Skip to main content

A Systematic Review and Meta-analysis of Exercise Intervention for the Treatment of Calf Muscle Pump Impairment in Individuals with Chronic Venous Insufficiency

Empirical Studies

A Systematic Review and Meta-analysis of Exercise Intervention for the Treatment of Calf Muscle Pump Impairment in Individuals with Chronic Venous Insufficiency

Index: Ostomy Wound Manage. 2017;63(8):30-43. doi: 10.25270/owm.2017.08.3043


Exercise intervention for individuals with chronic venous insufficiency (CVI) and venous leg ulcers (VLUs) is recommended to improve function of the calf muscle pump (CMP). A systematic review with meta-analysis was conducted to measure the effects of exercise (including function of the CMP, ankle strength, range of motion [ROM], and healing rates) for VLUs. Four (4) databases (Cumulative Index to Nursing and Applied Health Literature, 1981; MEDLINE, 1964; Scopus, 1966; and EMBASE, 1947) were searched for relevant articles from the date of index inception to January 2016. All study types that evaluated the effect of exercise on the function of the CMP in patients with CVI were included; narrative and systematic studies were excluded.All data were extracted by 1 reviewer onto a predetermined form and verified by a second reviewer. Data extracted included number of patients, primary diagnosis, patient demographics, study location, wound characteristics, inclusion/exclusion criteria, exercise group details, control group details, co-interventions, primary outcome measures, secondary outcome measures, wound healing measures, blinding, intention to treat, and study design. A total of 1325 articles was screened; 14 met the inclusion criteria (total study participants = 519). CMP hemodynamics were assessed using air plethysmography measurement results from the included studies. A meta-analysis of 8 articles was distilled to 3 with relevant data (83 participants) that found a significant increase on CMP ejection fraction in favor of the exercise group (Hedge’s g = 0.83; 95% CI 0.35-1.30, P <.001) compared to control. CMP residual venous fraction also favored the exercise group (Hedge’s g = 0.42; 95% CI -0.03-0.862, P = .066). Ankle ROM was higher in the exercise group (116 participants; Hedge’s g = 0.62; 95% CI -0.15-1.39, P = .116); however, these differences were not significantly different from controls. Exercise directed at improving calf muscle strength and ankle ROM for individuals with or at risk for VLUs improves CMP hemodynamics and function. Additional research using larger sample sizes to confirm the role of exercise in healing VLUs is warranted.


Chronic venous insufficiency (CVI) is a common, progressive condition that results in lower extremity edema, skin changes, and often ulceration. CVI is partially caused by faulty valves in the lower extremity, resulting in venous reflux and venous hypertension. Additionally, calf muscle dysfunction affects venous function; poor calf muscle hemodynamics reduce venous outflow, causing blood to pool in the lower extremity.1-4 

Venous reflux, caused by a combination of valve dysfunction and failure of the calf muscle pump (CMP), is the main cause of CVI.1-4 Multiple best practice guidelines5-7 and 1 systematic review8 recommend the use of compression dressings to minimize venous hypertension in patients with CVI who have or are at risk for venous leg ulcers (VLUs) given the high correlation of VLU healing with the use of compression stockings. 

Additionally, CMP impairment has a strong effect on VLU healing. The CMP is the primary mechanism by which blood returns to the heart from the lower extremities. Contraction of the calf muscles propels more than 60% of blood from the deep venous system toward the heart.9 In a 1993 case controlled study, Gross et al10 showed that in a group of 43 patients with leg ulcers, 60% of the patients who were considered to have valve deficiency also had an impaired CMP, 24% had no obvious signs of valve insufficiency, and 95% had a neuromuscular disorder that would impair CMP. A 2014 systematic review and meta-analysis by Williams et al11 showed patients with symptomatic venous disease were 1.37 times more likely to have CMP dysfunction and more than 70% of patients presenting with a VLU have an impaired CMP, which also serves as a prognosis for delayed healing of VLUs.12,13 This led to the realization that CMP function plays an imperative role in the development of VLUs.

