Skip to main content

The Effect of Nonweight-bearing Exercise and Protocol Adherence on Diabetic Foot Ulcer Healing: A Pilot Study

Empirical Studies

The Effect of Nonweight-bearing Exercise and Protocol Adherence on Diabetic Foot Ulcer Healing: A Pilot Study


  Innovative approaches to the prevention and treatment of foot wounds in persons with diabetes mellitus are needed and preliminary research suggests that exercise therapy may increase joint mobility and blood flow. A 12-week, prospective, quasi-experimental pilot study was conducted to evaluate the potential influence of nonweight-bearing ankle exercises, and adherence to same, on the size of neuropathic, diabetic foot wounds in community-dwelling older adults.

Nineteen patients were recruited. Of those, 10 (88.9% men) were randomized to ankle exercise treatments and nine (50% men) continued their previous care regimen. Patients randomized to the exercise program were younger and had smaller wounds than those in the control group (average age 62.2 ± 8.54 versus 74.25 ± 16.25 years and measurement 0.94 cm2 ±1.89 versus 2.53 cm2 ± 3.647, respectively). Thirty percent (30%) of the patients in the exercise and 33.3% in the control group healed. The percent wound reduction between groups was not significantly different (Mann Whitney U test, P = .696). Adherence to the recommended exercise program was variable but 70% performed some exercises. The results of this pilot study provide important lessons for future studies, including the need to enroll more patients and provide more exercise guidance. Considering the increasing prevalence of diabetes mellitus and resultant complications, as well as the potential benefits of this non-invasive treatment regimen, larger studies are warranted.

Potential Conflicts of Interest: none disclosed

 Type 2 diabetes is now regarded as the most common metabolic disease worldwide and the rate of newly diagnosed adults is increasing.1 Foot ulcers develop in approximately 15% of individuals with diabetes and precede 85% of lower limb amputations in these same individuals.2,3 Diabetes can affect the vascular, neurological, and immune systems as well as biochemical processes, resulting in altered wound healing.4,5

  Despite current best practice of aggressive debridement, offloading, and glucose control, a proportion of diabetic foot ulcers often remain unhealed.6-8 Researchers need to continue looking at innovative approaches that have the potential to increase healing and reduce limb loss.2,9,10

  Exercise often is contraindicated in persons with diabetic foot ulceration because offloading the foot for the purpose of pressure reduction is a key concept related to the healing these wounds.3,11 The literature search revealed one case-control study12 (n = 7) involving use of a motorized pump to perform ankle exercises in patients with diabetes with recalcitrant wounds; the results indicated a positive healing response. Although limited, some evidence suggests that exercise may affect lower limb wound closure.2,4,12 A 12-week, quasi-experimental longitudinal, prospective randomized group pilot study was conducted to explore the effect of exercise on healing diabetic foot ulcers.

Literature Review

  The effects of hyperglycemia and nitric oxide on peripheral circulation. Intracellular hyperglycemia causes an elevation in serum and plasma viscosity13,14; that, and an increase in fibrinogen levels, can contribute to an increased risk of clot formation in persons with diabetes13 A reduction in erythrocyte deformability in persons with diabetes, which may be a result of glycosylation, can result in a reduced ability for the cells to pass through the capillaries and increase flow resistance.15 Hyperglycemia results in the inhibition of nitric oxide,15 a potent vasodilator,16 and potentially reduces peripheral blood flow. Exercise has been shown to increase shear forces and enhance the vasodilatory capacity of arteries.17 Dimmeler and Zieher,18 following a review of the literature, concluded that normal exercise hyperemia can occur even though nitric oxide production is inhibited. This potential for increased blood flow in the presence of an altered vasodilatory response, such as that seen in diabetes, may have significance for the outcome of a study assessing exercise as a healing potentiator in diabetic lower limbs.

  Exercise therapy. Exercise therapy promotes the use of muscular contraction and bodily movement to ultimately improve overall functioning.19 Exercise has been shown to cause an immediate rise in blood flow in response to muscle contractions to provide oxygen and metabolites to the actively contracting skeletal muscle.20-22 In both animals and humans, muscle blood flow has been demonstrated to increase in fewer than 1 or 2 seconds after exercise initiation, typically peaking in less than 30 seconds.20 Results of reviewing the effectiveness of exercise as therapy in preparation for the current study indicated that the type of exercise prescribed may have a substantial impact on the outcomes of therapy. In assessing exercises for the promotion of venous return, a cross-sectional study23 (n = 20) using strain gauge plethysmography was reviewed. The study involved the use of ankle dorsiflexion, plantar flexion, and subtalar inversion and eversion and concluded that an exercise combining active ankle and subtalar movements produced the greatest increase in mean and peak velocity blood flow. Active ankle flexion resulted in a peak velocity increase of 33% and active subtalar movements increased peak velocity by 38%. This improvement in venous hemodynamics was maintained for up to 30 minutes after exercise therapy concluded.

  Another cross-sectional study24 using photoplethysmography identified increased blood flow up to 51% (P = 0.003, Wilcoxon signed-rank nonparametric) after 1 minute of isometric muscle contraction and by 59% (P = 0.005) after 1 minute of concentric muscle contraction.

  Arterial disease and diabetes. Atherosclerosis development is accelerated in diabetes and appears to progress with the increasing duration of disease.25-27 The American Diabetes Association (ADA)28 suggests that the prevalence of peripheral arterial disease in persons with diabetes over the age of 50 years is 29%.

  Contracting muscles produce and release IL-6, which induces lipolysis (the oxidation of fatty tissue) and is involved in the stability of blood glucose levels during exercise. Physical activity has the potential to suppress low-grade inflammation and so protect against diseases associated with chronic low-grade systemic inflammation29 such as atherosclerosis. A meta-analysis30 and a randomized controlled trial31 outlining the effects of exercise in the management of peripheral vascular disease discuss the improvement of circulation to the lower limb.

  The effect of diabetes on joint mobility. A general reduction in joint mobility has been noted 2 years after diagnosis in persons of all ages with diabetes.32 One cross-sectional trial33 compared joint mobility in 70 persons with diabetes (35 with neuropathy and 35 without neuropathy) to 30 non-diabetic control subjects. Joint mobility, assessed at the ankle and first metatarsophalangeal (MTP) joints, was significantly reduced in the neuropathic diabetic group (P <0.0001, linear regression analysis) compared to the other two groups. Ankle joint mobility was linked to higher foot pressures in the neuropathic diabetic group and the potential for a reduction in foot wound healing. Reduced ankle motility has been found to have a deleterious effect on venous hemodynamics and impacts on potential for ulcer healing.34

  A randomized controlled study35 on the effects of an unsupervised home exercise program on joint mobility and plantar pressures was conducted in a population with diabetes (N = 10). Participants were asked to perform exercises up to three times a day and document the amount of physical therapy completed. No statistical differences were found between the treatment and control groups with regard to age, gender, duration or type of diabetes HgbA1c, body mass index, and concluding visits. Peak plantar pressures decreased an average of 4.2% in the treatment group and increased an average of 4.4% in the control group. A statistically significant difference (P = .024) was found between pressures on the dominant foot in the control and treatment groups.35 Although this was a small study of short duration (1 month), the concept of reducing peak plantar pressure in an at-risk population (persons with diabetes with sensory neuropathy on an at-home unsupervised exercise program) is tantalizing and demonstrates the need for further research.

Methods and Procedures

  Ethical considerations. The study was approved by the Behavioural Research Ethics Board of the University of Saskatchewan and considered a minimal risk trial because the use of a nonweight-bearing exercise regimen is not medically invasive and no deceptive techniques were used within the data gathering process. Participants received detailed information about potential participant benefit/risk, potential funding biases, and the random allocation procedures dictating the exercise component of the trial. Participants were provided the opportunity to refuse participation and no evidence of coercion by the researcher was noted. No financial compensation was provided to any of the study participants. Confidential information was stored in a locked cabinet and all identifying information obscured during the data gathering and dissemination process. Informed consent was obtained after the study was explained to the individual and before the data collection process was initiated. The consent procedure presented the study components in a standard grade school language format and participants could withdraw at any time.

  Design. The researcher used a longitudinal, prospective design to examine potential changes related to a specific intervention within two comparable populations over time.36 This type of quasi-experimental design has been used previously within wound healing research to allow for a comparison of a control group with the group being manipulated37,38 within a pragmatic environment. The intervention or independent variable was the use of nonweight-bearing exercise in a population of individuals with diabetic foot ulceration. The dependent variable was the effect of the exercise program on healing (percent reduction in ulcer size) in comparison to a non-exercising population of individuals with the same diagnosis.

  Methods. Participants were individuals with foot ulcers referred to podiatric services within the Saskatoon Health Region who met the following inclusion criteria: 18 years of age or older with diabetes, ulceration, sensory neuropathy, and the ability to provide informed consent in English. The exclusion criteria were cognitive impairment (cognition was not formally assessed — if patients were able to answer questions and demonstrate an appropriate understanding, they were included in the study); infection (measured by greater than four pathogens by Gram stain); and ischemia. The latter was assessed clinically and patients who did not have palpable pulses were excluded from study participation.

  Protocols of wound care were not standardized. It was assumed that involving podiatry patients would limit variability in practitioner approach to wound care, permitting evaluation of the effects of exercise or no-exercise between groups. The use of this directional hypothesis is an attempt to reduce one of the confounding variables: the differences in approach to diabetic foot ulcer management among different healthcare service providers. The use of a single group of healthcare providers across both the control and experimental groups should allow for comparison of the proposed difference — ie, exercise — between the two groups.

  Before initiating the data collection process, numbered envelopes starting with the experimental arm and alternating with the control arm of the study were prepared. The envelopes contained the data collection tools specific to the participant’s assignment. The assignment process involved including every other person who fit the study inclusion criteria. Once written informed consent was obtained and before study arm assignment, monofilament testing was performed to ensure the person had neuropathy. Participants assigned to the control arm of the study were asked to continue their care as they had done before study involvement. Those assigned to the experimental arm received a sheet describing a selection of exercises with explanations. The exercise protocol was established as a result off the literature search. Exercises demonstrating, in cross-sectional trials, an increase in blood flow and that did not involve weight-bearing, included simple ankle inversion, eversion, flexion, and extension were prescribed. The researcher had participants demonstrate the exercises before leaving the clinic to ascertain their ability to complete them without pain or other obvious limitations. No time frame was established for completion of the exercise routine and because the study was home-based and entirely voluntary, there were no restrictions placed on performance. An exercise journal with self-completion information also was provided. Participants were asked to perform four ankle exercises 10 times each twice a day and note the frequency in the exercise journal. They were asked to bring the journal back with them for review on subsequent clinical visits. The exercise program used in this project was designed to be self-supervised. Studies using a model of self-supervised exercise have been conducted35 but in this study the selection of self-supervision was related to manpower limitations.

Data collection
   Wound measurement. Wounds were traced onto an acetate grid; length and width were determined with the use of a ruler and area was calculated by using the formula of an ellipse (0.5 x length of major axis) x (.0.5 x length of minor axis) x (π).39 The wounds were measured every 4 weeks for a total individual assessment period of 12 weeks. The use of a 4-week time frame for the routine assessment of wound healing, as well as the total time of 12 weeks, has been used in other studies.37,39,40 It has been suggested that simple measurements such as length and width and the calculation of area based on the formula for an ellipse provide an excellent representation of planimetric wound area.39 Planimetric measurements have been shown to be more accurate than ruler diameter-based measurements or square counting on acetate grids.39-42 A study comparing computerized planimetric wound measurements (n = 260) to simple wound measurement of length times, width, or even the length or width alone concluded that simple wound measurement represented an adequate and appropriate measure of wound size in the clinical setting (P <0.001, Pearson, Spearman and Kendall correlation coefficients).39

