study outcomes were not assessed according to the site of revascularisation, with the exception of the study by Greenhalgh et al. [30] where patients were stratified to femoropopliteal or aortoiliac trials. In the combined therapy versus revascularisation trials [31–35], the treatment sites were reported as being similar across groups in four out of the five trials [31, 32, 34, 35], and not reported in the other [33]. Five trials (63 %) [30, 31, 34, 35] reported that standard post-operative care (drug therapy and cardiovascular risk factor control) and standard advice (e.g. smoking cessation, exercise, and diet) were provided for all patients during the subsequent exercise intervention or control period.

Table 3 Intervention characteristics
Full size table

For patients undergoing combined therapy, the reported recovery time from revascularisation to the commencement of exercise training ranged from one [32] to 12 weeks [30]. For patients undergoing revascularisation therapy only, the reported time between recovery from revascularisation and post-intervention follow-up visit ranged between six [33] to 44 weeks [36]. Walking exercise was prescribed in all trials [30–36], which in some cases was performed in addition to upper/lower limb exercises incorporated into dance classes [34], or leg exercises as part of a circuit training [30, 32, 33]. The average duration of the supervised exercise training programs was 22 ± 13 weeks (ranged from 6 [31] to 48 weeks [36]). Exercise frequency ranged from 1 to 3 days per week, with each session lasting 22 min [32] to 60 min [34]. The prescribed exercise intensity was either beyond the onset of claudication pain [33] or to a submaximal [35, 36] or maximal pain threshold [30, 36]. One trial [34] prescribed exercise intensity (walking plus dance classes) as 15–17 on the Borg 6–20 perceived exertion scale, which equates to a vigorous to high intensity [37]. All exercise training programs were completely supervised, with the exception of one trial by Bo et al. [34] where participants performed one-third of the weekly sessions without supervision. Exercise programs were usually hospital-based [30–34]. Attendance at exercise training sessions (62 ± 13 %) was reported in four trials (50 %) [30, 32, 36], with only one trial [32] reporting a priori criteria for exercise compliance (≥85 % completion of supervised exercise sessions), and three trials (38 %) reporting that exercise programs were not delivered as planned (<75 % attendance) [30, 36].

Complications following revascularisation occurred in most trials [30–33, 35, 36] and ranged from minor haematomas and sensory deficits to major complications requiring reinterventions such as persistent bleeding, graft occlusion, aortic rupture, contralateral leg embolism and toe amputations. One trial [36] reported that 2 % of the patients in the supervised exercise training group underwent minor amputation due to deterioration of claudication and progression of lower limb ischaemia. No complications related to exercise training were reported. Four trials (50 %) [32, 34–36] reported rates of reintervention (further invasive treatment required to manage complications, restenosis or for persistent/recurrent symptoms) across groups. Overall, rates of reintervention were similar in the groups of patients undergoing revascularisation alone (range 6–15 %), supervised exercise training alone (10 %) or combined therapy (range 0–14 %) over a follow-up period of 3–12 months. In one trial [36], reintervention rates were significantly lower (8 %) in the combined therapy group compared to supervised exercise training (23 %) at 1-year follow-up.

3.3 Outcome Measures

A complete dataset of the outcome measures, mean differences and ES are presented in Electronic Supplementary Material Tables S9–S15.

3.3.1 Pain-Free Walking Distance

The mean difference in pain-free walking distance between combined therapy and revascularisation therapy alone ranged from 38 to 391 m in favour of combined therapy (Fig. 2; Electronic Supplementary Material Table S9). Between-group ES were large (mean ES 1.28; range 0.45–1.73) and indicated a significant benefit of combined therapy over revascularisation alone, with the exception of one trial [33] where ES analysis revealed no significant differences between groups (p = 0.14). The mean difference between combined therapy and exercise training alone ranged from 51 to 408 m in favour of combined therapy. Between-group ES were again large (mean ES 1.14; range 0.61–1.65) and significant in favour of combined therapy (p < 0.01).