Ankle movement or flexibility, also called range of motion (ROM), plays an important role in CMP function.9 Case controlled studies2,14 have shown reduced ankle ROM has been linked to severe venous insufficiency as well as active ulceration. Without adequate ankle ROM, the CMP cannot optimally squeeze the blood out of the deep venous system back toward the heart.2 Given these findings, it is important to discern what clinical interventions can be effective in maximizing CMP function in patients with chronic venous disease who have or are at risk for chronic VLUs.

The aim of this systematic review was to examine the effects of exercise on CMP function in patients with CVI with or without VLUs. Secondary outcomes evaluated included the effect of exercise on ankle strength, ankle ROM, and wound healing. 


Data sources. The following electronic databases were searched from date of inception until January 2016: the Cumulative Index to Nursing and Applied Health Literature, (CINAHL) (1981), MEDLINE (1964), Scopus (1966), and EMBASE (1947) (see Table 1). Search strategies were modified as needed for the database utilized. An experienced librarian was consulted to develop an appropriate search strategy for each database. Reference lists of retrieved and review articles were hand-searched for potential secondary sources. Unpublished literature databases also were searched and included The New York Academy of Medicine: grey literature report; MEDLINE plus: National Institute of Health; National Information Center on Health Services Research and Health Care Technology;; The International Standard Randomised Controlled Trial Number registry; the World Health Organization: International Clinical Trials Registry platform search portal, European Union; and Google Scholar. 

Selection of studies. This systematic review only included studies where exercise intervention was used as the treatment variable in adult humans with CVI or those at risk for developing VLUs. Before initiating the literature search, inclusion and exclusion criteria were established (see Table 2) according to the 4-factor (Problem/Population, Intervention, Comparison, and Outcome [PICO]) framework. All randomized and nonrandomized controlled trials, case studies, or case series assessing exercise intervention and its effect on the CMP in patients with CVI were included. Controlled trials comparing exercise intervention (alone or in conjunction with standard wound care and compression) to a control group receiving standard wound treatment were examined for inclusion in the meta-analysis. Narrative articles were not included in the meta-analysis. CMP outcomes could include calf muscle hemodynamics, ankle strength, ankle ROM, and wound healing. Studies were not excluded based on methodological quality. 

Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), an evidence-based minimum set of items for reporting in systematic reviews and meta-analyses, was completed.15 The PRISMA flow diagram is shown in Figure 1; 4 authors were involved in the process. In order to reduce selection bias for titles, abstracts, and full articles, a paired consensus process was utilized. Relevant citations were independently selected, then a paired consensus was conducted to decide which full articles were to be retrieved. This process was applied using the preset list of inclusion and exclusion criteria. Discrepancies were settled through discussion and, when in doubt, authors erred on the side of caution and full articles were obtained for assessment. From citations that met inclusion criteria, full articles were retrieved and evaluated. A similar paired consensus process was repeated after reading all of the full articles that were retrieved in order to select a final list of included studies. 

Outcome measures. The primary outcome measure was the effect of exercise on calf hemodynamics in persons with CVI with or without VLUs. Secondary outcomes included ankle strength, ankle ROM, and wound healing (evaluated by mean change in ulcer area, months to wound recurrence, or healed versus not healed after a duration of time). Calf hemodynamics were assessed using air plethysmography (APG) measurements — specifically, ejection fraction (EF) and residual venous fraction (RVF). APG is a simple, noninvasive, and reproducible quantitative method of evaluating overall lower extremity CVI.16,17 The measurements that quantify the effectiveness of the CMP include the EF and the RVF.18 The EF estimates the percentage of calf blood volume emptied with a single calf muscle contraction during a tiptoe exercise. The RVF measures the blood volume remaining in the calf after 10 consecutive tiptoe exercises. Ankle ROM reported in degrees, ankle strength, and endurance (using a variety of objective tools), with a statement of healed versus not healed were used to evaluate secondary outcomes. 

Data extraction. All data were extracted by 1 reviewer onto a pre-determined form and verified by a second reviewer. Data extracted included number of patients, primary diagnosis, patient demographics, study location, wound characteristics, inclusion/exclusion criteria, exercise group details, control group details, co-interventions, primary outcome measures, secondary outcome measures, wound healing measures, blinding, intention to treat, and study design. Where data within a study were not fully available, the corresponding author was contacted and complete data were requested. 

 Methodological quality. Assessment of methodological quality of all included studies was completed using the Physiotherapy Evidence Database (PEDro) scale, a validated 11-item scale with a score ranging from 0 to 10 (the first criterion is not being counted numerically).19,20 Detailed PEDro results can be found in Table 3. No studies were excluded based on the PEDro result. 

Data analyses. Where outcome variables among studies were similar, a meta-analysis was performed. In the event that study outcomes could not be pooled, results were described narratively. Due to the variation in exercise programs administered as the treatment intervention, a random effects model was used for data analysis. Dichotomous outcomes were expressed as risk ratio (RR) and 95% confidence interval (CI). RR was interpreted as: RR <1, risk of healing is lower in the exercise group; RR = 1, risk of healing is the same between the control and exercise groups; and RR >1, risk of healing is greater in the exercise group. Hedge’s g was used to interpret the effect size of continuous variables due to the small sample sizes of the studies evaluated. A statistically significant difference was denoted as P <.05. All statistical analyses were conducted by 1 reviewer using the Comprehensive Meta-Analysis (CMA) software (Biostat, Englewood, NJ). Standard errors of means were converted to standard deviations when data were not available from the authors. Assessment of heterogeneity was completed by evaluating study characteristics, interventions, outcome measurements, and statistical analysis. Statistical homogeneity was assessed using the I2 statistical test. I2 >50% indicated significant heterogeneity, and the results were interpreted with caution. 


Systematic review. As outlined in Figure 1, electronic searches yielded 1325 results from MEDLINE (395), CINAHL (110), EMBASE (661), and Scopus (159), of which 330 were duplicates and 995 potentially relevant citations. Of these, 972 citations were excluded through a paired consensus process, leaving a total of 23 citations for which full articles were retrieved. An additional 9 full articles were identified by reviewing secondary sources. Of the 32 full articles retrieved, 14 met the inclusion criteria (Table 4A, 4B, 4C, 4D, 4E)4,12,21-32: 10 from primary sources and 4 from secondary sources. Eighteen (18) studies were excluded for the following reasons: no intervention (7), nonexercise-based intervention (4), no calf muscle or ankle ROM exercises (1), review article (1), industry report (1), incorrect patient population (3), or not available through the Western University Library (authors’ facility) (1). 

Population. The 519 participants in all of the included studies (14) were community-dwelling individuals who walked independently. Minimal adverse events were reported in the included studies. The majority of studies (9) included participants with both CVI and VLUs. Three (3) studies included patients with no open wounds,4,21,23 and 1 study included persons with post thrombotic syndrome.27 The age range for study participants was 27 to 91 years. All but 2 studies22,32 included both men and women in the data collection. Eight (8) studies had participants lost to follow-up for various reasons.21-25,27,28,30 Wound duration for the studies that included open ulcers varied; 1 study included a wound that was present for 51 years.24 The study participants were recruited while receiving standard wound care either through a regional home care program or outpatient wound clinic for active or recently healed VLUs. One (1) of the studies required participants to have full ankle ROM to be included.12 

Study types. Eight (8) studies were randomized controlled trials (RCTs),22,25-31 1 was a prospective nonrandomized study,12 1 used a “randomized matched pairs” design,21 and 4 were prospective single-arm studies.4,23,24,32 Control groups consisted of no treatment,21 education only,29 usual care and compression,12,25,28,31 and compression only.22 Three (3) of the studies included exercise as concurrent treatment in the control group with the addition of more intensive, supervised exercises as part of the intervention exercise program.26-28

Leg compression was used at all times in 9 studies12,22,23,25,26,28-31 and during exercise only in 1 study.21 Compression was not used during exercise in 2 studies,4,24 recommended as optional or at the discretion of the health care provider in 1 study,27 and not at all in 1 study.32 The types of compression identified included compression stockings (108 study participants),21,22,23,27 compression bandaging (104 study participants12,26,29,30) either (184 study participants),28 or not described (82 study participants).25,31 

Intervention. Exercise programs in the treatment groups included calf muscle strengthening programs,4,12,25 calf muscle strengthening and ankle stretching ROM programs22-25,27,28,30,32, walking programs,28,29 and 1 unspecified.21 Of the 14 studies assessed, 13 discussed some form of supervision and education for the exercise intervention group; 5 noted the supervision and education for the exercise intervention were given by a health care provider with expertise in exercise prescription, such as a physiotherapist or exercise physiologist.22-24,28,30 The duration of the exercise program also varied among the studies: exercise intervention ranged from 2 times per week to daily, 5-10 minutes to 1 hour per session, and varied between 7 days and 18 months.

Outcomes. Seven (7) studies (159 participants) measured calf hemodynamics using APG.4,12,21-23,25,30 CMP function was measured using venous hemodynamic values through APG.13,16-18,33 Participants in all 7 had an improvement in CMP function after exercise. Six (6) of the 7 studies reported a greater increase in EF and decrease in RVF for the treatment group in comparison to the control group. The seventh study21 did not measure EF or RVF but also noted an improvement in CMP hemodynamics, with a 16% decrease in mean venous capacity in the treatment group, which was greater than the control group that did not show any change. 

Changes in ankle ROM were measured in 6 studies (137 participants).22,24,26,27,30,32 All but 1 of the studies found an increase in ankle ROM in favor of the exercise intervention group.22 However, ankle ROM measurements were performed using different assessment tools including an inclinometer,27 a goniometer,26,27,30,32 and a biodex machine.22 Positioning during ankle ROM measurements was inconsistent: 3 studies assessed participants with their knees bent,22,27,30 2 with the participant’s knee straight,26,32 and 1 was unspecified.24 One (1) study evaluated the effects of exercise on ROM of left and right ankles separately and then utilized the more conservative measurement for data analysis.32 

Calf muscle strength was assessed in 4 studies (114 participants).4,12,22,27 All of the included studies demonstrated an increase in calf muscle strength and/or endurance in the treatment group compared to the control group. Strength measurements were completed using different measurement tools, including a biodex machine,4,22 percentage of max plantar flexion with an ergometer,12 and the heel lift test.27

Five (5) studies (317 participants) assessed the effects of calf muscle exercise on wound healing.25,28,29,30,31 Four (4) demonstrated improved healing rates in the treatment group compared to the control group,28-31 and for the study where healing rates showed no improvement the author stated the results may be due to the small study sample and that the study was not powered to show an effect on ulcer healing, allowing high risk for type II error.25 The 2013 RCT (N = 40) by Ahmed et al31 showed a significant decrease in the Pressure Ulcer Scale for Healing (PUSH tool) score and reduction of the surface area in the treatment group versus control after 12 days. In the 2012 RCT by Heinen et al (N = 184),28 patients with leg wounds occurring after initiating an exercise trial had significantly fewer wound months than patients who had wounds present at baseline in both the control and treatment groups. These findings suggest a potentially positive impact of exercise intervention on healing times. 

Methodological quality of the controlled studies. Four (4) studies did not have PEDro appraisal because they were prospective, single-arm studies.4,23,24,32 The remaining 10 studies scored between 4 and 7 out of 10 on the PEDro scale (see Table 3). The mean PEDro score was 5.8, indicating low overall methodological quality.19 Authors of all 10 studies conducted between-group statistical comparisons and reported both average values and measures of variability for at least 1 key outcome. None of the studies was able to meet the criteria of blinding of all therapists or of all assessors due to the nature of the intervention. All of the studies except Kan and Delis12 had participants that were randomly assigned to groups.

Meta-analysis results. The present meta-analysis evaluated the evidence in the literature related to the effects of exercise intervention on the CMP in participants with CVI with or without VLUs. Although 14 articles met the inclusion criteria for review, only 8 were included for meta-analysis due to the variability of the study methodology and data available to calculate effect sizes.22,25-31 Of the 7 studies that reported calf hemodynamics,4,12,21-23,25,30 3 were included in a meta-analysis to calculate the effect of exercise on CMP function (83 participants), specifically for the values of EF and RVF.22, 25, 30 The results from the meta-analysis for EF showed a significant increase in favor of the treatment group with a large effect size (Hedge’s g = 0.83; 95% CI 0.35-1.30, P <.001) (see Figure 2). RVF also favored the intervention group although it was not statistically significant (Hedge’s g = 0.42; 95% CI -0.03-0.862, P = .066) (see Figure 3). 

Eight (8) studies4,12,22,24,26,27,30,32 evaluated the effects of exercise intervention on ROM and/or ankle strength, both of which affect the efficiency of the CMP. Due to the heterogeneity of the outcomes used to assess CMP strength, none of the studies was included in meta-analysis. Four (4) RCTs (116 participants) assessing ankle ROM were similar and facilitated combining results.22,26,27,30 The results from the meta-analysis for ankle ROM showed a moderate effect size toward increased ROM in the treatment group compared to control group, but it was not statistically significant (Hedge’s g = 0.62; 95% CI -0.15-1.39, P = .116) (see Figure 4). 

Included in this review are 5 studies assessing the effect of exercise on wound healing.25,28,29,30,31 Three (3) measured whether wounds were healed or not healed after 12 weeks and were combined in a meta-analysis to calculate the effect of exercise on VLU healing (N = 93).25,29,30 Results showed no significant difference in healing rates between exercise groups and control groups after 12 weeks of CMP exercises (RR: 1.027; 95% CI 0.77 to 1.38, P = .860) (see Figure 5). 


This systematic review and meta-analysis demonstrate the benefit of simple exercises for people with CVI and/or VLU. All of the 14 included studies (N = 519) produced results that showed incorporating exercise into standard practice with compression improves calf muscle function. Specifically, ankle exercises performed in a variety of ways resulted in better calf muscle hemodynamics compared to control treatments and increased ankle ROM compared to pre-treatment values. 

A 2012 narrative review by O’Brien et al34 identified the relationship between CMP exercises and improved venous return and ankle ROM. The authors also noted that the current literature contains only a few RCTs with small sample sizes. A 2016 Cochrane review also concluded the evidence available to assess the efficacy of physical exercise in people with CVI is insufficient.35 Unfortunately, the authors’ search of relevant literature revealed only 2 RCTs (Padberg22 and Hartman21), and outcomes were not able to be combined using meta-analytical methods. These authors did not provide rationale as to why studies that were identified in the current research were not included in their review. 

Using a systematic approach, the current authors were able to identify additional published work and complete a meta-analysis. By pooling the results, the overall effect of exercises on calf muscle function was determined. In this effort, all 7 studies that evaluated the effect of exercise on CMP functions reported positive findings — specifically, exercise was shown to significantly increase CMP EF and decrease RVF as compared to control treatments. 

Of note, only 5 of 14 studies utilized health care professionals who were trained in exercise prescription (eg, physiotherapists) to provide or supervise the exercise intervention. All 5 of these studies reported positive changes in CMP function and ankle strength or ROM.21,22,23,27,32 The link between CMP function and gait impairments has not been established; however, it is clear that expertise is required to carry over gains in ankle ROM and strength to produce improved walking efficiency and proper foot biomechanics. 

The strong association between CMP function and VLU severity and healing is well known. However, few studies have directly examined the effect of exercise on healing rates of people with VLUs. Four (4) of 5 studies in this systematic review found adding ankle exercises improved VLU healing. The method of measuring VLU healing was quite different across these studies. O’Brien et al30 and Meagher et al29 reported the proportion of people with healed ulcers after 12 weeks of exercise, whereas Ahmed et al31 used the PUSH tool to evaluate changes in wound appearance. Heinen et al28 conducted by far the largest RCT in this review; however, they measured VLU recurrence rate. Only 3 studies measured VLU size before and after an exercise intervention; one, a study by Jull et al,25 did not show exercise was associated with better healing outcomes. Therefore, when results were combined in the current research, an overall effect that favored exercise was not noted.

The current results are based on studies that had low to moderate overall scores for methodological quality. Most studies did not blind patients, therapists, or assessors, which is to be expected given the nature of this intervention. However, compliance with exercise prescriptions and drop-out rates across the studies were found to be acceptable (446 of 519 completed).

Of patients with a VLU, 70% also have CMP dysfunction; therefore, interventions that optimize ankle strength and flexibility should be considered mainstay treatment for people with a VLU.12 Although compression has been associated with better healing of VLU, it may interfere with the ankle’s ability to move through full ROM during ambulation or activity causing an altered gait pattern.34,35 CMP activation with ambulation is most effective with a heel-toe gait pattern with activity-appropriate footwear, which provides credibility to structured education sessions on exercise programs with expert clinicians. 

Minimal adverse events were reported in the included studies, a finding that indicates exercise is a low-risk, feasible intervention for people who have CVI and/or VLU. Furthermore, physical activity is known to reduce hypertension and maintain fitness in this population.36,37 Considering CMP exercises can produce favorable results with minimal risk, clinicians treating this relatively sedentary group of patients should prescribe these exercises or make a referral to a qualified professional who can improve ankle function and address gait impairments. 


Few studies exist in the published literature regarding the effects of exercises on individuals with CVI and persons at risk for VLU. Even fewer exist that recorded results using similar outcomes. In particular, studies assessing the outcomes of ROM, strength, and wound healing for CVI patients are lacking. A wide array of outcome measures is used in the included studies, which prevented the pooling of data from all of the studies into the meta-analysis. Considerable variation also exists in the exercise intervention, which includes exercise duration, frequency, and intensity. This makes it difficult if not impossible to assess the type of exercise that was most effective at improving CMP function. 

Using a very conservative approach to the meta-analysis, the authors were able to combine results of 3 smaller studies and produce a pooled effect in favor of exercise. As with most systematic reviews, the authors were restricted to information provided by the original researchers (which was limited in some cases); data were converted when authors could not be contacted. This study also only included community-dwelling individuals (eg, no long-term care patients).


A review and meta-analysis showed exercise can increase CMP function. Ankle ROM and strength also were found to be improved, although by a statistically insignificant amount (P >.05). Because the conclusions reached were based on low to moderate quality controlled clinical trials, they need to be interpreted with caution, but evidence suggests exercises can increase EF. Additional research in larger studies that explore the effect of exercise on the healing rates of VLU is warranted. Once the body of literature on this subject increases, it will be possible to determine whether certain exercise regimens are optimal, as well as the connection between structured exercises and supervised versus unsupervised programs. This research could help support the role of clinicians specialized in exercise training within the interprofessional wound care team. 


Ms. Orr is a physiotherapist and PhD student, School of Physical Therapy, Western University, Elborn College, London, Ontario, Canada. Ms. Klement is a Canadian Certified pedorthist, SoleScience Inc, Fowler Kennedy Sports Medicine Clinic, 3M Centre, Western University. Ms. McCrossin is a physiotherapist and instructor, School of Physiotherapy, Dalhousie University, Halifax NS, Canada. Ms. Drombolis is a physiotherapist, Riverside Health Care, Fort Frances, Ontario, Canada. Dr. Houghton is a Professor, School of Physical Therapy; Dr. Spaulding is a Professor, School of Occupational Therapy; and Dr. Burke is an associate professor, School of Health Studies, Western University, Elborn College.


Please address correspondence to: Lyndsay Orr, BScPT, MCISC-WH, PhD(c), School of Physical Therapy, Western University, Elborn College, 1201 Western Road, London, Ontario, Canada N6G 1H1; email: