Criteria for considering studies for this review
Types of studies
Parallel-group RCTs comparing supervised exercise programs with non- supervised exercise programs, in participants with PAD, were included. Trials were included irrespective of whether an intention- to-treat analysis was carried out.
Types of participants
Adults (18 years and older) with intermittent claudication (Fontaine II/Rutherford 1 to 3) (Fontaine 1954; Rutherford 1997) due to atherosclerotic disease, diagnosed by questionnaire or clinically, who were considered for conservative treatment, were included. Studies of participants with asymptomatic lower limb atherosclerosis identified by testing were excluded.
Types of interventions
Supervised exercise programs
A supervised exercise program needed to consist of more than six consecutive weeks of training, with more than 50% of the exercise time spent on walking or training the lower limbs. Inclusion of trials was not limited by frequency or intensity of the exercise program.
Non-supervised exercise programs
Non-supervised exercise programs were defined as those in which control group members were advised to walk (with or without a predefined exercise scheme or logbook) or were treated according to a structural home-based exercise program (walk advice with regular contact and exercise support from medically trained personnel). Studies with control groups, which did not receive exercise or walking advice or which received usual care (maintained normal physical activity), were excluded.
Types of outcome measures
The primary outcome measure was the effect on maximal treadmill walking distance or time. Outcomes had to be available at baseline.
The following secondary outcome measures were also considered:
- Pain-free treadmill walking distance, or time;
- Adherence to exercise program; and
- Quality of life.
Search methods for identification of studies
No restriction on language of publication was applied.
For this update, the Cochrane Peripheral Vascular Diseases Group Trials Search Coordinator (TSC) searched the Specialized Register (last searched September 2012) and the Cochrane Central Register of Controlled Trials (CENTRAL) 2012, Issue 9, part of The Cochrane Library (www.thecochranelibrary.com). See Appendix 1 for details of the search strategy used to search CENTRAL.
The Specialized Register is maintained by the TSC and is constructed from weekly electronic searches of MEDLINE, EM- BASE, CINAHL, and AMED, and through hand searching of 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 Specialized Register section of the Cochrane Peripheral Vascular Diseases Group module in The Cochrane Library (www.thecochranelibrary.com).
Searching other resources
A hand search of references lists of relevant studies was carried out.
Data collection and analysis
Selection of studies
Two review authors (HJPF and BLWB) independently selected trials for this review, whereas two other review authors (JAWT and MHP) confirmed the suitability of selected trials for inclusion in the review. Disagreements regarding inclusion/exclusion of selected trials were resolved by discussion.
Data extraction and management
Two review authors (HJPF and BLWB) independently extracted data using a standard data collection form. Data were entered into Review Manager (RevMan 2011). When necessary, additional information was sought for included trials. For all continuous outcomes (ie, walking distance or walking time, adherence to exercise program, quality of life), we extracted the number of participants and the mean difference between treatment groups per follow up period with accompanying standard deviation for each treatment group, if available. We also recorded other details of the included trials, for example, the study set- ting, the statistical power, and the number of dropouts in each group. We contacted authors for missing information regarding their methods.
Trials that compared three or more different therapy programs
- supervised and non-supervised therapy programs - with other kinds of programs, were included. From these trials, only participants treated by supervised or non- supervised exercise therapy were taken into account.
Assessment of risk of bias in included studies
Three review authors (HJPF, BLWB, and GJL) assessed and one review author (MHP) confirmed the methodological quality of the trials, primarily for adequacy of allocation concealment and follow-up. In trials that compared exercise programs with walking distance as the primary outcome, blinding of staff and participants is not possible and was therefore not considered.
We graded study quality in a table of risk of bias on the basis of a checklist of design components. This checklist comprised random sequence generation, allocation?? concealment, incomplete outcome data, and selective reporting. We achieved consensus by informal discussion. We summarized the adequacy of each category as "low", "unclear", or "high" risk according to the criteria provided in the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
Measures of treatment effect
We standardized reported data on walking time and distance to effect sizes to enable the calculation of an overall standardized mean difference (SMD). If standard errors (SEs) had been reported (and authors did not reply to our request to send unpublished data), we converted these to standard deviations. We used a fixed- effect model to obtain summary statistics for all types of outcome measures. When significant heterogeneity was found (P < 0.05), we calculated overall efficacy using random-effects models.
Unit of analysis issues
Only RCTs with follow-up of at least six weeks' duration were included in this review.
Dealing with missing data
We contacted the study authors to request any missing information. For each trial, we extracted the number of participants originally allocated to each treatment arm from the data, and we per- formed an intention-to-treat analysis.
Assessment of heterogeneity
We used both a Chi2 test and an I2 statistic to test for heterogeneity between trial results. We used the tests of heterogeneity at a significance level of alpha = 0.05 to examine whether observed variation in study results was compatible with the variation expected by chance alone. When possible, we pooled trial results by meta- analysis.
Assessment of reporting biases
We contacted authors to request missing data. To prevent language bias, we did not impose a language restriction. Funnel plots were used to identify potential publication bias.
If data were available, we performed statistical analyses using the Review Manager software (RevMan 2011). We standardized re- ported data on walking time and distance to effect sizes to enable the calculation of an overall standardized mean difference. For this purpose, the difference in outcomes of each study was expressed in the standard deviation (SD) observed in the study (Altman 1991). For articles with non-parametric data, SDs were calculated by dividing the SD by 1.35, according to Chapter 220.127.116.11 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011). We calculated the standardized mean difference (SMD) using a fixed- effect model, as the results were presented on different scales (distance or time).
Subgroup analysis and investigation of heterogeneity
Potential heterogeneity in non-supervised exercise programs might exist as the result of differences in intervention components (ranging from structural "go home and walk" advice to a formal home- based exercise program). We performed a subgroup analysis by di- viding non-supervised exercise programs into studies that treated participants with "go home and walk" advice and trials that used a home- based exercise program. Home-based exercise was defined as walking advice supplemented with some form of observation method (eg, exercise logs, telephone support, educational videos, step monitors) used by medically trained personnel. In studies that compared both non-supervised types with a supervised exercise program, the "shared" supervised group was split into two groups and included in two comparisons according to Chapter 16.5 of the Cochrane Handbook for Systematic Reviews of Interventions (Higgins 2011).
Individual study effects on the results were examined by removing each study one at a time to examine whether removing a particular study changed the results significantly. Studies with apparent risk of bias were excluded.