  When comparing tracing to digital planimetry, both methods were found to have good inter- and intrarater reliability.43

  Sensory neuropathy. The number of insensate sites used to define sensory neuropathy is debatable; the definition of an insensitive foot ranges from one to four sites.44 The operational definition of sensory neuropathy for the purpose of this study was a loss of sensation to any single area of the foot. Nine sites on the foot were tested: the plantar aspects of the first, third and fifth digits; the first, third and fifth metatarsal heads; the medial and lateral areas of the midfoot; and the dorsum of the midfoot. Each site was tested a second time if initial application of the monofilament could not be felt. Sensory loss was estimated as a percentage rounded to the nearest tenth of a percent — eg, if the individual had sensation at only one of the nine sites, a 90% loss of sensation was noted. The plantar aspect of the forefoot provides the best discernment between individuals with or without ulceration, with the heel providing essentially no discrimination45 to establish the presence of sensory neuropathy; for this reason, the heel was not used as a test site. The specificity of the 5.07/10 g monofilament to detect sensation loss at five sites (great toe, dorsum, plantar aspect of the first and fifth metatarsal, and the plantar aspect of the arch) is approximately 90%, but results and sensitivity vary between sites from 44% to 71% and reported specificity is 98% to 100% at all five sites tested.45 All sensation testing was performed by the researcher in an attempt to improve the consistency of the testing procedure.

  Infection protocol. To assess study eligibility and ascertain the absence of clinical infection, initial assessment of the wound and surrounding tissues included clinical signs and symptoms of infection. After cleansing the wound, a swab culture was obtained. The use of a Gram stain has been shown to reliably predict a bacterial load of >10 colony forming units (CFU) per gram of tissue if one micro-organism could be viewed on a slide preparation.46,47 Research in chronic leg ulcers indicates that no single bacteria or group of micro-organisms is more detrimental to wound healing than any other but the possibility of healing is reduced with four or more pathogens present.46 If the swab results came back with a result of four or fewer pathogens, the individual was included in the trial.

  Dartmouth COOP Functional Assessment Chart/WOCNA.48 The Dartmouth COOP Functional Assessment Charts/WOCNA (COOP/WOCNA) was completed for all study participants. The six COOP/WOCNA charts are a generic health status measure that require the individual to subjectively rate the listed areas of his/her functional activity within the last 2 weeks using five numbered response categories within each of the charts. The scores are inversely related to individual function, with a high score indicating functional limitation. The COOP/WOCNA charts were initially developed for general practitioners to use as a quick and simple method to determine health problems within an outpatient population and have been used in several studies and administered to more than one patient population experiencing a chronic disease with varied results.49,50 No research was found within the literature search process to support the validity of the use or non-use of this functional assessment tool in the population of patients diagnosed with diabetes. The original charts have been tested against other quality-of-life questionnaires and validity was found to be good.51

  Chronic medical conditions or use of alternative therapies. Study participants were asked if they had any chronic medical conditions that might affect their ability to exercise or if they had ever used any type of alternative therapy in the past in an attempt to improve their health.

  Data analysis. Descriptive statistics (mean, median, and standard deviation) utilizing SPSS 15.0 for Windows Grad Pack (Chicago, IL) were used to analyze the data and compare patient demographic variables and functional testing outcomes. Frequency distributions were used to describe patient gender, Lower Extremity Amputation Prevention (LEAP) scores, swab results (number of pathogens), participant comorbidities, and history of using alternative therapies. For patients in the treatment arm of the study, exercise frequency and wound healing outcomes were evaluated, sans frequency distributions due to small sample. The Mann Whitney U test was used to compare change in wound size from week 0 to week 12 between the experimental and control groups.


  The sample comprised 19 participants — 10 in the experimental and nine in the control groups — which is the number of individuals seen by podiatry services for the management of a diabetic foot ulcer within a 3-month period meeting the eligibility criteria. The experimental group was predominantly male (eight, 88.9%) with a mean age of 62.2 years (median age 60, range 49 to 74 years) and a mean COOP/WOCNA chart score of 17.2 (median 16, range 13 to 24) (see Table 1). In the experimental arm of the study, swab culture results showed that three wounds (33.3%) contained one pathogen, one (11.1%) had two pathogens, and one (11.1%) had four pathogens (see Table 2). Lab results were unavailable for four (44.4%) participants. Use of alternative therapies within the experimental group included meditation (one person), therapeutic sheepskin (one), and seven (77%) who used no alternative therapies (see Table 1).

  The control group included four men and four women. Their mean age was 74.25 years (range 54 to 94 years) and the mean COOP/WOCNA chart score was 18.75 (median 19, range 12 to 24) (see Table 1). The swab cultures of two control group participants (25%) showed no pathogens, two (25%) had two pathogens, one (12.5%) showed three pathogens, and lab results for three (37.5%) individuals were not available (see Table 2).

  Prevalence of comorbidities between the two groups was substantially different (see Table 1). For example, in the experimental group, 30% of participants had arthritis, compared to 38% in the control group; experimental group members also had cardiovascular accident, previous back surgery, and history of disc herniation (10% each). No control group participants used alternative therapy.

  Exercise frequency in the experimental group varied from “unknown” to three times per day; two individuals performed the exercises two times per day as requested (see Table 3). Sensory neuropathic changes in both groups ranged from 20% to 100% loss of sensation with 50% of all participants having 100% loss (see Table 1).

  Of the 10 experimental group participants, nine (90%) experienced a reduction in wound size of 26% to 100% over the 12-week course of the data collection regardless of exercise frequency (see Table 4). The one individual whose wound deteriorated was diagnosed with osteomyelitis 4 weeks into the data collection. In the control group, three individuals had an increase in wound size (31%) over the course of the study (see Table 5). The difference in percent wound reduction between the two groups was not statistically significant (Mann-Whitney U test, one tailed, P = .696). The mean Dartmouth and LEAP scores also were not significantly different. (see Table 6).


  The purpose of this quasi-experimental longitudinal, prospective randomized group pilot study was to explore the effect of exercise on healing diabetic foot ulcers. Although the results are inconclusive, lessons from this pilot study may help guide future research. Because the sample size was small, the unexpected variation between the two groups with respect to age, gender distribution, comorbidities, and wound culture results may have affected the overall study outcome. Three of the older participants exhibited negative healing results compared to only one negative outcome in the parallel younger group. Other variables that may have affected healing outcomes such as the duration or type of diabetes, HgbA1c, body mass index, length of wound duration, diet, and lifestyle were not collected and should be included in future studies. Also, although the presence of bacteria within diabetic foot ulcers is not an unexpected finding,46 assessing the presence of infection in chronic wounds remains a clinical challenge.52 For reasons of laboratory workload, access, and timing, Gram-stain results were considered adequate to establish the presence of micro-organisms. Because a substantial number of swab results were never received by the researcher, the culture data results should be interpreted with caution.

  The length of this study also may have affected the outcome. Using a 12-week time frame to assess reduction in wound area was believed to be adequate to evaluate change, especially because it has been suggested that a 4-week time frame has been shown to be sufficient in other research to demonstrate efficacy.40

  Self-supervised exercise programs have the potential for adherence issues and the reporting system used as a measure is problematic. During this project, participants were asked to use a self-report method in the form of a journal; this same methodology has been reported in other research.53,54 A more accurate method of assessing adherence would have been routine observation of the study participants performing the exercises (ideally, in clinic with set time periods), rather than just relying on participant honesty. In this study, using the measure of some attempts to participate as adherence, 70% of the individuals were at least partially adherent. Results of a longitudinal correlational trial53 indicate that home-based exercise programs are more likely to achieve compliance, as are therapeutic exercises, when compared to preventative routines; the current study supports the fact that patients may adhere to exercise routines when they are therapeutic and home-based and less likely to be followed when they are simply preventative. None of the participants within the exercise cohort indicated an inability to complete the exercises, nor was performance ability used as an excuse for nonadherence.

  In addition to the potential effect on healing, the presence of comorbid conditions may negatively influence participation in a prescribed exercise regimen55; in the experimental group, six participants (60%) indicated the presence of a comorbid condition and of those two (20%) had looked to alternate therapies in the past, possibly indicating a willingness to try new things.

  Overall, the percentage of wounds healed was higher (79.8%) than reported in the literature.7 In the experimental group, 90% of the wounds progressed toward healing or healed, compared to 62.5% in the control group. Some research suggests wound size affects healing with increased percentage of closure for “small” (<1 cm2) wounds.56 Seven wounds in the experimental (70%) and four (50%) in the control group were <1 cm2 and the study sample was too small to control for wound size in the data analysis.

  Identifying wound margins has been established as one of the greatest sources of measurement error42; when tracing small wounds, the thickness of the writing instrument may interfere with interpretation of the wound margins43 and contaminate the measurements. However, if the inclusion criteria had been to only accept individuals with wounds >1 cm2, most of the wound patients in the clinic would not have been eligible to participate. A second potential source of measurement error was in the calculation of wound area from a tracing; access to either digital planimetry or stereophotogrammetry could potentially have increased the accuracy of the data collected. Attempts to use the Visitrak™ system (Smith& Nephew, Largo, FL) for this study failed because the device is unable to accurately measure wounds <1 cm2.

  The use of the disposable LEAP monofilaments also has the potential to limit study outcome interpretations. When evaluated in a research facility, variability with an average deviation of >18% between 50 LEAP monofilaments tested was noted, compared to only a 10% variation with the Semmes-Weinstein monofilaments.57 The disposable monofilaments were used because they were readily accessible and inexpensive and because disposable equipment minimizes the risk of cross-contamination. Future studies should weigh the cost/benefit of using disposable versus non-disposable monofilaments.

  It is known that the reliability of the Dartmouth charts may be affected by the method of administration. The literature proposes that self-completion of the document may reduce potential interviewer bias.58 In this study, the charts were administered by the researcher, which may influenced the reliability of the data collected. Because the functional assessment was conducted only at the start of the study, the researcher already had confounded the assessment process by interviewing the participants before recognizing this was a study limitation.

  Although using a group of patients who were treated by one type of specialist (ie, podiatrists) has the potential advantage of limiting the variability in practitioner approach to wound care, this assumption was not validated and limits the generalizability of the results. Finally, within the control group, the relationship between exercise frequency and changes in wound area was not explored.


  The results of this pilot study to compare the impact of exercise to no exercise on wound healing in the lower limb of persons with diabetes were inconclusive. Designed as a pragmatic trial involving community-dwelling participants receiving regular care, the sample size was too small to account for enrollment, standard practice, exercise adherence, and potential study assessment variations. However, innovative approaches to diabetic foot wound prevention and closure are needed, including research investigating the potential of exercise to increase joint mobility and reduce gait related foot trauma. The results of this pilot study may help researchers develop robust protocols to explore the effects of these interventions.

Ms. Flahr is an Equipment and Product Standardization Nurse—Skin and Wound with the Saskatoon Health region. Please address correspondence to: Donna Flahr, RN, BSN, MSc, 266 Highbury Court, Saskatoon, SK, S7H4W3, Canada: email: