What is the effect of exercise on wound healing in patients with venous leg ulcers? A systematic review

Abstract

Standard best practice for the treatment of venous leg ulcers (VLUs) is compression bandaging of the lower leg to reduce hydrostatic pressure. There is considerable variation in reported healing rates when using this gold‐standard approach; therefore, a systematic and robust evaluation of other interventions is required. Exercise interventions, in addition to standard compression therapy, could improve wound‐healing time and prevent their recurrence. We have conducted a systematic review to examine the effects of exercise on wound characteristics, including time to heal, size and recurrence, pain, quality of life, adverse events, and economic outcomes. This review was registered with PROSPERO 2016:CRD42016046407. A systematic search of Ovid Medline, Ovid EMBASE, Ovid CINAHL, The Cochrane Library, PsycINFO, Web of Science, and PEDro was conducted on January 30, 2017, for randomised control trials to examine the effects of exercise on time to heal, size and recurrence, pain, quality of life, adverse events, and economic outcomes. Six studies met the inclusion criteria, but all had design flaws leading to biases, most commonly performance and selective reporting bias. Three studies compared a progressive resistance exercise programme (PREG) plus compression with compression alone for a period of 12 weeks. Low‐quality evidence indicates the following: possibly no difference in the proportion of ulcers healed (risk ratio [RR] 1.14, 95% CI 0.71 to 1.84, I ² 36%; 3 trials, 116 participants); probably no difference in quality of life (mean difference [MD] 3 points better on 100 point scale with exercise, 95% CI −1.89 to 7.89, 1 trial, 59 participants); possible increase in the risk of adverse events with exercise (OR 1.32, 95% CI 0.95 to 1.85, 1 RCT, 40 participants); and no difference in ankle range of motion and calf muscle pump. Evidence was downgraded due to susceptibility to bias and imprecision. Recurrence, pain, and economic outcomes were not measured in these trials, and time to healing was measured but not fully reported in 1 trial. We are uncertain of the effects of other interventions (community‐based exercise and behaviour modification, ten thousand steps, supervised vs unsupervised exercise) due to the availability of low‐ or very low‐quality evidence only from single trials. The review highlights the need for further research, with larger sample sizes, to properly address the significance of the effect of exercise on VLU wound characteristics.

1. INTRODUCTION

1.1. Rationale

Chronic venous insufficiency is 1 of the least researched cardiovascular diseases, which left untreated causes venous leg ulcers (VLU).1 Venous insufficiency is the most frequent aetiology (75%‐80%) of VLUs,2 arising from venous valve incompetence and calf muscle pump (CMP) insufficiency. This insufficiency leads to venous stasis and hypertension, which are associated with reduced ankle range of motion.3 The CMP is an important regulator of venous blood flow and blood pressure in the lower limbs.3 During exercise, for example, walking, the calf muscles contract and compress the intramuscular and deep veins, raising venous pressure and propelling blood in the deep venous system to flow towards the heart, while the 1‐way valve function prevents reflux, thereby preventing blood from pooling.4 Over 70% of patients with VLUs have an impaired CMP, and related to this prolonged wound‐healing time.

Different studies estimate the prevalence of VLUs to be 1% to 3% of the population.2 Prevalence increases with age to 1% to 2% between the ages of 65 and 95 years5 and approximately to 2% in people aged 80 years and older.6 Difficult‐to‐heal VLUs are costly ($4500/patient), and after healing, up to 70% of VLUs may recur.7 The economic and social burden due to VLUs is expected to rise with the population aging, the global spread of smoking, and the growing epidemics of obesity and type 2 diabetes.8

Standard best practice treatment includes compression of the lower leg by tight bandaging to reduce hydrostatic pressure in the leg.1, 9, 10 There is considerable variation in reported healing rates when using the gold‐standard approach, compression bandaging. Multi‐layered compression bandaging is reported to heal 23% to 69%of patients after 12 weeks of treatment,11, 12, 13 increasing to 87% at 52 weeks.12 However, up to 15% to 30% of chronic VLUs do not respond to compression treatment14 and remain unhealed, even after a year of treatment.15

Patient compliance with the VLU treatment regimens is a well‐documented problem.16 Compression bandages are recommended as long as there is evidence of venous disease, which often means that the treatment is lifelong.16 Compliance with treatment regimens is considered to be an important determinant in wound healing and rate of recurrence.16 Currently, there is little evidence to suggest that health care system options, such as leg clubs and leg ulcer clinics, increase patients' compliance with long‐term compression therapy.16

Compliance issues with compression therapy, slow response to this treatment regime, and a high recurrence rate of VLUs suggest that alternative adjunct treatments are necessary, for example, lifestyle advice such as physical activity.17 Physical activity is any bodily movement produced by skeletal muscles that requires energy expenditure, such as working, playing, carrying out household chores, travelling, and engaging in recreational pursuits.18 Exercise forms a subset of physical activity and involves a structured, planned, repetitive approach, with the aim to improve or maintain 1 or more components of physical fitness.18 There is an urgent need to systematically and accurately evaluate exercise interventions to determine whether they improve ulcer healing time and prevent their recurrence. An evaluation is especially pertinent as exercise is a low‐cost intervention with the potential to reduce comorbidities and the burden of disease.19 Exercise, such as regular walking, if incorporated into the treatment plan for VLU patients, may assist in facilitating the CMP, which in turn can improve healing rates.

