Modes of exercise training for intermittent claudication

  • Review
  • Intervention

Authors


Abstract

Background

According to international guidelines and literature, all patients with intermittent claudication should receive an initial treatment of cardiovascular risk modification, lifestyle coaching, and supervised exercise therapy. In most studies, supervised exercise therapy consists of treadmill or track walking. However, alternative modes of exercise therapy have been described and yielded similar results to walking. Therefore, the following question remains: Which exercise mode gives the most beneficial results?

Objectives

Primary objective: To assess the effects of different modes of supervised exercise therapy on the maximum walking distance (MWD) of patients with intermittent claudication.
Secondary objectives: To assess the effects of different modes of supervised exercise therapy on pain-free walking distance (PFWD) and health-related quality of life scores (HR-QoL) of patients with intermittent claudication.

Search methods

The Cochrane Peripheral Vascular Diseases Group Trials Search Co-ordinator searched the Cochrane Peripheral Vascular Diseases Group Specialised Register (July 2013); CENTRAL (2013, Issue 6), in The Cochrane Lib rary; and clinical trials databases. The authors searched the MEDLINE (1946 to July 2013) and Embase (1973 to July 2013) databases and reviewed the reference lists of identified articles to detect other relevant citations.

Selection criteria

Randomised controlled trials of studies comparing alternative modes of exercise training or combinations of exercise modes with a control group of supervised walking exercise in patients with clinically determined intermittent claudication. The supervised walking programme needed to be supervised at least twice a week for a consecutive six weeks of training.

Data collection and analysis

Two authors independently selected studies, extracted data, and assessed the risk of bias for each study. Because of different treadmill test protocols to assess the maximum or pain-free walking distance, we converted all distances or walking times to total metabolic equivalents (METs) using the American College of Sports Medicine (ACSM) walking equation.

Main results

In this review, we included a total of five studies comparing supervised walking exercise and alternative modes of exercise. The alternative modes of exercise therapy included cycling, strength training, and upper-arm ergometry. The studies represented a sample size of 135 participants with a low risk of bias. Overall, there was no clear evidence of a difference between supervised walking exercise and alternative modes of exercise in maximum walking distance (8.15 METs, 95% confidence interval (CI) -2.63 to 18.94, P = 0.14, equivalent of an increase of 173 metres, 95% CI -56 to 401) on a treadmill with no incline and an average speed of 3.2 km/h, which is comparable with walking in daily life.

Similarly, there was no clear evidence of a difference between supervised walking exercise and alternative modes of exercise in pain-free walking distance (6.42 METs, 95% CI -1.52 to 14.36, P = 0.11, equivalent of an increase of 136 metres, 95% CI -32 to 304). Sensitivity analysis did not alter the results significantly. Quality of life measures showed significant improvements in both groups; however, because of skewed data and the very small sample size of the studies, we did not perform a meta-analysis for health-related quality of life and functional impairment.

Authors' conclusions

There was no clear evidence of differences between supervised walking exercise and alternative exercise modes in improving the maximum and pain-free walking distance of patients with intermittent claudication. More studies with larger sample sizes are needed to make meaningful comparisons between each alternative exercise mode and the current standard of supervised treadmill walking. The results indicate that alternative exercise modes may be useful when supervised walking exercise is not an option for the patient.

Plain language summary

Modes of exercise training for intermittent claudication

Intermittent claudication is a cramping leg pain that develops during exercise and is relieved by a short period of rest. It is caused by inadequate blood flow to the muscles of the leg because of atherosclerosis (hardening and narrowing of the arteries). Intermittent claudication is closely associated with other vascular diseases, such as a heart attack or stroke. Therefore, all patients with intermittent claudication should receive cardiovascular risk management and lifestyle coaching to reduce cardiovascular risk factors.

To improve the walking capacity and quality of life, supervised exercise therapy is the primary treatment according to the current scientific evidence. Community-based supervised exercise appears to be at least as effective as programmes provided in a hospital setting. In the literature, supervised exercise therapy usually consists of treadmill walking. However, alternative modes of exercise therapy (e.g. cycling, strength training) have been described, with beneficial effects on walking capacity and quality of life. Therefore, the following question remains: Which exercise mode gives the most beneficial results?

The present review shows that there are few studies comparing alternative modes of exercise training to the standard of supervised walking exercise. The review authors identified five studies that randomised a total of 135 participants. The alternative modes of exercise therapy included cycling, strength training, and upper-arm ergometry. Comparing these alternative modes of exercise with supervised walking exercise showed no clear evidence of a difference in maximum or pain-free walking distance between the groups. Quality of life measures showed significant improvements in both groups; however, we could not make a direct comparison because of limited data.

This review shows that alternative modes of exercise therapy seem to yield similar results to supervised walking therapy. Therefore, they may be considered useful when supervised walking exercise is not an option for the patient. However, more randomised controlled trials are needed to make a meaningful comparison between the different modes of exercise therapy.

Laienverständliche Zusammenfassung

Übungsformen bei Claudicatio intermittens („Schaufensterkrankheit“)

Claudicatio intermittens („Schaufensterkrankheit“) ist ein krampfartiger Beinschmerz, der sich während der Durchführung von Übungen entwickelt und sich durch eine kurze Pause löst. Ursache ist eine gestörte Blutzufuhr in die Beinmuskulatur durch Arteriosklerose (Verhärtung und Verengung der Arterien). Claudicatio intermittens steht in engem Zusammenhang mit anderen Gefäßerkrankungen, wie zum Beispiel Herzinfarkt und Schlaganfall. Folglich sollten alle Patienten mit Claudicatio intermittens ein kardiovaskuläres Risikomanagement (Steuerung von Risikofaktoren für Herz-Kreislauferkrankungen) und ein Lebensstil-Coaching (Beratung) erhalten, um kardiovaskuläre Risikofaktoren zu verringern.

Zur Verbesserung der Gehausdauer und Lebensqualität ist eine angeleitete Übungstherapie nach aktuellem wissenschaftlichen Kenntnisstand die vorrangige Behandlung. Wohnortnah angebotene angeleitete Übungsprogramme scheinen dabei mindestens genauso wirksam zu sein wie in Krankenhäusern durchgeführte Programme. In der Literatur umfasst die angeleitete Übungstherapie üblicherweise das Gehen auf dem Laufband. Es sind jedoch auch andere Formen der Übungstherapie (zum Beispiel Radfahren, Krafttraining) mit positiver Wirkung auf die Gehausdauer und Lebensqualität beschrieben worden. Daher bleibt die nachfolgende Frage offen: Welche Übungsformen erreichen die besten Ergebnisse?

Die vorliegende Übersichtsarbeit zeigt, dass es nur wenige Studien gibt, die andere Übungsformen mit den standardmäßig durchgeführten angeleiteten Gehübungen verglichen. Die Autoren der Übersichtsarbeit fanden fünf Studien mit insgesamt 135 zufällig zugeordneten Teilnehmern. Die anderen Formen der Übungstherapie beinhalteten Radfahren, Krafttraining und Ergometertraining für die Oberarme. Der Vergleich dieser alternativen Formen der Übungstherapie mit angeleiteten Gehübungen ergab keine eindeutige Evidenz (wissenschaftlichen Beleg) für einen Unterschied hinsichtlich der maximalen oder schmerzfreien Gehstrecke zwischen den Gruppen. Messungen der Lebensqualität zeigten signifikante Verbesserungen in beiden Gruppen; allerdings konnten wir aufgrund begrenzter verfügbarer Daten keinen direkten Vergleich anstellen.

Diese Übersichtsarbeit zeigt, dass alternative Formen der Übungstherapie ähnliche Ergebnisse zu erzielen scheinen wie angeleitete Gehübungen. Deshalb könnten diese als sinnvoll betrachtet werden, wenn angeleitete Gehübungen keine Option für den Patienten sind. Dennoch werden weitere randomisierte kontrollierte Studien benötigt, um einen aussagekräftigen Vergleich zwischen den verschiedenen Formen der Übungstherapie ziehen zu können.

Anmerkungen zur Übersetzung

übersetzt von M. Gooßes & T. Bossmann, Juli 2014

Background

Description of the condition

Peripheral arterial occlusive disease (PAOD) is a chronic arterial occlusive disease caused by progressive atherosclerosis. Several arterial segments can be affected, such as the aorta; iliac; and femoral, popliteal, and tibial arteries in the limbs. The most common symptom is intermittent claudication (IC), defined as a cramping pain in the muscles of the leg(s) that occurs during exercise and is relieved by a short period of rest. Because of this condition, patients have a diminished maximum and pain-free walking capacity, leading to diminished health-related quality of life (Dumville 2004).

The incidence of IC increases with age, with an annual incidence rate of 0.7%, 3.9%, and 10.6% among 35- to 44-year-old men, 45- to 54-year-old men, and 55- to 64-year-old men, respectively (Kannel 1985). In women, the incidence rates are approximately 50% lower (Kannel 1985). IC restricts patients' activity and mobility and considerably reduces their health-related quality of life (Dumville 2004; McDermott 2001). In addition, because of the ongoing generalised atherosclerotic process, IC is closely associated with cardiovascular morbidity and mortality. Patients with IC have a five-year all-cause mortality rate of 10% to 15% and a 20% chance of a non-fatal cardiovascular event (Hirsch 2006). When IC progresses to critical limb ischaemia, an even higher mortality rate of 25% after one year is found (Norgren 2007).

Description of the intervention

Because of the serious health risks, all patients with IC should receive a multicomponent therapy consisting of cardiovascular risk modification, lifestyle coaching, and exercise therapy (Norgren 2007). Several randomised controlled trials and systematic reviews compared walking exercise supervised by a physical or exercise therapist to non-supervised exercise, usual care, placebo, single walking advice, endovascular interventions, or bypass surgery (Creasy 1990; Fokkenrood 2013; Fowkes 1998; Lundgren 1989; Spronk 2009). The current evidence supports supervised exercise therapy as the primary treatment for improvement of walking capacity and health-related quality of life in patients with IC. Furthermore, community-based supervised exercise appears to be at least as efficacious as programmes provided in a hospital setting (Bendermacher 2007; Kruidenier 2009; Nicolaï 2010). However, less attention has been paid to the mode of (supervised) exercise. Besides walking, alternative modes of supervised exercise training, such as cycling, upper-extremity cycle ergometer exercise, and strength training exist and are associated with a significantly improved walking capacity (Hiatt 1994; Sanderson 2006; Treat-Jacobson 2009).

How the intervention might work

A number of potential mechanisms have been suggested for the reduced functional capacity in IC, such as blood flow limitation due to arterial obstruction, disruption of endothelial function, altered skeletal muscle phenotype by mitochondrial dysfunction, increased blood viscosity, and inflammatory activation (Hamburg 2011). Exercise has the potential to reverse these pathological events and thereby interrupt the clinical course toward disability (Hamburg 2011).

Why it is important to do this review

In most studies, supervised exercise programmes involve treadmill or track walking that is of sufficient intensity to bring on claudication pain. Walking exercise is alternated with rest over the course of a 30- to 60-minute session. Exercise therapy for IC is recommended at least three times a week for three months, although there does not seem to be a clear dose-response relationship between exercise volume or intensity and symptom relief (Norgren 2007; Parmenter 2011). Unfortunately, some groups of patients with IC are not capable of completing the exercise protocol because of concomitant comorbidities, such as arthrosis, chronic obstructive pulmonary disease, stroke, or cardiac complaints. For these patients, an adjusted protocol or alternative exercise regime may be proposed.

Recently, a systematic review comparing any mode of exercise, whether supervised or unsupervised, was published (Parmenter 2011). It suggested that alternative modes of aerobic exercise, other than walking, appear equally beneficial compared to walking exercise, while the effects of progressive resistance training and upper-body exercise seem only promising. The effect size of each exercise mode was calculated and compared to the effect size of walking exercise. However, because of heterogeneity, no meta-analysis was performed on specific randomised controlled trials (RCTs) comparing the standard of supervised walking exercise to alternative exercise regimes.

Therefore, the question regarding which exercise mode gives the most beneficial results in walking distance, health-related quality of life, or both, in patients with IC remains to be answered. Previous Cochrane systematic reviews focused on the effect of exercise compared with usual care and the value of a supervised exercise programme in relation to non-supervised exercise (Fokkenrood 2013; Watson 2008). This systematic review will determine the effect of alternative exercise modes by analysing randomised controlled trials comparing the current standard of supervised walking exercise to alternative modes of exercise. Studies focusing on this research topic are increasing, implicating the need for a regular update of this review in the upcoming years.

Objectives

Primary objective: To assess the effects of different modes of supervised exercise therapy on the maximum walking distance (MWD) of patients with intermittent claudication.
Secondary objectives: To assess the effects of different modes of supervised exercise therapy on pain-free walking distance (PFWD) and health-related quality of life scores (HR-QoL) of patients with intermittent claudication.

Methods

Criteria for considering studies for this review

Types of studies

We included parallel-group, randomised controlled trials (RCTs) comparing (combinations of) alternative modes of exercise training (for example, ergometry, strength training, aerobic exercise, etc.) with supervised walking exercise in patients with IC. We excluded cross-over, factorial, or cluster RCTs.

Types of participants

The study population consisted of adults (18 years and older) with clinically determined IC, according to Fontaine stage II or Rutherford stages 1 to 3, who were considered for conservative treatment. We excluded studies of participants with asymptomatic lower-limb atherosclerosis identified by testing. When studies described a mixture of asymptomatic and symptomatic participants, we contacted the authors to ask if a subanalysis was available. If not, we excluded these studies.

Types of interventions

We included all RCTs comparing alternative modes of exercise training (e.g. arm ergometry, strength training, cycling, etc.) or combinations of exercise modes with a control group of supervised walking exercise. Supervised walking exercise needed to be supervised at least twice a week for a consecutive six weeks of training. We excluded studies reporting an exercise programme with a duration of less than six weeks of training or with less than two supervised walking sessions a week.

Since different types of alternative exercise modes are associated with an increased walking capacity, we combined all studies with different alternative exercise modes in one analysis. However, to analyse the effect of each individual or combined alternative exercise modes, we performed a subgroup analysis on the exercise mode (or combination) if more than one study was available.

We excluded all types of mechanical intermittent compression treatments as we did not consider them to be exercise training. Furthermore, we did not include studies comparing different types of walking exercise (supervised versus unsupervised, community versus hospital-based) or comparisons of different walking protocols (low- versus high-frequency training, low- versus high-intensity training, different treadmill exercise protocols).

Types of outcome measures

Primary outcomes

The primary outcome measurement was the maximum walking distance (MWD) measured by a treadmill test. Outcome measurements needed to be available at baseline and after at least six weeks of follow up. In case of different treadmill test protocols, we converted walking times or distances to total metabolic equivalents (total METs or sum of METs during the walking period). 

Secondary outcomes

Secondary outcome measurements were the pain-free walking distance (PFWD) and health-related quality of life (HR-QoL) scores. Besides a baseline measurement, results needed to be available after at least six weeks of follow up. In case of different treadmill test protocols, we converted walking times or distances to METs. 

Search methods for identification of studies

We did not use language restrictions.

Electronic searches

The Cochrane Peripheral Vascular Diseases Group Trials Search Co-ordinator (TSC) searched the Cochrane Peripheral Vascular Diseases Group Specialised Register (last searched July 2013) and the Cochrane Central Register of Controlled Trials (CENTRAL) 2013, Issue 6, part of The Cochrane Library, www.thecochranelibrary.com. See Appendix 1 for details of the search strategy used to search CENTRAL. The Specialised Register is maintained by the TSC and is constructed from weekly electronic searches of MEDLINE, EMBASE, CINAHL, AMED, and through handsearching relevant journals. The full list of the databases, journals and conference proceedings which have been searched, as well as the search strategies used are described in the Specialised Register section of the Cochrane Peripheral Vascular Diseases Group module in The Cochrane Library www.thecochranelibrary.com.

The TSC searched the following trial databases (July 2013) for details of ongoing and unpublished studies using the terms exercise and claudication:

Searching other resources

The authors searched MEDLINE (1946 to 15 July 2013) and Embase (1973 to 15 July 2013) using the search strategies described in Appendix 2 and Appendix 3. We reviewed the reference lists of articles identified by the above search strategies to identify other relevant citations.

Data collection and analysis

Selection of studies

GJL and FF independently selected trials for this review. SS and JT confirmed the suitability of selected trials for inclusion in the review. We sought additional information for included trials, if necessary.

Data extraction and management

GJL and FF independently extracted data using a standard data collection form created for this review. We entered the data into Review Manager (RevMan 5.2). We resolved disagreements between the review authors by discussion. SS acted as arbiter if no consensus was achievable. The extracted study data consisted of the following:

  1. study characteristics, including study design, method of randomisation, exclusions postrandomisation, publication year, country, and study period;

  2. baseline characteristics, including number of participants, losses to follow up, mean age, gender distribution, and inclusion and exclusion criteria;

  3. type of interventions, including mode(s) of exercise, duration of programme, number of sessions, number of supervised sessions, and exercise protocol; and

  4. the mean maximum walking distance or time, mean pain-free walking distance or time, and mean quality of life scores at baseline and follow-up periods.

Assessment of risk of bias in included studies

The authors (GJL and FF) assessed the risk of bias for each study as described in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011) for each of the following domains:

  1. randomisation/sequence generation;

  2. allocation concealment;

  3. blinding (of participants, personnel, and outcome assessors);

  4. incomplete outcome data;

  5. selective outcome reporting; and

  6. publication and other sources of bias.

Measures of treatment effect

To analyse treatment effect, we assessed the MWD, PFWD, and HR-QoL scores after participation in the exercise programme. In the case of different treadmill test protocols, we converted walking times or distances to total METs (sum of METs during the walking period) using the American College of Sports Medicine (ACSM) walking equation (ACSM 2006). The ACSM walking equation includes the time, speed, and inclination of the treadmill test. Since direct conversion of the walking times or distances to METs was not possible because of the absence of individual participant data, we simulated a new dataset for each study. We used postintervention walking distances and variances from the intervention and control group to simulate a new dataset assuming a normal distribution. For each simulated individual, we calculated the number of METs using the ACSM walking equation. We used the summary measures (mean METs and variances) from each simulated dataset as postintervention outcomes from the included studies.

Unit of analysis issues

We searched for RCTs with at least six weeks' duration of training and a parallel-group design. We included no cross-over trials in this review.

In assessment of the primary and secondary outcomes, we considered the participant the unit of analysis.

Dealing with missing data

We expected missing standard deviations for walking distances. In the case of missing data, we requested data from the original investigators, if appropriate. We did not impute missing outcome data for the primary and secondary outcomes.

Assessment of heterogeneity

For all the outcome measures, we assessed statistical heterogeneity by calculating the Q-statistic or Chi² test (P < 0.10 considered as heterogeneous) and the I² statistic (I² statistic greater than 50% considered as moderate to substantial risk of heterogeneity) in order to assess to what degree the data from the included studies were heterogeneous (Higgins 2011).

Assessment of reporting biases

In case of sufficient studies (> 10 studies), we planned to assess publication bias with a funnel plot with the maximum walking distance on the X axis and the standard error of each study on the Y axis (Higgins 2011). If there is bias, for example, because smaller studies without statistically significant effects remain unpublished, this will lead to an asymmetrical appearance of the funnel plot. In this situation, the effect calculated by the meta-analysis will tend to overestimate the intervention effect. Therefore, we also planned to evaluate funnel plot asymmetry using Begg and Egger tests (Begg 1994; Egger 1997) performed with Stata statistical software (StataCorp 2011).

Data synthesis

To analyse treatment effect, we used the DerSimonian and Laird random-effects model. This model takes the variance between studies and the variance within a study into account (DerSimonian 1986). We summarised the data of each study in forest plots and calculated summary estimates with a 95% confidence interval. We considered a two-sided P ≤ 0.05 as statistically significant, except for the test of publication bias for which the recommended levels are P ≤ 0.10. We performed analyses using RevMan 5.2.

Subgroup analysis and investigation of heterogeneity

We performed a subgroup analysis of each type of alternative exercise mode if we found more than one trial comparing the specific exercise mode with walking exercise. Furthermore, because most studies and international guidelines advise a 12-week supervised exercise programme, we reported outcomes both at the end of training and at 12 weeks of training. Finally, we performed a subgroup analysis on the combination of alternative exercise modes in relation to supervised walking exercise.

Sensitivity analysis

We examined individual study effects on the results by removing each study one at a time to examine whether removing a particular study would significantly change the results. 

In addition, we planned to perform sensitivity analyses on the methodological quality of the studies. We planned to exclude studies with apparent methodological flaws and risk of bias and examine whether removing these studies significantly changed results.

Results

Description of studies

See: Characteristics of included studies; Characteristics of excluded studies; Characteristics of ongoing studies.

Results of the search

See Figure 1.

Figure 1.

Study flow diagram

Included studies

We identified 10 publications comparing an alternative exercise regime to supervised walking exercise for IC. Of these 10 publications, we identified five primary studies (McDermott 2009; Regensteiner 1996; Ritta-Dias 2010; Sanderson 2006; Treat-Jacobson 2009). These five randomised controlled trials published in peer-reviewed journals fulfilled the inclusion criteria, and we considered these for inclusion in this review (See: Characteristics of included studies). Five additional publications described the results of three of the primary studies (Regensteiner 1996; Ritta-Dias 2010; Treat-Jacobson 2009) more extensively.

Three trials compared supervised walking exercise to an exercise or progressive resistance regime (McDermott 2009; Regensteiner 1996; Ritta-Dias 2010); one trial compared arm ergometry to supervised walking exercise (Treat-Jacobson 2009); and one trial compared cycling exercise to supervised walking exercise (Sanderson 2006). These five trials randomised a total of 184 participants with IC, with 135 participants randomised to the treatment arms relevant to this review. The number of participants per study was small, with a variation of between 29 and 45 participants. The mean age of the participants in the included trials varied between 62.0 and 71.7 years, and all trials included both men and women. The trials were conducted in the United States (3), Brazil (1), and Australia (1).

Enrolment criteria were rather homogeneous. All trials included participants if a declination in ankle brachial index was present with coinciding limiting or disabling symptoms of IC. One trial (McDermott 2009) assessed claudication symptoms by a questionnaire (San Diego Claudication Questionnaire). In all trials, the presence of critical limb ischaemia was an exclusion criterion. Participants were also excluded if the exercise capacity was limited by another factor than IC (e.g. angina, chronic obstructive pulmonary disease, arthrosis). In three of the five trials, claudication symptoms needed to be stable for, respectively, three months (Regensteiner 1996), six months (Ritta-Dias 2010), or more than 12 months (Sanderson 2006). Two trials (Regensteiner 1996; Ritta-Dias 2010) excluded participants if a revascularisation procedure was performed in the previous year. One RCT (Treat-Jacobson 2009) excluded participants if a coronary or lower-extremity revascularisation procedure was performed within the past three months. The two remaining trials (McDermott 2009; Sanderson 2006) excluded participants if they recently underwent surgery or a cardiovascular event.

Treatment duration varied between studies ranging from six weeks' training (Sanderson 2006), 12-week training (Regensteiner 1996; Ritta-Dias 2010; Treat-Jacobson 2009), to 24-week training periods (McDermott 2009).

Excluded studies

After title/abstract screening, we excluded 3626 studies. After full-text assessment, we excluded another 18 studies for various reasons (See: Characteristics of excluded studies). We excluded three studies as they were not randomised controlled trials (Gardner 2011; Kim 2006; Roitman 2010). We excluded four studies (Dedes 2010; Kuwabara 2010; Ornelas 2011; Treat-Jacobson 2012) as they were meeting posters with a limited description of the methods and results; no articles of these meeting posters were published yet. We excluded nine studies (Nawaz 2001; Parr 2009; Saxton 2008; Saxton 2011; Tebbutt 2011; Treat-Jacobson 2012; Walker 2000; Wang 2008; Zwierska 2005) because they did not assess (adequate) supervised exercise therapy according to our inclusion criteria for this review. Five studies (Nawaz 2001; Saxton 2008; Saxton 2011; Treat-Jacobson 2011; Walker 2000) did not report the primary and secondary outcome measures of this review. We excluded one study (Jones 1996) because the outcome measures were unclearly described. We tried to contact the authors but did not receive additional information. We excluded one study (Collins 2012) because we did not consider the intervention (pole walking) an alternative exercise regime in comparison to supervised walking.

Risk of bias in included studies

See Figure 2 for a summary of the risk of bias in each included study.

Figure 2.

'Risk of bias' summary: review authors' judgements about each 'Risk of bias' item for each included study

Allocation

Two studies (McDermott 2009; Ritta-Dias 2010) described adequate sequence generation and allocation concealment by means of computer randomisation. Two studies (Treat-Jacobson 2009; Sanderson 2006) described an adequate sequence generation, but did not describe the allocation concealment. One study (Regensteiner 1996) did not describe the randomisation process.

Blinding

In the included studies, participants and direct personnel could not be blinded to the intervention (exercise). For this reason, bias could be introduced. However, since all studies experienced the same limitation, we considered the risk of bias to be low for all studies. Detection bias can be avoided by blinding the outcome assessors. Three studies achieved this (McDermott 2009; Ritta-Dias 2010; Treat-Jacobson 2009).

Incomplete outcome data

Most studies well-described reasons for missing data, and we considered these plausible and well distributed among the groups. One study (McDermott 2009) addressed asymptomatic participants with PAOD as well as symptomatic participants. After contacting the authors, we derived the data of the sole symptomatic group with IC. However, it is unclear which of the missing data described in the article related to the symptomatic group. For this reason, we considered the incomplete outcome data in this study to be unclear.

Selective reporting

All studies described relevant outcomes. For three studies (McDermott 2009; Sanderson 2006; Treat-Jacobson 2009), we retrieved additional outcome data by contacting the authors.

Other potential sources of bias

Since we identified only five studies, we could not detect publication bias with a funnel plot. Additional Begg and Egger tests did also exclude any indication of publication bias (Begg: adjusted Kendall's score = 4, P = 0.462; Egger: bias 1.11, 95% confidence interval (CI) 0.43 to -2.77, P = 0.430).

One of the studies (McDermott 2009) did not describe the baseline characteristics of the subgroup of participants with intermittent claudication. This was due to the study setting. (The study included participants with asymptomatic as well as symptomatic peripheral arterial disease.) However, we identified no other potential sources of bias in the included trials.

Effects of interventions

Walking exercise versus alternative exercise

Maximum walking distance [METs]

Data for the maximum walking distance (MWD) obtained at the end of each study were available in all of the included trials, with a total sample size of 135 participants. We calculated the pooled treatment effect after standardising the reported walking distances. For this reason, we converted all distances to metabolic equivalents (METs) using the American College of Sports Medicine (ACSM) formulas for metabolic calculations. We considered the impact of heterogeneity as low with an I² statistic of 13%. At the end of training, the pooled MWD increased with an overall non-significant effect size of 8.15 METs (95% confidence interval (CI) -2.63 to 18.94, P = 0.14) in favour of walking exercise. This is the equivalent of an increase of 173 metres (95% CI -56 to 401 metres) on a treadmill with no incline and an average speed of 3.2 km/h, which is comparable with walking in daily life (Analysis 1.1). However, because of the width of the CI of the pooled MWD, we could not identify any clear evidence of difference between interventions.

Furthermore, we calculated the effect size after 12 weeks of training. After this follow-up period, data from three trials on MWD were available (Regensteiner 1996; Ritta-Dias 2010; Treat-Jacobson 2009), with a sample size of 74 participants. We considered the impact of heterogeneity as moderate to substantial with an I² statistic of 52%. In these trials, the pooled MWD increased with a non-significant effect size of 13.05 METs (95% CI -11.43 to 37.54, P = 0.30) in favour of walking exercise. This correlates with 276 metres (95% CI -242 to 795 metres) on a treadmill with no incline and an average speed of 3.2 km/h (Analysis 1.2).

Pain-free walking distance [METs]

Data for the pain-free walking distance (PFWD) obtained at the end of each study were available in four of the five included trials (McDermott 2009; Ritta-Dias 2010; Sanderson 2006; Treat-Jacobson 2009), with a total sample size of 116 participants. We calculated the effect after converting the reported walking distances to standardised METs. We considered the impact of heterogeneity as moderate with an I² statistic of 49%. At the end of training, the PFWD increased with an overall non-significant effect size of 6.42 METs (95% CI -1.52 to 14.36, P = 0.11) in favour of walking exercise. This is the equivalent of an increase of 136 metres (95% CI -32 to 304 metres) on a treadmill with no incline and an average speed of 3.2 km/h, which is comparable with walking in daily life (Analysis 1.3). However, because of the width of the CI of the pooled PFWD, we could not identify any clear evidence of a difference between interventions.

Furthermore, we calculated the effect size after 12 weeks of training. After this follow-up period, data for two trials for PFWD (Ritta-Dias 2010; Treat-Jacobson 2009) were available, with a sample size of 55 participants. We considered the impact of heterogeneity as low with an I² statistic of 0%. In these trials, the pooled PFWD increased with a non-significant effect size of 2.37 METs (95% CI -5.81 to 10.56, P = 0.57) in favour of walking exercise. This correlates with 50 metres (95% CI -123 to 224 metres) on a treadmill with no incline and an average speed of 3.2 km/h (Analysis 1.4).

Health-related quality of life (HR-QoL) and functional impairment

Two of the included studies (McDermott 2009; Regensteiner 1996) described HR-QoL and functional impairment. One study (Regensteiner 1996) used the Medical Outcomes Study (MOS) SF-20; another study (McDermott 2009) used the SF-36 Physical Functioning score to collect HR-QoL data. Both studies used the Walking Impairment Questionnaire (WIQ) to describe functional impairment.

Data on HR-QoL and functional impairment (WIQ) from one study (Regensteiner 1996), with a total sample size of 19 participants, showed an increase in HR-QoL with a significant effect size of 26.50% (95% CI 2.67 to 50.33, P = 0.03) in favour of walking exercise. WIQ distance score increased with an effect size of 2.00% (95% CI -16.04 to 20.04, P = 0.83) in favour of walking exercise. WIQ speed score decreased with an effect size of -4.50% (95% CI -27.34 to 18.34, P = 0.70) in favour of alternative exercise. WIQ stair-climbing score decreased with an effect size of -29.50% (95% CI -51.65 to -7.35, P = 0.009) in favour of alternative exercise.

Unfortunately, both the SF-36 Physical Functioning score and WIQ data from the second study (McDermott 2009) were not normally distributed. SF-36 Physical Functioning score improved in both the strength (n = 14) and treadmill walking group (n = 17), with a median of, respectively, 12.5 points (interquartile range = -5.00 to 20.00) and 10.0 points (interquartile range = 5.00 to 20.00), P = 0.811. WIQ distance score improved in both the strength (n = 15) and treadmill walking group (n = 15), with a median of, respectively, 14.0 points (interquartile range = 1.56 to 26.6) and 7.46 points (interquartile range = -0.36 to 25.0), P = 0.431. WIQ speed score improved in both the strength (n = 15) and treadmill walking group (n = 16), with a median of, respectively, 3.26 points (interquartile range = -7.61 to 26.1) and 1.63 points (interquartile range = -3.80 to 28.8), P = 0.736. WIQ stair-climbing score improved in the strength training group (n = 15), with a median of, respectively, 12.5 points (interquartile range = 4.17 to 25.0), while we saw no improvement in the median score in the treadmill walking group (n = 16, median score of 0.00 points, interquartile range = 0.00 to 14.6), P = 0.136.

Because of the skewed data of one of the studies (McDermott 2009) and the small sample size of both studies, we did not transform these data to perform a meta-analysis of the two studies.

Sensitivity analysis

We performed a sensitivity analysis to assess whether excluding an individual study would significantly change the main results on MWD and PFWD at the end of the study. For MWD, removing any individual study did not alter the results significantly, although by removing one study (Ritta-Dias 2010), the MWD increased almost significantly, in favour of walking exercise, with an overall effect size of 10.16 METs (95% CI -0.40 to 20.71, P = 0.06). This is the equivalent of an increase of 215 metres (95% CI -8 to 439 metres) on a treadmill with no incline and an average speed of 3.2 km/h. For PFWD, removing one of the studies did not alter the results significantly, although by removing one study (Ritta-Dias 2010), the PFWD increased almost significantly in favour of walking exercise, with an overall effect size of 8.30 METs (95% CI -0.26 to 16.86, P = 0.06). This is the equivalent of an increase of 176 metres (95% CI -6 to 357 metres) on a treadmill with no incline and an average speed of 3.2 km/h.

We did not perform a sensitivity analysis on the methodological quality of the studies because of the limited number of studies.

Walking exercise versus strength training

Maximum walking distance [METs]

Data for MWD obtained at the end of each study were available in three of the five included trials, with a total sample size of 82 participants (McDermott 2009; Regensteiner 1996; Ritta-Dias 2010). We considered the impact of heterogeneity as moderate to substantial with an I² statistic of 52%. At the end of the study, the pooled MWD increased with an overall non-significant effect size of 8.11 METs (95% CI -11.68 to 27.90, P = 0.42) in favour of walking exercise. This is the equivalent of an increase of 172 metres (95% CI -247 to 591 metres) on a treadmill with no incline and an average speed of 3.2 km/h, which is comparable with walking in daily life (Analysis 2.1).

Furthermore, we calculated the effect size after 12 weeks of training. After this follow-up period, data for two trials for MWD were available (Regensteiner 1996; Ritta-Dias 2010), with a sample size of 49 participants. We considered the impact of heterogeneity as substantial with an I² statistic of 76%. The pooled MWD increased with a non-significant effect size of 14.10 METs (95% CI -32.10 to 60.29, P = 0.55) in favour of walking exercise. This correlates with 299 metres (95% CI -680 to 1277 metres) on a treadmill with no incline and an average speed of 3.2 km/h (Analysis 2.2).

Pain-free walking distance [METs]

Data for the PFWD at the end of the study were available in two of the included trials, with a total sample size of 63 participants (McDermott 2009; Ritta-Dias 2010). We considered the impact of heterogeneity as low with an I² statistic of 0%. At the end of the study, the pooled PFWD increased with an overall non-significant effect size of 2.66 METs (95% CI -5.06 to 10.39, P = 0.50) in favour of walking exercise. This is the equivalent of an increase of 56 metres (95% CI -107 to 220 metres) on a treadmill with no incline and an average speed of 3.2 km/h, which is comparable with walking in daily life (Analysis 2.3).

Walking exercise versus a combination of exercise modes

Maximum walking distance [METs]

Two trials (Regensteiner 1996; Treat-Jacobson 2009) analysed the effect of a combination of exercise modes in relation to supervised walking exercise on the MWD. One trial (Treat-Jacobson 2009) compared a combination of walking exercise and arm ergometry to sole walking exercise. The second trial (Regensteiner 1996) compared a combination of walking exercise and strength training to sole walking exercise. In total, the two trials had a sample size of 41 participants for analysis of the MWD at the end of the studies. We considered the impact of heterogeneity as low with an I² statistic of 0%. The pooled MWD increased with an overall non-significant effect size of 9.31 METs (95% CI -12.77 to 31.40, P = 0.41) in favour of sole walking exercise. This is the equivalent of an increase of 197 metres (95% CI -270 to 665 metres) on a treadmill with no incline and an average speed of 3.2 km/h, which is comparable with walking in daily life (Analysis 3.1).

Pain-free walking distance [METs]

Data for PFWD obtained at the end of each study were available in one of the included trials with a total sample size of 25 participants (Treat-Jacobson 2009). At the end of this study, the PFWD increased with an overall non-significant effect size of 3.30 METs (95% CI -6.01 to 12.61, P = 0.49) in favour of sole walking exercise. This is the equivalent of an increase of 70 metres (95% CI -127 to 267 metres) on a treadmill with no incline and an average speed of 3.2 km/h, which is comparable with walking in daily life.

Discussion

Summary of main results

We included five randomised controlled trials (RCTs) in this review, with a total of 135 participants. We could not find any clear evidence of a statistical difference in maximum or pain-free walking distance when comparing supervised treadmill walking with an alternative exercise regime. Furthermore, sensitivity analysis did not significantly alter the results. Regarding quality of life, two studies (McDermott 2009; Regensteiner 1996) described health-related quality of life and functional impairment. However, the data in one of the two studies (McDermott 2009) were not normally distributed. Therefore, we could not make any meaningful comparison between studies.

Overall completeness and applicability of evidence

Although the topic of this review is contemporary (Parmenter 2011), this review identified only five RCTs, with a total of 135 participants. The five RCTs described only three alternative exercise modes: strength training (McDermott 2009; Regensteiner 1996; Ritta-Dias 2010), arm ergometry (Treat-Jacobson 2009), and cycling exercise (Sanderson 2006). In total, three of the five studies (McDermott 2009; Regensteiner 1996; Ritta-Dias 2010) compared strength training to supervised treadmill walking, for which we performed a subanalysis. Eventually, more studies are needed to make meaningful comparisons between each alternative exercise mode and the current standard of supervised treadmill walking. Therefore, the applicability of the current evidence is limited.

Quality of the evidence

The risk of bias of the included studies was, in general, low (see Figure 2), reflecting good methodological quality of the included studies. We could not detect publication bias because we could not assess asymmetry in a funnel plot with the limited number of studies. The quality of this review is however limited because of the small sample size of 135 participants. At analysis, we experienced some heterogeneity in Analysis 1.2, Analysis 2.1, and Analysis 2.2 using the I² statistic and in Analysis 2.2 using the Q-statistic or Chi² test. Sensitivity analysis did not significantly alter the results of the review.

Potential biases in the review process

We tried to limit all potential biases in the review process. To limit bias and make a meaningful comparison, we standardised maximum walking distance (MWD) and pain-free walking distance (PFWD) from each study by converting walking distances to total metabolic equivalents (METs) according to the American College of Sports Medicine (ACSM) formulas for metabolic calculations (ACSM 2006). However, direct conversion of the walking times or distances to METs was not possible due to the absence of individual participant data. We therefore simulated a new dataset for each study to make a meaningful comparison. It is unclear to what extent this could have biased our findings.

We excluded one study (Treat-Jacobson 2011) solely because it did not report the correct outcome measures. Although we were careful to ascertain that relevant outcomes were not available because they were not measured rather than not reported, this could be a potential bias in the review process. In future updates, we will pay further attention to this potential source of bias.

Agreements and disagreements with other studies or reviews

We agree with a previous published systematic review (Parmenter 2011) that there was no clear evidence of difference between alternative exercise modes and supervised walking exercise for intermittent claudication. Results seem promising, but additional studies are urgently needed to validate these exercise modes in relation to the standard of supervised walking exercise.

Authors' conclusions

Implications for practice

Few studies were found and there was no clear evidence of a difference between supervised walking exercise and alternative exercise modes. More studies with larger sample sizes are needed to make meaningful comparisons between each alternative exercise mode and the current standard of supervised treadmill walking. The results indicate that alternative exercise modes may be useful when supervised walking exercise is not an option for the patient.

Implications for research

More studies are urgently needed to make meaningful comparisons between each alternative exercise mode and the current standard of supervised treadmill walking.

Acknowledgements

We would like to thank Dr Karen Welch for searching the Cochrane Peripheral Vascular Diseases Group Specialised Register and the Cochrane Central Register of Controlled Trials (CENTRAL) in The Cochrane Library (www.thecochranelibrary.com).

We would like to thank Dr Demián Glujovsky for enabling us to use EROS (Early Review Organizing Software http://www.eros-systematic-review.org/) in our reviewing process.

Data and analyses

Download statistical data

Comparison 1. Walking exercise versus alternative exercise
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Maximum walking distance at the end of training5135Mean Difference (IV, Random, 95% CI)8.15 [-2.63, 18.94]
2 Maximum walking distance after 12 weeks of training374Mean Difference (IV, Random, 95% CI)13.05 [-11.43, 37.54]
3 Pain-free walking distance at the end of training4116Mean Difference (IV, Random, 95% CI)6.42 [-1.52, 14.36]
4 Pain-free walking distance after 12 weeks of training255Mean Difference (IV, Random, 95% CI)2.37 [-5.81, 10.56]
5 Change in health-related quality of life at the end of training1 Mean Difference (IV, Random, 95% CI)Totals not selected
6 Change in functional impairment (WIQ distance score) at the end of training1 Mean Difference (IV, Random, 95% CI)Totals not selected
7 Change in functional impairment (WIQ speed score) at the end of training1 Mean Difference (IV, Random, 95% CI)Totals not selected
8 Change in functional impairment (WIQ stair score) at the end of training1 Mean Difference (IV, Random, 95% CI)Totals not selected
Analysis 1.1.

Comparison 1 Walking exercise versus alternative exercise, Outcome 1 Maximum walking distance at the end of training.

Analysis 1.2.

Comparison 1 Walking exercise versus alternative exercise, Outcome 2 Maximum walking distance after 12 weeks of training.

Analysis 1.3.

Comparison 1 Walking exercise versus alternative exercise, Outcome 3 Pain-free walking distance at the end of training.

Analysis 1.4.

Comparison 1 Walking exercise versus alternative exercise, Outcome 4 Pain-free walking distance after 12 weeks of training.

Analysis 1.5.

Comparison 1 Walking exercise versus alternative exercise, Outcome 5 Change in health-related quality of life at the end of training.

Analysis 1.6.

Comparison 1 Walking exercise versus alternative exercise, Outcome 6 Change in functional impairment (WIQ distance score) at the end of training.

Analysis 1.7.

Comparison 1 Walking exercise versus alternative exercise, Outcome 7 Change in functional impairment (WIQ speed score) at the end of training.

Analysis 1.8.

Comparison 1 Walking exercise versus alternative exercise, Outcome 8 Change in functional impairment (WIQ stair score) at the end of training.

Comparison 2. Walking exercise versus strength training
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Maximum walking distance at the end of training382Mean Difference (IV, Random, 95% CI)8.11 [-11.68, 27.90]
2 Maximum walking distance at 12 weeks of training249Mean Difference (IV, Random, 95% CI)14.10 [-32.10, 60.29]
3 Pain-free walking distance at the end of training263Mean Difference (IV, Random, 95% CI)2.66 [-5.06, 10.39]
Analysis 2.1.

Comparison 2 Walking exercise versus strength training, Outcome 1 Maximum walking distance at the end of training.

Analysis 2.2.

Comparison 2 Walking exercise versus strength training, Outcome 2 Maximum walking distance at 12 weeks of training.

Analysis 2.3.

Comparison 2 Walking exercise versus strength training, Outcome 3 Pain-free walking distance at the end of training.

Comparison 3. Walking exercise versus a combination of exercise modes
Outcome or subgroup titleNo. of studiesNo. of participantsStatistical methodEffect size
1 Maximum walking distance at the end of the study241Mean Difference (IV, Random, 95% CI)9.31 [-12.77, 31.40]
2 Pain-free walking distance at the end of the study1 Mean Difference (IV, Random, 95% CI)Totals not selected
Analysis 3.1.

Comparison 3 Walking exercise versus a combination of exercise modes, Outcome 1 Maximum walking distance at the end of the study.

Analysis 3.2.

Comparison 3 Walking exercise versus a combination of exercise modes, Outcome 2 Pain-free walking distance at the end of the study.

Appendices

Appendix 1. CENTRAL search strategy

#1MeSH descriptor: [Arteriosclerosis] this term only894
#2MeSH descriptor: [Arteriolosclerosis] this term only0
#3MeSH descriptor: [Arteriosclerosis Obliterans] this term only72
#4MeSH descriptor: [Atherosclerosis] this term only407
#5MeSH descriptor: [Arterial Occlusive Diseases] this term only766
#6MeSH descriptor: [Intermittent Claudication] this term only720
#7MeSH descriptor: [Ischemia] this term only764
#8MeSH descriptor: [Peripheral Vascular Diseases] explode all trees2181
#9atherosclero* or arteriosclero* or PVD or PAOD or PAD17383
#10(arter* or vascular or vein* or veno* or peripher*) near (occlus* or reocclus* or re-occlus* or steno* or obstruct* or lesio* or block* or harden* or stiffen*)7989
#11peripheral near/3 dis*3299
#12claudic* or IC3417
#13isch* or CLI17025
#14#1 or #2 or #3 or #4 or #5 or #6 or #7 or #8 or #9 or #10 or #11 or #12 or #1341307
#15MeSH descriptor: [Exercise] explode all trees9253
#16MeSH descriptor: [Physical Therapy Modalities] explode all trees13193
#17MeSH descriptor: [Physical Exertion] this term only3168
#18MeSH descriptor: [Sports] explode all trees7474
#19MeSH descriptor: [Leisure Activities] this term only162
#20MeSH descriptor: [Fitness Centers] this term only19
#21MeSH descriptor: [Physical Exertion] this term only3168
#22(physical near/3 (exercise* or exertion or endurance or therap* or conditioning or activit* or fitness or train*)):ti,ab,kw17049
#23(exercise near/3 (train* or intervention* or protocol* or program* or therap or activit* or regim*)):ti,ab,kw7249
#24(fitness near/3 (train* or intervention* or protocol* or program* or therap* or activit* or regim* or centre* or center*)):ti,ab,kw944
#25((training or conditioning) near/3 (circuit or intervention* or protocol* or program* or activit* or regim*)):ti,ab,kw6667
#26(walk* or run* or treadmill or aerobic or swim* or danc* or weight or squat* or lunge or bend* or raise or cycling or step):ti,ab,kw81770
#27kinesiotherap*:ti,ab,kw207
#28((endurance or aerobic or cardio*) near/3 (fitness or train* or intervention* or protoco* or program* or therap* or activit* or regim*)):ti,ab,kw6185
#29#15 or #16 or #17 or #18 or #19 or #20 or #21 or #22 or #23 or #24 or #25 or #26 or #27 or #28107265
#30#14 and #29 in Trials6781

Appendix 2. Authors' MEDLINE search strategy (1946 to 15 July 2013)

#1exp intermittent claudication6963
#2exp peripheral arterial disease/1601
#3peripheral vascular diseases {No Related Terms}11222
#4(peripheral obstructive art* or iliac art* or femoropop* or claudic$ or dysvascular$ or peripheral vascular disease or leg isch?mia).ab,ti.24285
#5((arter$ or vascu$ or peripher$) adj5 (obstruct$ or occlus$ or steno$ or isch?em*)).ab,ti.97389
#6or/1-5126990
#7exp exercise therapy/29123
#8exp exercise/109247
#9(sport* or walk* or exercise* or rehabil$ or training$ or activit$).ab,ti.2593935
#10or/7-92624448
#11randomized controlled trial.pt.378440
#12controlled clinical trial.pt.88571
#13randomized.ab.275755
#14randomly.ab.193433
#15(trial or groups).ab.1457559
#16or/11-151808082
#17exp animals/ not humans.sh.3998626
#1816 not 171475758
#196 and 10 and 182434

Appendix 3. Authors' Embase search strategy (1973 to 15 July 2013)

#1exp peripheral occlusive artery disease/115535
#2peripheral vascular disease {No Related Terms}18261
#3leg ischemia {No Related Terms}10809
#4iliac artery obstruction {No Related Terms}26448
#5(peripheral obstructive art* or iliac art* or femoropop* or claudic$ or dysvascular$ or peripheral vascular disease or leg isch?mia).ab,ti.32163
#6((arter$ or vascu$ or peripher$) adj5 (obstruct$ or occlus$ or steno$ or isch?em*)).ab,ti.131292
#7or/1-6252382
#8exp kinesiotherapy/45442
#9exp exercise/198194
#10(sport* or walk* or exercise* or rehabil$ or training$ or activit$).ab,ti.3159880
#11or/8-103222285
#12(random$ or factorial$).ab,ti.848921
#13((doubl? or singl?) adj1 blind$).ab,ti.156646
#14assign$.ab,ti.229137
#15allocat$.ab,ti.78470
#16double-blind procedure/118323
#17randomized controlled trial/349997
#18single-blind procedure/17719
#19or/12-181158395
#207 and 11 and 192747
#21exp ANIMAL/ or NONHUMAN/ or exp ANIMAL EXPERIMENT/20113789
#22exp HUMAN/14647994
#2321 not 225465795
#2420 not 232087

Contributions of authors

  • Gert-Jan Lauret (GJL) wrote the protocol, selected relevant trials, assessed trial quality, extracted data, and wrote the review.

  • Farzin Fahkry (FF) contributed to the protocol, selected relevant trials, assessed trial quality, and extracted data.

  • Hugo Fokkenrood (HF) contributed to the protocol and to the text of the review.

  • Myriam Hunink (MH) contributed to the protocol and to the text of the review.

  • Joep Teijink (JT) contributed to the protocol, confirmed suitability of selected trials for inclusion, and contributed to the text of the review.

  • Sandra Spronk (SS) commented and contributed to the protocol, confirmed suitability of selected trials for inclusion, was consulted where disagreements occurred, checked data for accuracy, and contributed to the text of the review.

Declarations of interest

GJL: nothing to declare.
FF: nothing to declare.
HF: nothing to declare.
MH: MH's institution has received funding from ZonMW, Netherlands Organization for Scientific Research, National Institutes of Health, and Stichting Technische Wetenschappen for MH's research projects not related to this review. MH also reports receiving royalties from Cambridge University Press from the textbook 'Decision Making in Health and Medicine' and travel/meeting expenses from the International Society for Strategic Studies in Radiology (ISSSR), the European Society of Radiology (ESR) and the European Institute of Biomedical Imaging Research (EIBIR).
JT: nothing to declare.
SS: nothing to declare.

Sources of support

Internal sources

  • No sources of support supplied

External sources

  • Chief Scientist Office, Scottish Government Health Directorates, The Scottish Government, UK.

    The PVD Group editorial base is supported by the Chief Scientist Office.

Differences between protocol and review

  • We altered the primary outcome from mean change in maximum or pain-free walking distance to a postintervention maximum or pain-free walking distance. After extracting the data, most of the included trials only reported pre-intervention and postintervention walking distances and their variances. No data on mean change were reported. Although calculating the mean change in walking distance from the pre-intervention and postintervention data was possible, calculating a variance (standard deviation) of the mean change without having individual participant data was more challenging. To calculate a variance of the mean change from the reported summary pre-intervention and postintervention variances, we at least needed the correlation between the pre-intervention and postintervention variance (Higgins 2011), which was not reported. Furthermore, as we only included randomised trials and no significant difference in walking distance between the intervention and control group in each trial existed at baseline, we did not have any indication that our main outcome, the mean difference in postintervention walking distance between the intervention and control group might be biased.

  • We changed the P value that was considered statistically significant in case of heterogeneity from 0.05 in our protocol to 0.10 in our full review.

  • In the Background, we adjusted the description of the systematic review by Parmenter and colleagues (Parmenter 2011) after they contacted us and added why no meta-analysis was performed in this systematic review.

Characteristics of studies

Characteristics of included studies [ordered by study ID]

McDermott 2009

MethodsRCT
Participants1009 participants with asymptomatic and symptomatic PAOD were assessed for eligibility; 156 participants were randomised; after contacting authors, 33 randomised participants were affected by intermittent claudication
Interventions

Group 1: 24 weeks of supervised treadmill walking

Group 2: 24 weeks of supervised lower-extremity resistance training

Group 3 (control group): 11 nutritional information sessions over 6 months

Outcomes6-minute walk test, Short physical performance battery, Brachial artery flow-mediated dilation, Physical activity (accelerometry), Maximum treadmill walking time, Treadmill time to onset of leg symptoms, SF-36 physical functioning score, Walking Impairment Questionnaire, Knee extension isometric strength/power, Plantarflexion isometric strength
NotesWe contacted the authors for more information regarding study results for the subgroup of participants with intermittent claudication
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskParticipants were randomised by computer using a randomly permuted block method. Randomisation was stratified by the presence versus absence of intermittent claudication
Allocation concealment (selection bias)Low riskParticipants were randomised by computer using a randomly permuted block method
Blinding of participants and personnel (performance bias)
All outcomes
Low riskIn all of the included studies, participants and direct personnel could not be blinded to the intervention (exercise). For this reason, bias could be introduced. However, since all studies experienced the same limitation, we considered the risk of bias to be low for all studies
Blinding of outcome assessment (detection bias)
All outcomes
Low riskExaminers were blinded to participant group assignment
Incomplete outcome data (attrition bias)
All outcomes
Unclear riskIntention-to-treat analyses were performed for dropouts. Analyses were repeated using multiple imputation for persons who died or dropped out before completing 6-month follow-up testing. However, because of the study setting, it is not clear which participants with intermittent claudication (instead of asymptomatic peripheral arterial disease) dropped out. The paper stated that missing data at follow-up were more common in more frail participants. On the other hand, the authors performed a sensitivity analysis and showed that the missing data were not likely to have significantly altered the findings
Selective reporting (reporting bias)Low riskAll relevant outcomes were described. Although not all secondary outcomes were mentioned in the trial registration, all outcomes mentioned in the study protocol were reported in the final draft of the paper
Other biasUnclear riskBecause of the study setting (the study included participants with asymptomatic or symptomatic peripheral arterial disease), the baseline characteristics of the subgroup of participants with intermittent claudication were not described. We identified no other forms of bias

Regensteiner 1996

MethodsRCT
Participants44 participants evaluated: 15 were excluded before randomisation; 29 participants were enrolled and randomised
Interventions

Group 1: 12 weeks of supervised walking exercise. Secondly, 12 weeks of additional supervised walking exercise

Group 2: 12 weeks of strength training. Secondly, 12 weeks of additional supervised walking exercise

Group 3 (control group): no treatment for 12 weeks. Secondly, 12 weeks of a combination of supervised walking exercise and strength training

OutcomesPeak treadmill walking time, Ankle brachial indices (in rest and after exercise), Walking Impairment Questionnaire score, Physical Activity Recall score, Medical Outcomes Study Questionnaire score, Vitalog activity monitor
Notes-
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Unclear riskNot described
Allocation concealment (selection bias)Unclear riskNot described
Blinding of participants and personnel (performance bias)
All outcomes
Low riskIn all of the included studies, participants and direct personnel could not be blinded to the intervention (exercise). For this reason, bias could be introduced. However, since all studies experienced the same limitation, we considered the risk of bias to be low for all studies
Blinding of outcome assessment (detection bias)
All outcomes
Unclear risk

To minimise the potential for bias in the self-evaluation of walking ability and functional status, questionnaires were administered before the treadmill test. Thus, participants' questionnaire responses were not influenced by their treadmill exercise performance. In addition, the interviewer and participant were blinded to previous questionnaire scores.

The paper did not describe whether examiners were blinded to participant group assignment

Incomplete outcome data (attrition bias)
All outcomes
Low risk2 participants in the control group were not available for follow-up at the 12-week evaluation. They were excluded from further analysis. In the supervised walking and strength training group, no missing outcome data were described
Selective reporting (reporting bias)Low riskAll relevant outcomes were described
Other biasLow riskNo significant differences in baseline characteristics were found. We identified no other forms of bias

Ritta-Dias 2010

MethodsRCT
Participants34 participants were randomised; 4 participants did not complete training; 30 participants completed the study protocol (15 per group)
Interventions

Group 1: 12 weeks of supervised treadmill exercise

Group 2: 12 weeks of strength training

OutcomesInitial claudication distance, Total walking distance, Peak VO₂, VO₂ at the first stage of treadmill test, Ischaemic window, Leg strength with lower ABI, Leg strength with higher ABI
NotesNo intention-to-treat analysis
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskRandomisation was performed by computer random number generation
Allocation concealment (selection bias)Low riskRandomisation was performed by computer random number generation
Blinding of participants and personnel (performance bias)
All outcomes
Low riskIn all of the included studies, participants and direct personnel could not be blinded to the intervention (exercise). For this reason, bias could be introduced. However, since all studies experienced the same limitation, we considered the risk of bias to be low for all studies
Blinding of outcome assessment (detection bias)
All outcomes
Low riskParticipants were evaluated at baseline (pre-training) and after 12 weeks of exercise training (post-training) by a physician who was blinded to the exercise programme performed by participants
Incomplete outcome data (attrition bias)
All outcomes
Low risk

4 participants (2 in the strength training groups and 2 in the walking exercise groups) did not complete the training programmes for the following reasons: inguinal hernia (n = 1), gastrointestinal infection (n =1), ongoing treatment for lung cancer (n = 1), and diagnosis of abdominal aneurysm (n = 1)

Although no intention-to-threat analysis was performed, reasons for missing data were plausible and well distributed among intervention groups

Selective reporting (reporting bias)Low riskAll relevant outcomes were described in the study results
Other biasLow riskNo significant differences in baseline characteristics were found. We identified no other forms of bias

Sanderson 2006

MethodsRCT
Participants694 participants were assessed for eligibility; 43 participants were randomised
Interventions

Group 1: 6 weeks of supervised treadmill walking

Group 2: 6 weeks of supervised cycling

Group 3 (control group): cardiovascular risk management and exercise advice

OutcomesMaximum walking time, Pain-free walking time, Maximum cycling time, Pain-free cycling time, Submaximal and peak heart rate/VO₂/respiratory exchange ratio/minute ventilation
NotesWe contacted the authors and received relevant outcome data, which were not clearly described in the paper
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskA closed-envelope system was used to randomise participants from the stratified groups to a control group, a cycle-training group, or a treadmill training group
Allocation concealment (selection bias)Unclear riskNot described
Blinding of participants and personnel (performance bias)
All outcomes
Low riskIn all of the included studies, participants and direct personnel could not be blinded to the intervention (exercise). For this reason, bias could be introduced. However, since all studies experienced the same limitation, we considered the risk of bias to be low for all studies
Blinding of outcome assessment (detection bias)
All outcomes
Unclear riskNo blinding of outcome assessment was described
Incomplete outcome data (attrition bias)
All outcomes
Low risk1 participant in the treadmill training group withdrew after 1 week of training. The baseline data have been omitted
Selective reporting (reporting bias)Low riskRelevant outcome data for this review (maximum and pain-free walking time) were only presented in a figure; no numerical data were given. However, after contacting the authors, we collected these data
Other biasLow riskNo significant differences in baseline characteristics were found. We identified no other forms of bias

Treat-Jacobson 2009

  1. a

    ABI: ankle brachial index.
    RCT: randomised controlled trial.

MethodsRCT
Participants102 participants were assessed for eligibility; 45 participants were randomised
Interventions

Group 1: 12 weeks of arm ergometry

Group 2: 12 weeks of supervised treadmill walking

Group 3: 12 weeks of a combination of arm ergometry and supervised treadmill walking

Group 4 (control group): usual care (cardiovascular management and exercise advice)

OutcomesMaximum walking distance, Pain-free walking distance, Resting ankle-brachial index/heart rate/blood pressure, Functional capacity (Peak VO₂)
NotesThe study paper mentioned only the change in walking distances. To retrieve all data, we contacted the authors and received the exact outcome data at follow-up
Risk of bias
BiasAuthors' judgementSupport for judgement
Random sequence generation (selection bias)Low riskEligible participants were randomised by simple randomisation tables to 1 of the 4 study groups
Allocation concealment (selection bias)Unclear riskNot described
Blinding of participants and personnel (performance bias)
All outcomes
Low riskIn all of the included studies, participants and direct personnel could not be blinded to the intervention (exercise). For this reason, bias could be introduced. However, since all studies experienced the same limitation, we considered the risk of bias to be low for all studies
Blinding of outcome assessment (detection bias)
All outcomes
Low riskThe physician supervising the treadmill tests was blinded to treatment group assignment. However, other staff assisting with testing were not blinded. To mitigate for this limitation, care was taken to ensure standardisation of exercise-testing protocols
Incomplete outcome data (attrition bias)
All outcomes
Low riskIncomplete outcome data are well described and equally distributed among intervention groups. 4 of the 45 participants withdrew from the study before completing exercise training (2 participants in the arm- ergometry group and 2 participants in the treadmill walking group). Therefore, only 41 participants completed the 12-week follow-up. 31 participants completed the 24-week follow-up
Selective reporting (reporting bias)Low riskAll relevant outcomes were described in the study results
Other biasLow riskNo significant differences in baseline characteristics were found. We identified no other forms of bias

Characteristics of excluded studies [ordered by study ID]

StudyReason for exclusion
  1. a

    MWD: maximum walking distance.
    RCT: randomised controlled trial.
    SET: supervised exercise therapy.

Collins 2012Alternative exercise regime
Dedes 2010Meeting poster; no journal article available
Gardner 2011Commentary on other journal article; not a RCT
Jones 1996Outcome measures unclearly described; no information from authors
Kim 2006Commentary on other journal article; not a RCT
Kuwabara 2010Meeting poster; no article available
Nawaz 2001No intervention group with adequate supervised walking therapy; no correct outcome measures
Ornelas 2011Meeting poster; no article available
Parr 2009No intervention group with adequate supervised walking therapy
Roitman 2010Editorial; not a RCT
Saxton 2008No control group with adequate SET, no correct outcome measures
Saxton 2011No control group with adequate SET, no correct outcome measures (MWD assessed by shuttle-walk test instead of protocolised treadmill test)
Tebbutt 2011No control group with adequate SET
Treat-Jacobson 2011No correct outcome measures
Treat-Jacobson 2012Meeting poster; no journal article available; no control group with adequate SET
Walker 2000No control group with adequate SET; no correct outcome measures (no treadmill test)
Wang 2008No control group with SET
Zwierska 2005No control group with adequate SET

Characteristics of ongoing studies [ordered by study ID]

NCT00895635

  1. a

    mph: miles per hour.
    PAD: peripheral arterial disease.

Trial name or title

Exercise training to reduce claudication: arm ergometry versus treadmill walking.

Evaluating two exercise training programs to reduce leg pain in people with peripheral arterial disease (The EXERT study)

MethodsAllocation: randomised
End point classification: efficacy study
Intervention model: parallel assignment
Masking: single-blind (outcomes assessor)
Primary purpose: treatment
Participants

150 participants of both gender, 18 to 90 years old, who comply with the inclusion and exclusion criteria

Inclusion criteria

  • Has lifestyle-limiting claudication

  • Able to walk on a treadmill at 2 mph

  • Able to perform arm-ergometry exercise

  • Able to complete a 12-week exercise programme

Exclusion criteria

  • Physical activities are limited for reasons other than claudication

  • Uncontrolled high blood pressure

  • Uncontrolled diabetes

  • Unstable coronary heart disease

  • Ischaemic rest pain or tissue loss

  • Recent (in the 3 months before study entry) coronary or peripheral revascularisation

Interventions

Study arms:

Experimental: treadmill exercise training
Participants will take part in a 12-week supervised treadmill exercise training programme
Intervention: behavioural: treadmill exercise training

Experimental: arm-ergometry exercise training
Participants will take part in a 12-week supervised aerobic arm-ergometry exercise training programme
Intervention: behavioural: arm-ergometry exercise training

Active comparator: usual care control group
Participants will receive usual care for PAD from their doctor
Intervention: behavioural: usual care

Outcomes

Primary outcome measures

  • Maximal walking distance (time-frame: measured at baseline and weeks 6, 12, and 24)

  • Pain-free walking distance (time-frame: measured at baseline and weeks 6, 12, and 24)

Secondary outcome measures

  • Limb blood flow (time-frame: measured at baseline and weeks 6, 12, and 24)

  • Cardiovascular function (time-frame: measured at baseline and weeks 6, 12, and 24)

  • Quality of life (time-frame: measured at baseline and weeks 6, 12, and 24)

Starting dateJanuary 2009
Contact informationContact: Diane J Treat-Jacobson, PhD, RN612-624-7613treat001@umn.edu
Contact: Laura N Kirk, PhD, RN612-626-4687kirk0013@umn.edu
Notes-

Ancillary