Fig. 2
figure 2

Pain-free walking distance (m): change in walking distance with revascularisation alone/exercise training alone (control) versus combined therapy of revascularisation plus exercise training. a,bGreenhalgh et al. 2008 [30] was divided into two trials as patients were stratified according to the anatomical site of revascularisation (femoropopliteal or aortoiliac) prior to randomisation. Forest plot detailing between-group effect size and 95 % confidence intervals (CIs) of each trial. Δ change from baseline, UTC unable to calculate

Full size image

3.3.2 Maximal Walking Distance

The mean difference in maximal walking distance between combined therapy and revascularisation therapy alone ranged from 97 to 321 m in favour of combined therapy (Fig. 3; Electronic Supplementary Material Table S10). Between-group ES were large (mean ES 0.86; range 0.31–1.44), with ES analysis revealing a large and significant ES (p < 0.01) in favour of combined therapy over revascularisation in two trials [32, 35], and non-significant ES (p > 0.05) in another two trials [33, 34]. Combined therapy was generally superior to exercise training alone with a mean difference that ranged from 82 to 282 m, and a large and significant between-group ES (mean ES 0.98; range 0.41–1.75), except for one trial [33] where the ES did not reach significance (p = 0.07).

Fig. 3
figure 3

Maximal walking distance (m): change in walking distance with revascularisation alone/exercise training alone (control) versus combined therapy of revascularisation plus exercise training. Forest plot detailing between-group effect size (ES) and 95 % confidence intervals (CIs) of each trial. a,bGreenhalgh et al. 2008 [30] was divided into two trials as patients were stratified according to the anatomical site of revascularisation (femoropopliteal or aortoiliac) prior to randomisation. Greenhalgh et al. 2008 [30]a and Greenhalgh et al. 2008 [30]b did not report post-intervention data (standard deviation) for each comparison group but did report between-group ES data. Δ change from baseline, UTC unable to calculate

Full size image

3.3.3 Six-min Walking Distance

Only one trial measured 6-min walking distance [34]. The mean difference between combined therapy and revascularisation therapy alone was 7 m (95 % CI −48 to 63). The between-group ES was trivial with no significant difference between groups (ES 0.08, 95 % CI −0.49 to 0.64; p = 0.79; Electronic Supplementary Material Table S11).

3.3.4 Resting Ankle-Brachial Index

The mean difference in ABI between combined therapy and revascularisation therapy alone ranged from −0.05 to 0.13, and the between-group ES were mostly trivial (mean ES 0.06; range −0.35 to 0.53) with no significant difference between groups (Fig. 4; Electronic Supplementary Material Table S12), with the exception of one trial [32], which reported a significant effect in favour of combined therapy (p = 0.01). The mean difference in ABI between combined therapy and exercise training alone ranged from 0.13 to 0.31. Between-group ES were large (mean ES 1.26; range 0.64–1.82) and significant in favour of combined therapy (p < 0.01).

Fig. 4
figure 4

Resting ankle-brachial index (ABI): change in ABI with revascularisation alone/exercise training alone (control) versus combined therapy of revascularisation plus exercise training. Forest plot detailing between-group effect size (ES) and 95% confidence intervals (CIs) of each trial. a,bGreenhalgh et al. 2008 [30] was divided into two trials as patients were stratified according to the anatomical site of revascularisation (femoropopliteal or aortoiliac) prior to randomisation. Greenhalgh et al. 2008 [30]a and Greenhalgh et al. 2008 [30]b reported only 2-year follow-up data, which revealed a significant effect in favour of combined therapy over supervised exercise training in both trials (Greenhalgh et al. 2008 [30]a ES 0.11, 95 % CI 0.03–0.20; p = 0.01; Greenhalgh et al. 2008 [30]b ES 0.14, 95 % CI 0.03–0.26; p = 0.02). Δ change from baseline, UTC unable to calculate

Full size image

3.3.5 Resting Calf Arterial Blood Flow

Resting calf arterial blood flow was only reported in one study by Lundgren et al. [33]. The mean difference in resting calf blood flow between combined therapy and revascularisation therapy was 2.30 ml·100 ml−1·min−1 (95 % CI −3.48 to 8.08), which resulted in a small and non-significant between-group ES (ES 0.24, 95 % CI −0.36 to 0.84; p = 0.43; Electronic Supplementary Material Table S13) [33]. The mean difference between combined therapy and exercise training alone was 6.90 ml·100 ml−1·min−1 (95 % CI 1.40–12.40). The between-group ES was moderate (ES 0.76, 95 % CI 0.14–1.38) and significant in favour of combined therapy (p = 0.02).

3.3.6 Health-Related Quality of Life

The mean difference between combined therapy and revascularisation in the scores of the physical health component of quality of life ranged from −2 to 44 (Fig. 5; Electronic Supplementary Material Table S14). Between-group ES were mostly negligible and non-significant (mean ES 0.00; range −0.20 to 0.96), except for one trial [32] where the ES revealed a large and significant effect in favour of combined therapy (p < 0.01). Based on this single trial [32], combined therapy had a greater effect on scores of physical domains of quality of life with a mean difference of 25 (95 % CI 8–42) and a moderate between-group ES (ES 0.59, 95 % CI 0.18–0.99; p < 0.01) over exercise training alone.

Fig. 5
figure 5

Quality of life (SF-36 Physical Health): change in scores with revascularisation alone/exercise training alone (control) versus combined therapy of revascularisation plus exercise training. Forest plot detailing between-group effect size (ES) and 95 % confidence intervals (CIs) of each trial. a,bGreenhalgh et al. 2008 [30] was divided into two trials as patients were stratified according to the anatomical site of revascularisation (femoropopliteal or aortoiliac) prior to randomisation. Greenhalgh et al. 2008 [30]a and Greenhalgh et al. 2008 [30]b reported only 2-year follow-up data, which revealed that combined therapy had no added effect over supervised exercise training in the femoropopliteal trial (Greenhalgh et al. 2008 [30]a ES −0.4, 95 % CI −4.2 to 3.4; p = 0.82), and a significant effect in favour of combined therapy in the aortoiliac trial (Greenhalgh et al. 2008 [30]b ES 7.8, 95 % CI 1.5–14.1; p = 0.02). Δ change from baseline, UTC unable to calculate

Full size image

The mean difference between combined therapy and the revascularisation therapy in the scores of the mental health component of quality of life ranged from 0 to 5 (Fig. 6; Electronic Supplementary Material Table S15). Between-group ES were small (mean ES 0.23; range 0.00–0.41) with no significant difference between groups (p > 0.05). The mean difference in scores between combined therapy and exercise training, calculated for one trial [32], was −33 (95 % CI −75 to 8). The between-group ES was negative and small, with no significant difference between groups (ES −0.31, 95 % CI −0.71 to 0.08; p = 0.12).

Fig. 6
figure 6

Quality of life (SF-36 Mental Health): change in scores with revascularisation alone/exercise training alone (control) versus combined therapy of revascularisation plus exercise training. Forest plot detailing between-group effect size (ES) and 95% confidence intervals (CIs) of each trial. a,bGreenhalgh et al. 2008 [30] was divided into two trials as patients were stratified according to the anatomical site of revascularisation (femoropopliteal or aortoiliac) prior to randomisation. Greenhalgh et al. 2008 [30]a and Greenhalgh et al. 2008 [30]b reported 2-year follow-up data, which revealed that combined therapy had no added effect over supervised exercise training in both trials (Greenhalgh et al. 2008 [30]a ES 2.4, 95% CI −1.7 to 6.5; p = 0.25; Greenhalgh et al. 2008 [30]b ES 4.9, 95% CI −1.3 to 11.1; p = 0.12). Δ change from baseline, UTC unable to calculate

Full size image

4 Discussion

This review identified a total of eight randomised controlled trials investigating the effect of combined therapy of revascularisation and supervised exercise training in patients with PAD and symptoms of IC. The summary of findings indicate that combined therapy leads to greater improvements in pain-free and maximal walking distances compared with revascularisation or supervised exercise training alone. Combined therapy improved ABI to a similar extent as revascularisation, and to a greater extent than supervised exercise training. In addition, findings from a small number of trials suggest that: (1) combined therapy may lead to greater improvements in leg blood flow than supervised exercise training (one trial), and no added effect compared to revascularisation alone (one trial); and (2) combined therapy may lead to greater gains in physical quality of life than supervised exercise training (one trial), with no further improvements in the physical and mental domains of quality of life than revascularisation alone (three trials).

There has previously been significant interest in comparing the effects of revascularisation with exercise therapy for patients with PAD, and reviews of this literature generally conclude that both forms of treatment are equally beneficial for improving walking ability [38, 39]. Our current systematic review offers a new perspective by showing that patients with PAD, including those with symptoms of more severe disease, benefit from the combination of revascularisation and supervised exercise therapy, with an additional mean difference of 38 to 408 m in treadmill walking distances over revascularisation or supervised exercise training alone. It is possible that the benefits of combined therapy are sustained over the longer term, as evidence suggests that within 1–2 years of the completion of interventions, patients treated with combined therapy maintain greater pain-free and maximal walking distances compared with patients treated with revascularisation [34] or supervised exercise training alone [30]. However, further study is required as the trial by Mazari et al. [32] showed that revascularisation therapy, exercise therapy and the combined therapy approach were equally effective in improving walking distances, with no significant difference between groups 1 year after the completion of interventions.

It is noteworthy that the single study that measured 6-min walking distance [34] failed to find a significant difference between combined therapy and revascularisation, which is in contrast with the generally large differences observed for pain-free and maximal treadmill walking distances found in the current review. While large and clinically meaningful changes in 6-min walk distance (>50 m) [40] were observed with combined therapy and revascularisation, there was no difference in the effect between the groups. This study also assessed treadmill walking distances, and there was a significant gain in pain-free walking distance with combined therapy over revascularisation alone (MD 203 m, p < 0.01). However, the effect for maximal walking distance was not significantly different between groups (MD 98 m, p = 0.18), which is somewhat consistent with the 6-min walk test finding.

The data presented here suggest that combined therapy leads to large gains in walking capacity over revascularisation or exercise therapy alone, and it does so over revascularisation alone without any further improvement in lower limb haemodynamics. While it is well established that lower limb revascularisation achieves improvements in walking capacity and ABI by restoring arterial patency and improving limb blood flow capacity [33, 41, 42], increases in walking capacity with exercise training occur with negligible changes in lower limb blood flow and ABI [15]. Thus, the significant effect of combined therapy compared with revascularisation alone on walking capacity indicates that exercise has an additive effect, and reinforces the notion that revascularisation and exercise therapy improve walking capacity via different mechanisms [43]. Although the primary pathology of PAD is impaired blood flow, other factors such as altered skeletal muscle phenotype, also contribute to patients’ poor exercise tolerance [44]. Exercise training induces adaptive changes in skeletal muscle phenotype, which include increases in muscle cross-sectional area [45], alterations in metabolic activity [46] and enhanced capillary supply [47, 48], thus potentially improving the ability of the working muscles to optimise the utilisation of the limited oxygen delivered to the lower limbs. These intramuscular changes, in conjunction with the haemodynamic benefits of revascularisation, likely explain the significant improvements in walking capacity achieved with combined therapy over exercise or revascularisation therapy alone. This is somewhat speculative as the current understanding of the mechanisms that underpin the improvements in exercise tolerance with exercise training is limited to studies of patients with IC who have not undergone a previous revascularisation intervention. Whether the mechanism of the improvement with exercise therapy differs following revascularisation (i.e. where limb blood flow has been restored) remains to be determined.

Based on our findings, it appears that patients undergoing either revascularisation alone or combined therapy perceive greater improvements in their physical and mental health than patients undergoing exercise therapy alone. This may be explained by the fact that patients are more likely to perceive improvements in quality of life as a result of immediate improvements in IC and other symptoms after revascularisation, compared with more gradual improvements that result from exercise therapy [36, 42]. It is known that exercise-induced leg pain is the most prevalent barrier to physical activity in patients with IC [49]. It is plausible to suggest that revascularisation may increase the patient’s ability to walk independently by providing symptom relief, and thereby enhance their confidence and motivation to engage in exercise and lifestyle change [50]. From a clinical perspective, given that physical activity is strongly associated with overall physical function [51] and all-cause and cardiovascular mortality in patients with PAD [52, 53], it is possible that the improvement in walking capacity achieved with combined therapy, in comparison with either revascularisation or exercise alone, would not only enhance the capacity to undertake activities of daily living, but would also be associated with a better patient prognosis as a result of increased physical activity.

With the advancement of revascularisation techniques and devices, the rates of lower limb interventions in patients with PAD have rapidly increased [54]. However, revascularisation therapy is associated with a substantial risk of complications and associated reinterventions due to initial and subsequent treatment failure [55]. Only four (50 %) trials reported reintervention rates [32, 34–36], and these limited data suggested that the need for reintervention associated with combined therapy was comparable, or even lower, than that for revascularisation alone or supervised exercise training alone. Combined therapy may improve the durability of revascularisation, which would have substantial social and cost-benefits. Fakhry et al. [36] reported that one-third of the patients undergoing combined therapy developed restenosis of their initially revascularised lesion at one-year follow-up, yet only 4 % of these patients required a secondary intervention because of recurrent symptoms. A cost-effectiveness analysis of the Mazari et al. [32] trial showed that, within 1 year, the cost per quality-adjusted life-year (QALY) was significantly higher for the group of patients undergoing revascularisation (€11,777.00), compared with supervised exercise training (€6147.04) or combined therapy (€10,649.74) [56]. Secondary interventions following revascularisation contributed significantly to the total cost of care in this study, supporting the suggestion that supervised exercise training should be offered to patients undergoing revascularisation with the aim of reducing potential reintervention costs in the long-term [56].

Supervised exercise therapy is a well-established effective treatment for improving walking capacity in patients with IC [57, 58]. While there is strong evidence to support current recommendations for the prescription of exercise as part of the initial treatment of patients with IC [57, 59, 60], there are currently no recommendations for exercise prescription in patients with more severe disease undergoing revascularisation. Further research is needed to establish whether the current exercise recommendations for patients with IC are also suitable for use post-operatively. For instance, although the majority of the included trials prescribed aerobic interval exercise at a moderate-high level of IC pain, the use of pain as a marker of intensity may not always be appropriate in post-surgical patients because it is expected that revascularisation would relieve symptoms [36, 42], and therefore enable patients to exercise with little or no ischaemic leg pain. There does not seem to be a clear relationship between exercise program length and improvement in outcomes. Fakhry et al. [36] found that the mean differences in walking distances and ABI between combined therapy and supervised training remained significant in favour of the combined therapy at the beginning (1 month), midpoint (6 months) and at completion of 12 months of the exercise program, but decreased in magnitude over time. This may be explained by the fact that the prescribed volume of exercise was reduced after the first 6 months of training, which is a somewhat unconventional approach as the goal of exercise prescription is usually to gradually progress the intensity and/or volume of exercise towards levels that can be maintained for the longer term. While the heterogeneity of exercise prescribed in the current trials make it difficult to provide guidelines on exercise prescription at this time, the observed improvements in pain-free and maximal walking distance across a range of exercise interventions would suggest that exercise training should be routine care post revascularisation. Further research is needed to develop optimal and specific exercise training programs for PAD patients who have undergone revascularisation, including guidance on the ideal time to commence exercise following revascularisation, as well as on the progression of exercise following revascularisation. These trials should aim to clarify how each component of exercise, such as intensity, frequency and duration of exercise sessions, contributes to improvements in exercise tolerance.

Exercise attendance rates were only reported in four (50 %) [30, 32, 36] of the trials included in this review, and therefore it is unclear whether exercise programs were always delivered as planned. Incomplete description of exercise attendance is highly prevalent in the PAD literature [61], and this has negative implications for the interpretation of results, investigation of dose-response effects, and replication of protocols in future studies. Low exercise attendance in one of the trials was reported to be associated with lack of motivation, working commitments, insurance-related issues and orthopaedic co-morbidities [35]. Future research is needed to identify barriers that should be addressed to enhance delivery/attendance of post-operative exercise interventions for PAD. For example, structured post-operative home-based exercise programs may be effective as an alternative to hospital-based programs or as an extended maintenance phase of such programs [62].

The efficacy of the combined therapy approach should be considered in light of potential biases associated with exercise testing procedures. First, only three of the included trials [34–36] used a graded-test protocol, which is considered the more reliable treadmill protocol for determining walking capacity in patients with PAD [63]. Despite differences in the treadmill protocols used between studies, the present review demonstrates improvements with combined therapy in both pain-free and maximal walking distances. Uniformity in the assessment of walking distance is desirable in future research to facilitate comparison of study outcomes through meta-analyses. Second, none of the included trials reported that patients were familiarised with the walking tests prior to baseline assessments. The introduction of any learning effect, particularly in patients who performed treadmill walking as part of the therapy program, has the potential to overestimate the true therapy effect [64, 65]. Third, in all of the included studies, a cap was placed on the duration of treadmill walking tests, ranging from 5 min (215 m) to 30 min (1600 m). It is likely that such caps limit the ability to accurately assess walking distances, and possibly contribute to an underestimation of the treatment effect. For example, Greenhalgh et al. [30] reported that walking distances in patients who improved beyond the allocated cap of 1000 m were taken as equal to 1000 m. Similarly, Kruidenier et al. [35] reported that 11 patients in the combined therapy group and one patient in the revascularisation group improved beyond the allocated cap of 30 min. Therefore, the implementation of open-ended, maximal graded test protocols for the assessment of walking distances should be considered in future studies [66]. The 6-min walking test is also a valid and reliable alternative for the assessment of walking capacity, and has been shown to be representative of mobility loss among patients with PAD [67, 68]. In addition, it is not associated with a learning effect when repeat testing is performed [67]. Despite its widespread use in clinical practice, only one of the included trials assessed patient’s self-paced walking ability using the 6-min walking test. Therefore, future studies should consider inclusion of the six-minute walk test as an additional outcome measure.

This review has some limitations. As expected, the protocols for the assessment of primary outcomes were not uniform across studies, and the small number of trials within each outcome subgroup precluded the ability to undertake a meta-analysis. Non-peer-reviewed publications, such as conference abstracts, were excluded from the current review, and there was no attempt to identify registered clinical trials that have not been published. Therefore, we cannot fully exclude the possibility of publication bias. Despite an attempt to request missing data, few of the data could be obtained, and therefore, analyses proceeded for continuous variables using data that were converted from median/range and standard error/confidence intervals using established estimates [20, 69]. In most instances authors reported that they were no longer able to access the study data, which was not unexpected considering that results from five of the nine trials were published >5 years ago. Whether this reveals selective reporting and whether the use of pooled data contributed to the variance in ES is unclear. The mean ES calculated for each outcome did not consider the grade of evidence for each outcome or any differences in the sample size of each study, which differs from the overall (weighted) ES generated through a meta-analysis approach. Future studies examining the effects of combined therapy should clearly report between-group comparative data and effects in accordance with standard recommendations [69].

5 Conclusion

This systematic review found that patients with PAD and symptoms of IC treated with combined therapy achieve greater improvements in walking distance than those treated with either revascularisation or exercise therapy alone. The combination of revascularisation with supervised exercise training appears to be as beneficial as revascularisation alone for improving leg haemodynamics. The limited number of studies to date makes it difficult to conclude whether combined therapy may yield better results than revascularisation or supervised exercise training alone for health-related quality of life.

The use of supervised exercise training as an adjunct to revascularisation in patients with PAD is a promising treatment strategy and the improvements in walking distance suggest that exercise training should be routine care post revascularisation. Future research is needed to investigate whether the benefits achieved with combined therapy are sustained, and whether combined therapy is cost-effective in the long-term. There is also a need to define optimal and specific exercise training programs for PAD patients who have undergone revascularisation, as well as strategies to improve compliance with exercise programs. Uniformity in the assessment of outcomes, such as treadmill walking distance, is desirable in future research to facilitate comparison of study outcomes. Revascularisation and exercise therapy appear to have an additive effect, whereby independent mechanisms possibly contribute to the improvements in walking capacity when these treatments are combined. Further research is required to fully understand the mechanisms of this beneficial effect.