1.2. Objective

The aim of this study was to conduct a systematic review to examine the effects of exercise in addition to standard compression therapy on VLU characteristics, including time to heal, size and recurrence, pain, quality of life, adverse events, and economic outcomes.

2. METHODS

2.1. Protocol and registration

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta‐Analyses (PRISMA).20 The protocol for this review was registered with PROSPERO 2016:CRD42016046407.21

2.2. Eligibility criteria

2.2.1. Inclusion criteria

Studies that met the following criteria were included in the review:

• the study design was a randomised controlled trial (RCT) comparing exercise with other interventions, in addition to standard compression therapy, in all participants.

• people with a VLU.

• exercise included:

  • ‐
    strength or resistance training and aerobic activity that involved the lower limb.
  • ‐
    land‐ and water‐based.
  • ‐
    structured and not structured.
  • ‐
    supervised and unsupervised (ie, undertaken at home).
• studies that reported 1 or more of the following outcomes were eligible for inclusion:

  • ‐
    primary outcomes: time to complete healing, proportion of VLUs healed, recurrence of VLUs.
  • ‐
    secondary outcomes: health‐related quality of life, wound pain, proportion with adverse events, economic outcomes, calf muscle pump (CMP) function, and range of ankle mobility (ROAM).

2.2.2. Exclusion criteria

Studies were excluded where the exercise comprised flexibility, household chores, and activities of daily living.

2.3. Information sources and Search

An electronic search was conducted on Ovid Medline (1946—January 30, 2017), Ovid EMBASE (1974—January 30, 2017), Ovid CINAHL (1982—January 30, 2017), The Cochrane Library (January 30, 2017), PsycINFO (1806—January 30, 2017), Web of Science (1900—January 30, 2017), and PEDro (1929—January 30, 2017). Medical subject heading search terms and text or keywords associated with the concepts of “physical activity” and “venous leg ulcers” were used (Table 1).

Table 1.

Search strategy

Database: Ovid MEDLINE(R) 1946 to present, with daily update Search strategy: Varicose Ulcer/ Venous Insufficiency/ Leg Ulcer/ ((varicose adj3 ulcer*) or (venous adj3 ulcer*) or (leg adj2 ulcer) or (stasis adj2 ulcer*) or (lower adj3 extremit* adj3 ulcer*) or (crural adj2 ulcer*)).mp. [mp = title, abstract, original title, name of substance word, subject heading word, keyword heading word, protocol supplementary concept word, rare disease supplementary concept word, unique identifier] (ulcus adj2 cruris).mp. 1 or 2 or 3 or 4 or 5 Exercise Movement Techniques/ or Exercise Therapy/ or Exercise/ Resistance Training/ (resistance adj2 training).mp Physical Therapy Modalities/ Sports/ sport.mp. sports.mp kinesiotherap*.mp physiotherap*.mp exercise*.mp. (exercise adj2 therap*).mp (Physical adj2 (therap* or fitness or exercise* or activit* or train*)).mp 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14 or 15 or 16 or 17 or 18 6 and 19 limit 20 to yr = “1990 ‐Current” limit 21 to (addresses or autobiography or bibliography or biography or case reports or comment or historical article or interactive tutorial or interview or lectures or legal cases or legislation or letter or news or newspaper article or personal narratives or portraits or "review" or "scientific integrity review" or systematic reviews or technical report or validation studies or video‐audio media or webcasts) 21 not 22

2.4. Study selection

Search results were collated in a reference database (EndnoteX5, Thomson Reuters, 2010), duplicates were deleted, and initial screening of titles was independently conducted by 2 reviewers (DS and RMG). A priori inclusion and exclusion criteria were applied at this stage. Two reviewers independently applied inclusion and exclusion criteria to the full texts of remaining references to select studies for this review (D.S. and R.L.). Manual searches of reference lists and citation tracking of papers identified as potentially relevant were also conducted. Discordance between reviewers was resolved by discussion, and when necessary, by a third reviewer (J.O.B.).

2.5. Data collection process and data items

A full‐text review of each included study was conducted by 1 reviewer (DS), and the following data items were extracted: study characteristics, population characteristics, and research design. The following outcome data were also extracted: number of events and participants per treatment group for dichotomous outcomes (proportion of VLU healed, recurrence of VLU, proportion with adverse events) and mean, standard deviation, and number of participants per group for continuous outcomes (time to complete healing, health‐related quality of life, wound pain, CMP function and ROAM). A second author (RL) reviewed the data extraction, and disagreements were resolved via consensus.

2.6. Risk of bias in individual studies

Risk of bias for the included articles was assessed using the Cochrane Collaboration's tool for assessing risk of bias22 using Review Manager (RevMan) software.23 We assessed the risk of selection, performance, detection, attrition, and selective reporting biases and other biases (baseline imbalance or uneven application of co‐interventions) as low, high, or unclear. Two reviewers independently rated each study against the criteria (D.S. and R.L.). Discordance was resolved by discussion, and when necessary, by a third senior reviewer (C.W. or R.J.).

2.7. Quality of evidence

We used GRADE to assess the overall quality of the evidence (ie, trials) underpinning each outcome (R.J. and C.W.) (GRADE Working Group).24 GRADE considers 5 factors in the quality: study limitations (risk of bias), consistency of effect, imprecision, indirectness, and publication bias. High‐quality evidence has no or inconsequential study limitations across trials, consistent effect across trials, little imprecision, little indirectness, and no or unlikely publication bias. Evidence is downgraded to moderate if there is limitation in 1 of the 5 factors (eg, bias that may affect the result), low (for 2 factors), or very low (more than 2).

2.8. Summary measures

The included studies' results were plotted as point estimates, that is, RR with corresponding 95% CI for dichotomous outcomes; mean difference (MD); and 95% CI for continuous outcomes using RevMan software.23 Authors of trials were contacted to obtain relevant information when this was missing in the published article. If no response was received from the authors of trials, the results were presented as reported in the original trial report.

2.9. Synthesis of results

When possible, outcomes were presented using forest plots. For clinically homogeneous studies, with similar participants, comparators, and using the same outcome measure, we pooled outcomes in a meta‐analysis, using the random‐effects model as the default.23 Heterogeneity across studies was assessed by the I ² statistic.

3. RESULTS

3.1. Study selection

The combined searches yielded 1267 records, of which 90 studies were selected for full‐text review. Of 90 studies, 6 met our inclusion criteria and were assessed in the quantitative synthesis (Figure 1); 84 studies were excluded for the following reasons: 67 studies were not RCTs; 7 did not include participants with venous leg ulcers; 5 did not include exercise interventions; 4 were not journal articles; and 1 was a duplicate (Appendix).

Figure 1.

PRISMA flowchart of search results

3.2. Study characteristics

Studies were published between 2009 and 2017. All studies were RCTs that involved patients with VLUs. Details of each of the included studies are summarised in Table 2. Settings were diverse, including outpatient clinics,25, 26 university hospitals,27 and patient homes28, 29, 30 (Table 2). Primary outcomes differed among the studies, although all studies reported wound characteristics (time to heal, proportion of ulcers healed, and ulcer recurrence). Intervention and control groups are defined in Table 3, and the participants' baseline characteristics are outlined in Table 5. Economic outcomes were not reported in any of the studies (Table 4). Other missing outcomes are outlined in Table 4.

Table 2.

Characteristics of included studies

Author (y)	Country	Study design	Setting	Study period (wk)	Population	No. of participants (n)	Intervention types	Outcomes	Funding
Heinen (2012)25	NL	RCT	OPC	12	Adults up to 65 y+ with VLU with mixed aetiology of VI and arteriolar or minor arterial insufficiency	184 IG: 92 CG: 92	IG: Lively Legs programme CG: CAU	Adherence to therapy, ulcer size, comorbidity, proportion healed¥, and ulcer recurrence¥	Radbound University
Jull (2009)28	NZ	Open‐label RCT	HB	12	Adults 18 y+ with VLU	40 IG: 21 CG: 19	IG: PREG in addition to compression. CG: compression only	CMP function¥, ulcer size, time to complete healing¥, and proportion healed¥	Health Research Council of NZ
Meagher (2012)27	IRL	RCT	University hospital	12 (or until fully healed)	Patients with VLU	35 IG: 18 CG: 17	IG: Exercise programme in addition to compression CG: Compression only	Time to complete healing¥, proportion healed¥, pain¥, ulcer size, and duration	No external source of funding
O'Brien (2013)29	AUS	Open‐label RCT	HB	12	Adults 18 y+ with VLU	13 IG: 6 CG: 7	IG: PREG in addition to compression. CG: CAU	Ulcer area and size, time to complete healing¥, proportion healed¥, CMP function¥, and ROAM¥	Queensland University of Technology
O'Brien (2017)30	AUS	Open‐label RCT	HB	12	Adults 18 y+ with VLU	63 IG: 31 CG: 32	IG: PREG in addition to compression. CG: CAU	Ulcer area and size, ROAM¥, adherence to therapy, time to complete healing¥, proportion healed¥, and health‐related quality of life¥	Queensland University of Technology
Szewczyk (2010)26	PL	RCT	OPC	9	Patients with VLU	(32) IG: (16) CG: (16)	IG: Extensive supervised exercise programme CG: Unsupervised exercise programme	ROAM¥, pain¥, ulcer swelling, venous claudication, lipodermatosclerosis, number of ulcers, size and duration, ulcer recurrence¥	—

Abbreviations: —, not specified/not reported; ¥, indicates outcome was reported in this review; AUS, Australia; CAU, care as usual; CG, control group; CMP, calf muscle pump; HB, home based; IG, intervention group; IRL, Ireland; PREG, progressive resistance exercise programme; NL, Netherlands; NZ, New Zealand; ROAM, range of ankle motion; OPC, outpatient clinic; PL, Poland; RCT, randomised control trial; VI, venous insufficiency; VLU, venous leg ulcer.

Table 3.

Description of intervention and comparison groups

Author (y)	Number of intervention groups	Intervention group	Comparison group
Heinen (2012)25	1	Nurse‐led self‐management counselling programme (Lively Legs), consisting of 2‐6 sessions on physical activity and adherence to compression therapy. Adherence to physical activity and compression were assessed, barriers and facilitators for behaviour changed discussed using motivation interviewing, and individual tailored goals were set. Lively Legs programme was based on Social Cognitive Theory, the Precaution Adoption Process Model, and motivational interviewing. The first session involved assessing patients' lifestyles, and subsequent sessions involved evaluation of behaviour change and giving feedback.	Treatment according to CBO (2005) guidelines for VLU (wound care and compression). No structured assessment/counselling with regards to lifestyle. Treatment performed at outpatient clinic weekly (on average).
Jull (2009)28	1	Home‐based, individually tailored 12‐wk progressive resistance exercise programme in addition to compression. 3‐5‐min warm up of walking followed by heel raises. Heel raises performed with 1 or 2 arms against a wall for support before participant slowly raises onto tip‐toes. Dosage was 3 sets of reputations at 80% of the participants' maximum (assessed by a registered nurse every 3 wk during the intervention phase). Regimen undertaken on alternate days.	Compression only, visited on same schedule as intervention. All participants received normal district nursing care.
Meagher (2012)27	1	Home‐based, 12‐wk (or until fully healed) exercise programme whereby participants were asked to perform a target of 10 000 steps per day.	Compression only, asked not to alter walking behaviour and received patient information leaflet regarding study aims and objectives. No specific advice regarding exercise was given.
O'Brien (2013)29	1	Home‐based, individually tailored 12‐wk unsupervised progressive resistance exercise programme in addition to compression. The exercise protocol is split into 3 stages Seated heel raises Standing heel raises (both legs) Standing heel raise (1 leg) The participants progress through each stage by the various 4 levels 10 repetitions × 3 sets × 3 times per day 15 repetitions × 3 sets × 3 times per day 20 repetitions × 3 sets × 3 times per day 25 repetitions × 3 sets × 3 times per day Participants moved on to the next stage once they comfortably completed the current stage for at least 3 days/until ready to progress based on their feedback and in consultation with the principal researcher. Participants were also encouraged to walk at least 3 times per week for 30 min if possible.	Compression bandaging and wound care. Controls received a brochure about the importance of exercise for adults with VLUs and a pedometer to observe the amount of daily steps.
O'Brien (2017)30	1	See O'Brien 201329	See O'Brien 201329
Szewczyk (2010)26	2	9‐wk exercise programme. All participants walked 3 km daily and completed 3 sets of 15 repetitions of circular foot movements, heel raises, and alternate performance of foot dorsiflexion and plantar flexion each day. All participants performed exercises after applying multi‐layer compression. Intervention group additionally exercised on training bikes under nurse supervision during outpatient clinic appointment, twice a week for 20 min each time at a moderate pace.	All participants walked 3 km daily and completed 3 sets of 15 repetitions of circular foot movements, heel raises, and alternate performance of foot dorsiflexion and plantar flexion each day. All participants performed exercises after applying multi‐layer compression.

VLU, venous leg ulcer.

Table 5.

Baseline Patient Characteristics

Author, y	Age in years (mean)	Gender (%)	BMI	Ulcer area (mean cm2)	Ulcer duration	ABPI (mean)	Mobility n (%)
Heinen (2012)25	IG: 65 CG: 67	B: 60 F	IG: 31 CG: 29	IG: 9 CG: 8.4	Mean months at baseline: IG: 7 CG: 7.3	—	—
Jull (2009)28	IG: 54.6 CG: 53.3	IG: 23.8 M CG: 72.7 M	—	Median: IG: 3.4 CG: 3.1	Median wk: IG: 23 CG: 28	B: 1.1	W/O aids: IG: (85.7) CG: (84.2)
Meagher (2012)27	IG: 66a CG: 78a	IG: 33 M CG: 24 M	>25 kg/m2 IG: 78% CG: 59%	<10 cm2 IG: 94% CG: 65% >10 cm2 IG: 5.6% CG: 35%	Median wk: IG: 8.5 CG: 15	IG: 1.17 CG:1.00	Independently mobile: IG: (94) CG: (82)
O'Brien (2013)29	IG: 66 CG: 63.6	IG: 50 M CG: 42.8 M	IG: 31.7 CG: 36.7	IG: 5.1 CG: 3.2	Mean wk: IG: 19.5 CG: 34.8	B: 0.07	With aids: B: 2
O'Brien (2017)30	IG: 71.3 CG: 71.7	51.6b	—	IG: 8.8 CG: 6.0	Median wk IG: 16 CG: 14	—	—
Szewczyk (2010)26	IG: 76 CG: 71	IG: 75 M CG: 56.3 M	—	IG: 43.1 ± 52.6 CG: 47.7 ± 73.2	IG: 7·2 ± 9.55 y CG: 8.5 ± 9.91 y	—	Total range of ankle‐leg mobility: IG: 22.7 ± 12.3 CG: 24.6 ± 12·7

Abbreviations: —, not specified/not reported; ABPI, ankle brachial pressure index; B, both; BMI, body mass index; CG, control group; F, female; IG, intervention group; M, male.

^aMedian years.

Frequencies and percentages within groups not reported.

Table 4.

Summary of study results

Author (y)	Time to healing (wk 12)	Proportion healed	Recurrence	Quality of life	Pain	Adverse events	Economic outcomes	Calf muscle pump function	Range of ankle mobility
Heinen (2012)25	—	IG: (51/92) 55% CG: (41/92) 45% by 18 mo.	IG: (32/69) 46% CG: (38/67) 56% at 18 mo HR 0.61, 95% CI: 0.35‐1.06, P = .07.	—	—	—	—	—	—
Jull (2009)28	IG: (8/21) 38% CG: (10/19) 53% Unadjusted HR 0.68, 95% CI: 0.28‐1.68, P = .49.	IG: (8/21) 38% CG: (10/19) 53% by wk 12.	—	—	—	IG: 59% CG: 41% OR 1.32,95% CI: 0.95‐1.85.	—	VV MD 3.7, 95% CI: −29.9‐22.5, P = .8 EV MD 21.8, 95% CI: −3.3 to 46.8, P = 0.09 EF MD 18.5, 95% CI: 0.03‐36.6, P < .05 VFI MD −0.1, 95% CI: −2.2 to 2.1, P = .9 RV MD 6.1, 95% CI: −13.2‐25.4, P = .5 RVF MD −10.4, 95% CI: −29.2‐8.4, P = 0·3.	—
Meagher (2012)27	IG: (15/18) 83% CG: (13/17) 76% P = .13 Number of steps was associated with time to healing at wk 1 and wk 4 (P = .052, P = .008, respectively).	IG: (15/18) 83% CG: (13/17) 76% by wk 12 P = .13.	—	—	Pain not associated with increasing the number of daily steps (P = 0.45). Relationship b/w pain and healing time not statistically significant (P = .88).	—	—	—	—
O'Brien (2013)29	—	IG: (3/6) 50% CG: (3/7) 40% by wk 12 χ2 = 0.11, P = .74.	—	—	—	No adverse effects in either groups.	—	EF F 1,5 = 6.7, P = .05 RVF F 1,5 = 7.02, P = .04 VFI F 1,5 = 0.29, P = .61.	IG: 31.8 ± 6 CG: 23.5 ± 8
O'Brien (2017)30	—	IG: (24/31) 77% CG: (17/32) 53% by wk 12 χ2 (1, n = 59) = 2.75, P = .09.	—	IG: 46 ± 10.2 CG:43 ± 8.9 RR 3.00, 95% CI: −1.89‐7.89, P = .2.	—	—	—	—	IG: 32 ± 10.75 CG: 24 ± 10.25
Szewczyk (2010)26	—	—	—	—	No association b/w pain and range of ankle mobility.	—	—	—	Statistically significant increase in the total range occurred later CG compared with IG.

Abbreviations: —, not specified/not reported; b/w, between; CG, control group; CI, confidence interval; EJ, ejection fraction (%); EV, ejection volume (mL); HR, hazard ratio; IG, intervention group; RR, risk ratio; RV, residual volume (mL); RVF, residual volume fraction (%); VFI, venous filling index (mL/s); VV, venous volume (mL).

3.3. Risk of bias within studies

Three studies described adequate methods of allocation sequence and concealment and were at low risk of selection bias28, 29, 30 (Figure 2). The remainder (n = 3) reported these methods unclearly25, 26, 27 and were at risk of selection bias. All studies were at risk of performance bias, with none of the studies reporting adequate participant blinding. Studies were at low risk of detection bias for physician‐ or investigator‐reported outcomes (time to complete healing, proportion of VLUs healed, recurrence of VLU, range of motion and CMP function) as we judged that knowledge of the intervention would likely have little effect on measurement of these objective outcomes. Studies that included self‐reported outcomes (pain, quality of life, and adverse events) were at risk of detection bias in the measurement of these outcomes.27, 28, 30Attrition bias was unlikely in the majority of the studies due to few participant losses to follow up (n = 5),25, 26, 28, 29, 30 and 1 study did not clearly report their losses to follow up.27 We judged 4 studies at low risk of selective reporting bias 26, 28, 29, 30 as they reported primary outcomes fully. However, 2 studies selectively reported outcomes; they measured risk in 1 study (n = 1)27 and high risk in 1 study (n = 1).25

Figure 2.

Risk of bias table for included studies

3.4. Results of studies

3.4.1. Effects of interventions

Three studies compared a PREG with compression alone.28, 29, 30 Single studies reported the following comparisons: community‐based exercise and behaviour modification (Lively Legs) vs compression25; Ten Thousand Steps exercise programme vs compression27; and supervised exercise programme plus compression compared with unsupervised exercise plus compression.26

3.4.2. PREG compared with compression care

Three studies compared a PREG intervention with compression (CC) for a period of 12 weeks (n = 116).28, 29, 30 Wound recurrence, pain, and economic outcomes were not measured in these trials, and the difference between the groups related to time to complete healing only was reported in a single trial without means and measures of variance per treatment group. Quality of evidence for all other outcomes was downgraded to low quality or very low quality due to risk of biases in studies and imprecision due to low event rates and low number of participants.

3.4.3. Time to complete healing

Time to complete healing was only reported inadequately in 1 study (n = 40)28 (Table 4). Means and standard deviations could not be extracted from the data, nor were they provided upon request; thus, we could not verify the result. Quality of evidence for this outcome was therefore judged as very low due to failure of studies to report the outcome or to report the outcome selectively, presence of performance, and detection biases and imprecision.

3.4.4. Proportion healed

Proportion of VLUs healed was reported at up to 12 weeks for all 3 studies (n = 116) (see Table 5). Proportions of people with healed VLUs were similar in the compression only group (30/58) and the compression plus exercise (35/58), and there was no statistical difference between groups (RR: 1.14, 95% CI 0.71‐1.84, I ² 36%). Evidence for this outcome was of low quality due to design flaws that rendered the studies susceptible to performance bias and imprecision due to the low number of events and participants.

3.4.5. Quality of life

Low‐quality evidence for quality of life was available from 1 study30 (n = 59) using the Medical Outcome Survey Short Form‐8 questionnaire (SF‐8). Physical‐ and mental health‐related quality‐of‐life scores (PCS) did not differ between PREG and CC (MD 3.00, 95% CI −1.89‐7·89, P = .23) (Table 4).

3.4.6. Adverse events

Low‐quality evidence from 1 study (n = 40) indicates a possible higher proportion of adverse events in the PREG compared with controls28 (OR 1.32, 95% CI 0.95‐1.85) (Table 4). Evidence was downgraded due to the risk of bias and imprecision.

3.4.7. CMP function

Two papers reported CMP function (n = 47).28, 29 Air plethysmography (APG) was used to measure venous volume (VV), ejection fraction (EF), venous filling index (VFI), and residual volume fraction (RVF) in both studies. There was no difference between groups for VV scores (MD: 4.13, 95% CI −20.97‐29·24, I ² 0%, P = .75), EF scores (MD: 13.74, 95% CI −3.34 to 30.82, I ² 27%, P = .11), VFI scores (MD: −0.21, 95% CI −1.36‐0.93, I ² 0%, P = .72), or RVF scores (MD: −6.37, 95% CI −19.33‐6.60, I ² 0%, P = .34). Evidence was downgraded to low quality due to the risk of bias and imprecision.

3.4.8. Range of ankle mobility

Two papers reported ROAM (n = 76),29, 30 finding a difference between groups for ROAM in favour of the intervention group (MD: 8.08, 95% CI 3.63‐12.54, I ² 0%,). Evidence was downgraded to low quality due to risk of bias and imprecision.

3.5. Lively Legs exercise programme plus care as usual compared with CC only

One trial of 184 participants compared community‐based exercise and behaviour modification (Lively Legs) plus usual care (wound care, compression bandages at an outpatient clinic) with usual care alone for a period of 18 months.25 Only 2 outcomes of interest were reported (proportion of ulcers healed and recurrence of ulcers), and only low‐quality evidence underpinned these outcomes due to imprecision and risk of bias. Time to healing, quality of life, pain, adverse events, economic outcomes, CMP, and ankle range of motion were not reported.

3.5.1. Proportion healed

There is uncertainty regarding the differences of proportions of people healed between groups (RR 1.24, 95% CI 0.93‐1.67) due to possible imprecision of the results and risk of selection bias. See Table 4 for the proportions of people healed with regards to the Lively Legs exercise programme plus care as usual and CC only.

3.5.2. Recurrence

There is uncertainty regarding the differences of the number of participants who experienced a recurrent ulcer between treatment groups (RR 0.82, 95% CI 0.59‐1.14) due to possible imprecision around the results and risk of selection bias. See Table 4 for recurrence rates between the Lively Legs exercise programme plus care as usual and CC only.

3.6. Ten Thousand Steps exercise programme plus care as usual compared with CC only

One trial with 35 participants compared a Ten Thousand Steps exercise programme plus care as usual (compression bandages) with CC alone27 and reported only 1 outcome of interest (proportion healed) at various time points. Evidence was low quality for these outcomes due to risk of biases and imprecision due to the small number of participants. Ulcer recurrence, quality of life, adverse events, economic outcomes, CMP, and ankle range of motion were not measured, and time to healing was measured, but mean data were not reported.

3.6.1. Time to complete healing

Means and standard deviations could not be extracted from the data, and at the time of publication, the authors did not respond to our request to provide the data (Table 4).

3.6.2. Proportion healed

Proportion of ulcers healed is reported at up to 12 weeks and is included in Table 4. There was no difference between groups for proportion of ulcers healed (RR: 1.09, 95% CI 0.78‐1.52).

3.6.3. Pain

The visual analogue scale ranging between 0 and 10, whereby a score of 0 indicates no pain and a score of 10 indicates severe pain, was used to assess the participants' pain levels. Means and measures of variance were not reported nor provided upon request; thus, we could not analyse this outcome (Table 4).

3.7. Supervised exercise programme plus compression compared with unsupervised exercise programme plus compression

One trial with 32 participants compared an extensive programme of physical exercises plus compression and a basic programme of physical exercises plus compression for a period of 9 weeks26 and reported only pain and ankle range of motion at 9 weeks. Time to healing, proportion with healed ulcers, ulcer recurrence, quality of life, adverse events, economic outcomes, CMP, and ankle range of motion were not reported. Due to limited reporting of outcomes, risk of bias and small sample size, the evidence was downgraded to very low quality.

3.7.1. Pain

Pain scores ranging between 0 and 2, whereby a score of 0 indicated no pain, a score of 1 indicated moderate pain, and a sore of 2 indicated severe pain requiring pain killers, was used to assess participant pain levels. Number of events and frequencies were not reported and not provided upon request (Table 4); thus, we were unable to substantiate the results.

3.7.2. Range of ankle mobility

Means and standard deviations could not be extracted from the data nor were provided upon request (Table 4); thus, we could not substantiate the results.

3.8. Protocol deviations

The databases searched differed from what was specified in our protocol. The search result initiated by researchers was checked and finalised by a librarian with substantial experience in database searching. On the suggestion of a librarian, we decided to replace the Health Technology Assessment Database with PEDro, a physiotherapy evidence database. We have also amended the date range to optimise the likelihood of identifying all previously published studies, removing the initially imposed restriction to 1990. It was decided to include only RCTs and cluster RCTs to allow for better quality of evidence. Results were presented as per protocol when available, although only outcomes and end of treatment and last follow up were reported. Additional outcomes, such as CMP function and ROAM, were also extracted from the included studies as they were relevant to the assessment of benefits of exercise in VLU populations.

4. DISCUSSION

4.1. Summary of evidence

This systematic review examined 6 RCTs published between 1990 and 2017 that measured the effects of exercise interventions on VLU healing rates, time to healing, recurrence, quality of life, pain, adverse events, and economic outcomes. The key findings are as follows: low‐quality evidence from 3 trials indicates there may be no difference in the proportion of ulcers healed after 12 weeks of a PREG plus compression compared with compression alone. Low‐quality evidence from a single trial indicates there was probably no difference in quality of life and a possible increase in the risk of adverse events. Low‐quality evidence from 2 trials also indicates that there was possibly no difference in ankle range of motion and CMP. Evidence was downgraded due to susceptibility to bias in the trials and imprecision due to the small number of participants. Recurrence, pain, and economic outcomes were not measured in these trials, and time to healing was measured only in a single trial but not fully reported.

Low‐quality evidence from a single trial indicates that a community‐based nurse counselling and behaviour modification and exercise program (Lively Legs) probably did not improve healing rates and VLU recurrence compared with standard care (attendance at an outpatient wound clinic and compression bandaging alone). The evidence was downgraded to low quality due to potential for selection bias and imprecision in the results. Therefore, there is uncertainty about the effect estimates, which further trials are likely to change. This trial did not report time to healing, adverse events, quality of life, pain, or secondary outcomes, such as CMP and ankle range of motion. Low‐quality evidence from a single trial suggests that the Ten Thousand Steps exercise programme does not result in benefits in terms of time to healing, proportion healed, and pain. Very low‐quality evidence from a single trial suggests that a supervised exercise programme plus compression may not confer additional benefits over an unsupervised exercise programme plus compression in terms of ROAM and pain.

The evidence underpinning the results is low quality at best due to bias in the trial design and imprecision. All RCTs included in this review were susceptible to selection, performance, or reporting biases. All of them were small‐scale trials, and even for the few outcomes for which we could pool data, the number of participants or events were small, leading to imprecise effect estimates.

4.2. Strengths and limitations

A noteworthy feature of trials included in this review was the wide variability in study interventions. Many different exercise regimens were prescribed. The diversity in exercise prescription is not surprising given the lack of consensus on the optimal exercise prescription for this patient population. Conversely, the wide variety in study outcomes and measurement methods is surprising. This variation precluded pooling studies and made overall conclusions regarding the relative effectiveness of exercise difficult. The short duration or complete lack of follow‐up data examining the effect of exercise on quality of life and economic outcomes in the long term is also noted. Moreover, data are lacking to support the use of exercise in preventing ulcer recurrence. A further limitation is the non‐specificity with respect to the timing and setting of the exercise intervention. Finally, poor adverse event reporting in most of the studies limits any conclusions about the relative safety of exercise, and the small samples provide insufficient power to detect meaningful differences in rates of rare adverse events.

5. CONCLUSIONS

Chronic VLU healing is a complex clinical problem that requires intervention from skilled, often costly, multidisciplinary wound care teams. There is little evidence from RCTs about which exercise interventions improve healing rates. There are suggestions that exercise interventions may result in the patients more consciously following advice, including performing exercises on regular basis and as a result improving VLU healing rates. This review summarises the best available evidence regarding the effects of exercise on time to complete healing, proportion of VLU healed, recurrence of VLU, health‐related quality of life, wound pain, reported adverse events, and economic outcomes.

There was no statistically or clinically significant change in wound‐healing outcomes. However, this systematic review provides low‐quality evidence that exercise as an intervention improves healing for adults with VLUs or at least trends in a positive direction. It is not known, however, whether the trend towards healing for those in exercise intervention groups occurred as a result of the exercise intervention because there was a lack of studies using direct measures of time to healing. This evidence suggests that positive changes in wound healing may occur despite non‐significant changes in the proportion healed and time to healing. The results show consistent support for the application and feasibility of exercise programmes using self‐management principles and/or nurse‐led counselling programmes. Direct comparisons of RCTs are not possible due to the lack of homogeneity. Future trials should focus on adequate randomisation, concealment of allocation, and blinding of outcome assessors throughout the study.

The evidence suggests that exercise may improve wound healing, quality of life, and physical functioning in adults with VLUs. Although these preliminary results are promising, the findings are based on a relatively small number of trials with significant methodological weaknesses. Furthermore, there is currently no evidence to support the use of exercise regimens to improve wound healing. We emphasise the importance of methodologically rigorous studies, which examine different exercise regimens (eg, moderate vs low‐intensity), which would help to better understand the role of exercise among adults with VLUs and to prevent VLU recurrence. We suggest that the exercise prescription should be reported in detail (frequency, intensity, time and type of exercise) to allow for determination of exercise dose response. To this end, adherence to exercise should be reported for both completion of exercise sessions (attendance) and exercise prescription (intensity and duration). Furthermore, monitoring of activity in the comparison group(s) is necessary to assess potential contamination. Consensus is required on standardised methods of assessing physical fitness and wound healing to allow for pooling of data and for comparisons across studies. Future trials should formally monitor for and report the incidence of the adverse events, such as bleeding or additional wound opening.

1.

Reason for exclusion	Study author, y
Study design was not a randomised control trial	Alberti, 2008 Azoubel, 2010 Bannister, 1968 Bartholomew, 1952A Bartholomew, 1952B Berard, 2002 Bergendahl, 1972 Brown, 2012 Brunner, 1992 Bunkan, 1963 Clarke‐Moloney, 2007 Curwen, 1953 Davies, 2007 Davies, 2008 Dembowski, 1985 Dissemond, 2005 Eifell, 2006 Falconer, 1955 Finlayson, 2014 Freund, 1981 Gottschalk, 1987 Halper, 1951 Heinen, 2006 Heinen, 2007A Heinen, 2007B Hosoi, 2002 Junger, 1994 Junger, 1998 Kan, 2001 Kelechi, 2010 Kelechi, 2014 Klyscz, 1997A Klyscz, 1997B Klyscz, 1999 Koenitz, 1984 Larcinese, 2008 Leal, 2015 Lindhagen, 1983 Lurie, 2011 McCulloch, 1998 McCulloch, 2015 Muller, 1976 Muller, 1979 Nadland, 2011 O'Brien, 2014A O'Brien, 2014B Pasynkov, 1978 Peschen, 1996 Pieper, 2012 Pillot, 1974 Pirner, 1969 Ramelet, 1991 Roaldsen, 2009 Roaldsen, 2006 Seredenko, 1992 SkavbergRoaldsen, 2011 Steins, 2000 Strolin, 2004 Stucker, 2011 Szewczyk, 2006 Tew, 2015 vonArnim, 1966 Weiss, 1991 Wissing, 1999 Wissing, 1997 Yang, 1999 Zhou, 2015
Participants did not have a venous leg ulcer	Ciuffetti, 1999 Kahn, 2011 Mangiafico, 1993 Obermayer, 2007 Padberg, 2004 Ramos‐Gonzalez, 2012 Zajkowski, 2006
Intervention was not exercise	Agu, 2004 Aquino, 2016 Araujo, 2016 Burnand, 1981 Taradaj, 2012
Article type was not a journal article	Klonizakis, 2014 Klonizakis, 2014 Anonymous, 2008 Heinen, 2010
Duplicate	Pieper, 2014

Smith D, Lane R, McGinnes R, et al. What is the effect of exercise on wound healing in patients with venous leg ulcers? A systematic review. Int Wound J. 2018;15:441–453. 10.1111/iwj.12885