Characteristics of treatment responses to physical exercise and diet in diabetic peripheral artery disease patients: a systematic review

Abstract

Purpose

Peripheral artery disease (PAD) affects walking performance and increases risk of amputation. Diabetes mellitus (DM) not only contributes to disease progression but may also affect treatment effects. Though physical exercise is crucial in PAD management and diet optimization may reduce risk of atherosclerosis, their effects and variations in diabetic PAD patients remain unclear. The present systematic review aims to summarize features of treatment response to physical exercise and diet management in diabetic PAD patients, and provide crucial information on optimization of PAD management of these patients in clinical practice.

Methods

A systematic review was performed on human studies investigating effects of physical exercise and diet on development and improvement of diabetic PAD patients by searching PubMed, Embase, Scopus, Web of Science and CINAHL. The review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA). The certainty level of evidence was assessed using the Grading of Recommendations, Assessment, Development, and Evaluation (GRADE).

Results

A total of 18 studies were included, 15 of which evaluated effects of exercise on diabetic PAD, and three evaluated correlation between diet and diabetic PAD. Supervised exercise therapy (SET) and home-based exercise were able to improve ankle-brachial index (ABI) and blood flow velocity in diabetic PAD. In contrast, walking performance such as the six minute walk distance (6MWD), maximal walking time and distance, peak walking time (PWT) and claudication onset time (COT) were not significantly improved or presented conflicting results. Administration of multiple exercises or alternative interventions such as shock wave therapy or low-frequency pulsed electromagnetic field therapy (LFPMT) may enhance improvement of ABI. Adherence to Mediterranean diet was associated with blood flow and intermittent claudication, as well as risk of development of PAD. GRADE analysis revealed very low to moderate grade evidence for findings on ABI, flow velocity and functions, mainly due to study types, number of studies available, small sample size, moderate risk of bias and conflicting results. Similarly, low grade evidence was also revealed for effects and association of diet on risk of PAD due to study type and number of studies available.

Conclusion

Diabetic PAD patients may present reduced treatment response to physical exercise in walking performance and other functions, which may be improved by combination of multiple physical exercises or alternative interventions with or without combination of exercise. Mediterranean diet may be associated with beneficial effects, which requires further investigations. These findings were of very low to moderate level of certainty, and therefore, further validation is warranted in the future.

Supplementary information

The online version contains supplementary material available at https://doi.org/10.1007/s12020-026-04836-9.

Introduction

Peripheral artery disease (PAD) is a common disease characterized by stenosis or occlusion of arteries other than coronary arteries and cerebral vessels, affecting over 200 million people worldwide [1]. The condition could be categorized into asymptomatic PAD, chronic symptomatic PAD, chronic limb-threatening ischemia (CLTI) and acute limb ischemia [2]. PAD affects walk performance and functional status, increases risks of amputation [2], myocardial infarction and stroke [3, 4]. Effective treatments may assist with prevention of major adverse limb events as well as cardiovascular events [2].

Diabetes mellitus (DM) was present in around 55% of PAD patients [5]. The disease is not only a risk factor for PAD [6] but also a contributor for progression to CLTI and increased risk of amputation [7]. In addition, PAD patients with DM present approximately two-fold higher risk of mortality than PAD patients without DM [8], highlighting the necessities of efficient interventions. Exercise and diet are considered main components of lifestyle that affect development and progression of PAD [9, 10]. An efficient tailored intervention consist of these components was desired to combat for the cohort.

Structured exercise including supervised exercise therapy (SET) and community-based programs constitutes a core component in PAD management [2] to improve walk performance and functional status [11, 12]. Though much emphasis on efficient and effective interventions has been addressed in diabetic PAD patients, difficulties are present in management of this population due to pathophysiologic features of the disease. For instance, peripheral neuropathy causes extra damage to vessels and muscles, affects gait and increases risk of fall, which cause variations in treatment response and adherence. In addition, the frequently present diabetic foot ulcers may affect tolerability and adherence to the interventions.

A healthy diet reduces risk of atherosclerosis [13], and Mediterranean diet, which is high in plant-based food, olive oil as fat, and low in red and processed meat, is associated with improvement in glycemic control and reduced cardiovascular risk in type 2 DM (T2DM) [14]. However, its direct contribution to development and presentation of diabetic PAD remains to be determined.

Currently the characteristics of treatment response to physical exercise and diet, as well as potential factors affecting these responses specifically in diabetic PAD patients remain unclear. Understanding these response features is critical in optimizing treatment strategies to achieve a variety of goals in this population in clinical practice. The present systematic review aimed to summarize characteristics of response to physical exercise and diet in clinical studies evaluating diabetic PAD patients, determine whether selection and combination of these interventions lead to achievement of different goals, and identify factors to be considered for optimization of strategies in management of diabetic PAD patients.

Methods

Search strategies and selection of eligible studies

The literature research was performed on PubMed, Embase, Scopus, Web of Science and CINAHL using search terms (‘lower extremity artery disease’ OR ‘lower extremity arterial disease’ OR ‘peripheral artery disease’ OR ‘peripheral arterial disease’) AND (‘diabetes’ OR ‘diabetic’ OR ‘hyperglycaemia’ OR ‘hyperglycemia’ OR ‘impaired glucose tolerance’ OR ‘impaired fasting glucose’ OR ‘insulin resistance’) AND (‘exercise’ OR ‘training’ OR ‘diet’ OR ‘dietary’ OR ‘nutrition’) to identify literatures published in English until 14 May 2025 (Supplementary Table S1). A total of 2461 publication records were identified and imported to EndNote 20. Duplications were removed, followed by exclusion of irrelevant types of articles, including book and book chapters, case report and series, comment, opinion and communication, conference paper, editorial, erratum and correction, retraction, guideline and consensus, language, letter and response, perspective and hypothesis, protocol, rational and design, short survey, note and report, as well as review, systematic review and meta-analysis. Subsequent screening was achieved by reviewing titles and abstracts to further exclude irrelevant studies. In addition, studies with full text unavailable were also excluded. Irrelevant studies and non-human studies were further excluded. Studies evaluating treatment response from exercise and diet in diabetic PAD, including diabetic PAD with peripheral neuropathy or polyneuropathy, as the total cohort or subcohort were included to determine features of treatment response to exercise and diet in diabetic PAD and non-diabetic PAD. Studies comparing response from exercise and other interventions in diabetic PAD were also included to determine features of treatment response to exercise in diabetic PAD as compared to other interventions or standard care. The present systematic review was reported according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) (Fig. 1, Supplementary Table S2), but was not prospectively registered. Two reviewers reviewed the records independently. When discrepancies were present, a third reviewer was consulted, and a final agreement was reached following discussion.

Fig. 1.

PRISMA flow chart for database search. *Excluded article types include book and book chapters, case report and series, comment, opinion and communications, conference papers, editorial, erratum and corrections, retractions, guideline and consensus, language, letter and response, perspective and hypothesis, protocol, rational and design, short survey, note and report, review, systematic review and meta-analysis. **irrelevant studies are those met exclusion criteria

Data extraction

Demographic information including age (mean ± standard deviation [SD] or median) and sex proportion (presented as proportion of female participants), as well as sample size, disease type, grouping, intervention, assessment and outcomes were extracted using a standardized form. When information on average age and SD was available for subgroups in the study, the average age presented as mean ± SD for the total population was calculated. Data extraction was established by two reviewers independently. As study characteristics such as study types, populations, interventional components, training protocol and comparators, as well as outcome measurements varied greatly among studies, presenting great heterogeneities among these studies, results were synthesized narratively, and no quantitative synthesis was performed.

Quality assessment

The risk of bias was evaluated using the revised Cochrane risk-of-bias tool for randomized trials (RoB2) for randomized controlled trials (RCTs), risk of bias in non-randomized studies of interventions (ROBINS-I) for non-RCTs, Joanna Briggs Institute (JBI) critical appraisal for cohort studies and cross-sectional studies, and Newcastle-Ottawa Scale (NOS) for case control studies. Evaluation was performed by two reviewers independently. When discrepancies were present, a final agreement was reached following discussion or consultation of a third reviewer.

The Grading of Recommendations, Assessment, Development, and Evaluation (GRADE) was utilized to assess level of certainty of evidence. Main results were classified as high, moderate, low or very low [15].

Results

Study characteristics

Eighteen studies were included, of which nine were RCTs [16–24], three were interventional but not randomized studies [25–27], four were cohort studies [28–31], one was case control study [32], and one was cross-sectional study [33]. Of these eligible studies, 15 evaluated effects of exercise on diabetic PAD with a total of 1501 samples [16–29, 31], three studies evaluated relationships between diet and PAD with T2DM with a sample size of 1718 [30, 32, 33]. Within studies evaluating effects of exercise on diabetic PAD, 14 described age in the format of mean ± SD, median or age range, with average age ranging from 46.9 to 77.9 years [17–26, 28, 29, 31, 33]. Two studies reported age as median [16, 30], one study reported as age range [27], while one study did not specify information on age [32]. Constitution of sex was reported in 17 studies. In studies evaluating exercise, the proportion of female participants ranged from 20.5% to 100%. In studies evaluating nutrition, the proportion of female participants ranged from 28.5% to 49.5% (Table 1).

Table 1.

Observational studies involving dietary-related components

Ref	Disease type	Sample size	Age (years)	Sex (%F)	Observation	Outcome 1	Outcome 2	Outcome 3
Case control study
[32]	DM +/-PAD (33.33% DM + PAD)	432	/	28.5	Mediterranean dietary pattern	FFQ dietary score • higher FFQ associated with reduced PAD risk	Associations • Diabetes duration, hypertension, butter consumption associated with increased PAD risk	FFQ dietary score • Improve predictive value of DM duration and HTN-based models
Cohort study
[30]	DM +/- PAD 2.11% DM + PAD	1112	DM + PAD 61.37 (6.26) DM-PAD 60.72 (6.90)	49.5	Dietary intake and lifestyle	Prevalence • 12.2% individuals developed PAD	Lifestyle • Smoking associated with increased PAD risk	Dietary intake • Higher intake of fish and shellfish associated with decreased PAD risk
Cross-sectional study
[33]	DM +/-PAD (44.83% DM + PAD)	174	69.6 ± 8.9	38.5	Mediterranean diet adherence	Adherence to mediterranean diet • Associated with ABI, DFU, intermittent claudication and pedal pulse absence	MEDAS-14 final score • Associated with DFU, intermittent claudication and posterior tibial pulse	Food associated with lower PAD • Nut, white meat preference, sautéed dish

+: with; -: without/: not indicated

ABI = ankle brachial index; DFU = diabetic foot ulcer; DM = diabetes mellitus; FFQ = food-frequency questionnaire; HTN = hypertension; MEDAS = mediterranean diet adherence screener; PAD = peripheral artery disease

Types of exercise interventions

SET was administered by 9 studies [16, 19, 23–27, 29, 31], and home-based exercise was utilised by 4 studies [16, 17, 20, 28]. Three studies administered institutional exercise training but did not specify whether under supervision throughout the training program [18, 21, 22]. For institutional exercises, the main type of exercise was walking training on a treadmill [16, 18, 19, 24–26, 29, 31] or unspecified walking [21]. Among these studies, 4 studies performed walking in combination with other exercises or interventions [19, 21, 24, 29]. Specifically, walking was performed with resistance training [24], limb strengthening and balance [21], low-frequency pulsed-electromagnetic field therapy (LFPMT) [19] or functional electrical stimulation (FES) [29]. Other interventions included active exercise (leg movements) combined with resistance [23] and Buerger-Allen exercise (BAE) [27]. For the 4 home-based exercise, 3 studies employed walking as the type or major type of exercise [20, 22, 28], while one study employed exercises for muscle work [17]. In addition, one administrated exercise in combination with injection of glargine and oral intake of glimepiride [22] (Table 2).

Table 2.

Demographic, intervention and outcomes of studies involving exercise interventions

Ref	Disease type	Sample size	Age (years)	Sex (%F)	Intervention	Frequency	Duration	Comparator	Outcome 1	Outcome 2	Outcome 3	Outcome 4
Randomized controlled trials (RCTs)
[16]	PAD +/-DM (41.67% PAD + DM)	60	66 (12)	53.0	SET (treadmill walking)	3 days/week; 15 min/session for week 1–2; plus 5 min biweekly until 40 min/session	3 months	Home-based exercise	COT after exercise • Improved in PAD-DM • Improved in male but not female PAD + DM • Change in COT: male> female • Change in COT associated with No. of strides/session	PWT after exercise • Improved in PAD-DM • Improved in male but not female PAD + DM • Change in PWT: male> female	Non-response rate • PAD + DM > PAD-DM • Female> male	/
[17]	DM + PAD	68	53.0 ± 12.0	44.1	Exercise	2 times/day	20 weeks	Placebo (Sham ultrasound electrotherapy)	ABI after 20 and 24 weeks • Improved in exercise group • No improvement in placebo group	Flow velocity after 20 and 24 weeks • Improved in exercise group • No improvement in placebo group	Fibrinogen, Hb, HDL, HbA1c • Improved in exercise group • No improvement in placebo group	Cholesterol • Improved in exercise group after 20 but not 24 weeks
[18]	DM + PAD	60	51.2 ± 4.1	38.3	Shock wave	2000 pulse, 4 Hz, twice/week	8 weeks	Treadmill walking	WD & WT (pain-free & max) • Improved in both groups • Value following treatment: Shock wave > treadmill walking	ABI • Improved in both groups • Value following treatment: Shock wave < treadmill walking	/	/
[19]	DM + PAD#	24	54.5 ± 7.3	50	Exercise (M-HIIT + LFPMT or LFPMT + M-HIIT)	M-HIIT (treadmill walking): 4 intervals, 4 min each, twice weekly LFPMT: 24 min/session, twice weekly	4 weeks	Control (no M-HIIT or LFPMT)	FB after 4 and 8 weeks • Improved in both exercise groups • Level of improvement: M-HIIT + LFPMT group > LFPMT + M-HIIT group	ABI after 4 and 8 weeks • Improved in both exercise groups • Level of improvement: LFPMT + M-HIIT group > M-HIIT + LFPMT group	/	/
[20]	DM + PAD	145	66.5 ± 10.1	31	Exercise (home-based walking)	50 min/session, minimum 4 days/week	24 weeks	Control (biweekly phone call)	Change in maximal treadmill WD after 6 months: • No difference between groups	Change in treadmill WD until onset of pain after 6 months: • No difference between groups	Change in walking speed • Exercise > control	Change in mental health • Exercise > control
[24]	DM + PAD	80	46.9 ± 2.9	/	Concurrent training	30–60 min/ session, 3 session/week	12 weeks	Aerobic training	HbA1c • Improved (value decreased) in both groups • Improved (value decreased) more in concurrent group	ABI • Improved in both groups • Improved more in concurrent group	/	/
[21]	DM + PAD	145	60.4 ± 11.4	36	Aerobic exercise +strengthening	3 sessions/week	12 weeks	Balance exercise	ABI & TBI after 12 weeks • Improved in aerobic group • Level of improvement: aerobic > balance	6MWD and WIQ components after 12 weeks • Improved in aerobic group • Level of improvement: aerobic > balance	WHOQOL-BREF (physical functioning, psychological and social relationships) after 12 weeks • Improved in both groups • Level of improvement: aerobic > balance	FPG&HbA1c after 12 weeks • Improved (value decreased) in aerobic group • Level of improvement: aerobic > balance
[22]	DM + PAD	130	63.4 ± 5.4	43.1	Intervention (Glargine injection + Glimepiride oral + aerobic exercise)	/	12 weeks	Control (Glargine injection + Acarbose oral)	FPG, 2hPG, HbA1c, HOMA-islet, HOMA-IR after treatment • Improved in both groups	Thrombelastogram indices after treatment • Better in Intervention group	ABI after treatment • Intervention > control	Rate of FMA improvement: • Higher in Intervention group
[23]	DM + PAD	15	77.9 ± 6.2	100	HVES + CSWD + PE	/	/	Crossover	Blood flow velocity • Reduced in femoral and popliteal arteries after HVES and PE • Reduced in posterior tibial artery after PE	/	/	/
Interventional but not randomized studies
[26]	PAD +/-DM (SET: 36.6% PAD + DM; Attention control: 39.74% PAD + DM)	309	67.9 ± 9.3	42.7	SET (treadmill walking)	50 min/session; 3 times/week	24 weeks	Attention control (education)	6MWD • Improved in SET group • No change in PAD+DMLevel of improvement: higher in those with baseline 6MWD < 334 m	Change in max treadmill WT • Improved in SET groupNo change in PAD + DM	/	/
[25]	PAD +/-DM (43.18% PAD + DM)	44	72.3 ± 7.1	47.7	SET (treadmill walking or alternatives)	2–3 sessions/week	12 weeks	/	SPPB total score • Improved after 12 weeks	6 MW- claudication onset distance and total distance • Improved after 12 weeks	WIQ-distance, WIQ-speed, PROMIS-mobility • Improved after 12 weeks	Sedentary behavior variables and activities • No change after 6 & 12 weeks
[27]	DM +/- PAD* (74% DM+ right PAD; 70% DM+left PAD)	50	30–75	42	BAE	5 times/day	4 days	/	ABI • Improved after BAE	MNSI • Improved after BAE	/	/
Cohort studies
[31]	PAD +/-DM (48.15% PAD + DM)	27	67.6 ± 3.0	40.7	SET (treadmill walking)	30–40 min /session; 3 times/week	12 weeks	/	BAFMD after SET • Improved in PAD-DM • No change in PAD + DM	COT after SET • Improved in PAD-DM • No change in PAD + DM Correlation between COT change and plasma nitrite flux change • Positive correlation in PAD-DM • No correlation in PAD + DM	PWT after SET • Improved in PAD-DM & PAD + DM	Pain-free WT after SET • Improved in PAD-DM • No change in PAD + DM
[28]	DM + PAD	202	69.0 ± 10.5	31.5	Exercise (TiTo home-based exercise + intermittent walking)	10 min/session; twice/day	24–48 weeks	Control (healthy & active lifestyle)	5-year survival • Exercise > control	All-cause hospitalization • Exercise < control	Peripheral revascularization • Exercise < control	Lower limb amputation • Exercise < control
[29]	PAD +/-DM (50% PAD + DM)	142	70.5 ± 6.4	20.5	FES + walking	5 days/week	8 weeks	/	6MWD • Improved immediately but no change after 3 months in PAD + DM	PPI • Improved after 3 months in PAD + DM	PADQOL • Improved after 3 months in PAD + DM	ICQ • Improved after 3 months in PAD + DM

+: with; -: without; /: not indicated; *DM patients with peripheral neuropathy symptoms included; #DM patients with polyneuropathy included

2hPG = 2 h postprandial blood glucose; 6MWD = six minute walk distance; 6 MW = six minute walking; ABI = ankle brachial index; BAE = Buerger-Allen Exercise; BAFMD = brachial artery flow-mediated dilation; COT = claudication onset time; CSWD = continuous short wave diathermy; DM = diabetes mellitus; FB = functional balance; FES = functional electrical stimulation; FMA = Fugl-Meyer motor function score; FPG = fasting plasma glucose; Hb = hemoglobin; HbA1c = glycosylated hemoglobin; HDL = high-density lipoprotein; HOMA-islet = insulin function index; HOMA-IR = insulin resistance index; HVES = high-voltage electrical stimulation; ICQ = intermittent claudication questionnaire; LFPMT = low-frequency pulsed-electromagnetic field therapy; M-HIIT = moderate-to-high-intensity interval-training; MNSI = Michigan neuropathy screening instrument; PAD = peripheral artery disease; PADQOL = quality of life questionnaire for peripheral arterial disease; PE = physical exercise; PPI = perceived pain intensity; PROMIS = patient-reported outcomes measurement information system; PWT = peak walk time; SET = supervised exercise therapy; SPPB = short physical performance battery; TBI = toe brachial index; TG = triglyceride; TiTo = test in – train out; WD = walking distance; WIQ = walking impairment questionnaire; WT = walking time

Intervention intensity, frequency and duration of exercise

For studies that employed walking during institutional exercise, training intensity was mostly determined according to tolerability of onset or change in severity of claudication pain [16, 21, 31], though gradual increase of intensity was reported by two studies [18, 26]. The training duration for each session varied from 30 to 60 min per session [16, 19, 21, 24, 26, 29, 31], while two studies did not specify training duration per session [18, 25]. Training frequency was more uniform, with 7 studies reporting three times per week [16, 18, 21, 24–26, 31]. A frequency of twice or 5 times per week have also been employed by two studies [19, 29]. A total training duration of 3 months was employed by 5 studies [16, 21, 24, 25, 31]. Duration of one month [19], two months [18, 29], and 6 months [26] have also been utilised (Table 2).

For studies that administrated combination of active and resistance exercise, the active exercise included 3 sets of 15 repetitions of a series of movements, and the resistance exercise included 3 sets of 10 repetitions with resistance of 50%, 75% and 100% one-rep max (1-RM) [23]. For the BAE, the exercise was administrated 5 times per day for four consecutive days [27] (Table 2).

For home-based exercise that employed walking exercise, training duration for each session varied from 10 min [28] to 50 min [20] per session, and frequency varied from twice per day [28] to 4 days per week [20]. The total treatment duration for home-based exercise was longer than walking on a treadmill, ranging from three to 12 months [20, 22, 28] (Table 2).

Measurements

Lower extremity perfusion was evaluated by 11 studies [17–19, 21–27, 31]. Specifically, ankle-brachial index (ABI) was measured by 11 studies [17–19, 21–24, 26, 27, 31, 33], by using a handheld Doppler probe and calibrated mercury sphygmomanometer [17–19, 21, 23, 24, 26, 33] or manually measured using a blood pressure cuff for lower extremity perfusion [22, 27]. Meanwhile, toe brachial index (TBI) was observed in one study [21], by placing a photoplethysmography (PPG) sensor at the great toe along with the blood pressure cuff for measuring the systolic pressure for the brachial and calculating the TBI [21]. Arterial blood flow velocity was measured for femoral [23], popliteal [23], posterior tibial [17, 23] and dorsalis pedis [17] arteries, by using a Doppler probe connected to a plethysmograph [17, 23]. In addition, brachial artery flow-mediated dilation (BAFMD) of brachial artery was also evaluated by one study [31], which was obtained using high resolution ultrasound and a 7.5 MHz linear array transducer as described in a previous publication of this research group [34] (Table 2).

Severity of claudication and walking ability was assessed by 7 studies [16, 18, 20, 25, 26, 29, 31]. Maximal treadmill walking distance and time was measured using Gardner protocol [18, 19, 26]. In addition, treadmill claudication pain distance [18] and six minute walking distance (6MWD) [26, 29] were also measured to evaluate walking ability. Peak walking time (PWT) and claudication onset time (COT) [23] were assessed to evaluate severity of claudication [16, 31]. PAD-related and health-related quality of life was evaluated by PAD quality of life (PADQOL) [29], intermittent claudication questionnaire (ICQ) [29], medical outcomes study (MOS) 36-Item short form survey (SF-36) and World Health Organization quality of life – BREF (WHOQOL-BREF) quality of life questionnaire [21]. Furthermore, other PAD-related outcomes evaluated include PAD-related lower limb revascularization, lower limb amputations, all-cause hospitalizations and five-year survival by one study [28], and index of multiple chronic conditions [25].

Functional outcomes were evaluated by 4 studies [19, 20, 22, 25]. 6MWD, timed up and go (TUG), short physical performance battery (SPPB) [25], as well as self-reported walking impairment questionnaire (WIQ) [20, 25] and patient-reported outcomes measurement information system physical function – mobility (PROMIS-Mobility) [25] were measured to reflect physical function. Other measurements include functional balance using Berg balance scale (BBS) [19] and lower extremity motor function using Fugl-Meyer motor function scoring [22]. In addition, sedentary time and physical function were also evaluated by accelerometer [25].

Measurements on diabetes control and metabolism have also been performed by 5 studies [17, 21, 22, 24, 25]. Blood glucose control, including fasting plasma glucose (FPG) [21, 22], 2 h postprandial blood glucose (2hPG) [22] and glycosylated hemoglobin A1c (HbA1c) [17, 21, 22, 24] were measured. Insulin functions including homeostatic model assessment of insulin function index (HOMA-islet) and HOMA for insulin resistance (HOMA-IR) [22], as well as diabetes complications index [25] were also evaluated. In addition, metabolic related indicators including body mass index (BMI) [17, 26], cholesterol, high density lipoprotein (HDL), low density lipoprotein (LDL) and triglycerides were evaluated [17]. Furthermore, peripheral neuropathy symptoms were also evaluated using Michigan neuropathy screening instrument (MNS) [27].

Other measurements include ischemic pain after six minute walking time (6MWT) using perceived pain intensity (PPI) [29], geriatric depression score [20], fear of falling using falls efficacy scale-internation (FES-I), motivation using self-efficacy for exercise (SEE) questionnaire, self-efficacy using self-efficacy for managing chronic disease scale [20, 25] and exercise behaviors survey [20]. Blood measurement including fibrinogen, hemoglobin, nitrite and thrombelastogram indices were also utilized [17, 22, 31].

Nutrition analysis

As for nutrition studies, mediterranean diet pattern [32] and adherence [33] have been evaluated. Other specific types of food have also been evaluated [30]. The 14-item Mediterranean Diet Adherence Screener (MEDAS-14) questionnaire and the 24-item food-frequency questionnaire (FFQ) were utilized [32, 33]. In addition, a 7-day food diary and 168-item food frequency questionnaire with an interview was also performed [30]. Diet component investigated include fresh fruit, vegetables, whole grains, bread, whole-wheat bread, pasta, legumes, nuts and olive oil, milk, eggs, poultry, fish, dairy products, red meat and sweets, as well as red wine, beer spirits and coffee [30, 32] (Table 1).

Outcomes

Outcomes were summarized and categorized according to the following framework: vascular, functional, metabolic and patients-reported outcomes. Specifically, improvement in perfusion and vascular functions such as ABI, TBI and arterial blood flow velocity are considered as vascular outcomes. Severity of claudication and walking ability such as 6MWD, PWT and COT were considered as functional outcomes. Parameters such as FPG, 2hPG and HbA1c were considered as metabolic outcomes, while questionnaires such as WIQ and FFQ were considered as self-reported outcomes.

Improvement of perfusion and vascular functions in diabetic PAD

Studies revealed that ABI was significantly improved after 2–3 months of treadmill walking [18, 24] or 5 months of home-based exercise [17], with improvement maintained at 1 month after completion of exercise [17] in diabetic PAD patients. Flow velocity of posterior tibial and dorsalis pedis arteries was improved (value increased) after 5 months of home-based exercise in diabetic PAD patients [17]. In contrast, the flow velocity of posterior tibial artery was reduced after once-off physical exercise, as reported by one study [23]. Buerger-Allen exercise, which targeted at improving arterial insufficiency of lower limb, may improve lower extremity perfusion after treatment as short as 4 days in a mixed cohort of diabetes patients with peripheral neuropathy in the presence or absence of PAD [27]. Limited study evaluated endothelial function. Existing study revealed that BAFMD was not improved in PAD patients with diabetes, while was improved significantly in PAD patients without diabetes [31] (Table 2).

Combination of multiple physical exercises may achieve or enhance improvement in perfusion. Specifically, combination of aerobic, strengthening and balance exercise was able to improve ABI and TBI for 3 months of exercise in diabetic PAD patients [21]. In addition, combination of treadmill walking and resistance training showed a higher improvement in ABI compared to treadmill walking alone in diabetic PAD patients [24]. Administration of shock wave therapy or LFPMT may show better improvement of ABI compared to administration of physical exercise. Administration of 2 months of shock wave therapy showed higher increase of ABI compared to administration of treadmill walking [18]. Similarly, administration of LFPMT for one month showed better improvement in ABI compared to moderate-to-high-intensity interval training (M-HIIT) [19]. In addition, administration of LFPMT for one month followed by M-HIIT for one month showed better improvement in ABI compared to administrating the two methods in the reverse sequence [19], indicating that the sequence of administration may exert different effects on improvement in ABI in diabetic PAD patients. Furthermore, administration of physical exercise in addition to glucose management showed better improvement in ABI than applying glucose management alone in these patients [22]. It may take time for these interventions to exert their maximum effects, as once-off physical exercise or high-voltage electrical stimulation (HVES) led to decreased flow velocity of femoral and popliteal arteries in PAD patients with diabetes [23] (Table 2).

Improvement in walking performance in diabetic PAD

Despite a significant improvement after 3–6 months of SET in a mixed cohort of PAD patients with and without diabetes [25, 26], 6MWD was not improved significantly after SET in diabetic PAD patients compared to controls [26]. Though improvement of PWT was reported after 3 months of SET or home-based exercise in PAD without diabetes [16, 31], the improvement of PWT in diabetic PAD patients remains controversial, and may vary between sexes. Specifically, PWT was reported to be improved after SET by one study [31], while was reported by another study to be improved after SET or home-based exercise only in male diabetic PAD patients [16]. In addition, a high non-response rate was observed after SET or home-based exercise in female diabetic PAD patients [16]. Furthermore, maximal treadmill walking time was not improved after SET in diabetic PAD patients compared to controls [26] (Table 2).

COT was improved after 3 months of SET or home-based exercise in PAD patients without diabetes [16, 31], but was not improved in diabetic PAD patients [31] or only be improved in male diabetic PAD patients after SET or home-based exercise [16]. In addition, claudication onset distance in 6MWT was improved, with value increased, after 3 months of SET (treadmill walking) in a mixed cohort of PAD patients with and without diabetes [25], but was not evaluated specifically in diabetic PAD patients alone. Walking speed was improved, while maximal treadmill walking distance and walking distance until onset of pain were not significantly improved after 6 months of home-based walking exercise in diabetic PAD patients [20]. Pain free walking time was also not significantly improved after SET in PAD patients with diabetes, which was in contrast with a significant improvement reported in PAD patients without diabetes [31] (Table 2).

Combination of different types of physical exercises including aerobic, strengthening and balance exercise showed significant improvement of 6MWD after 3 months of training in diabetic PAD patients [21]. In addition, combining treadmill walking with functional electrical stimulation for 2 months was also effective in improving 6MWD [29]. Administration of shock wave therapy for calf muscle for two months improved pain free walking distance and time as well as maximum walking distance and time more significantly than administration of treadmill walking in diabetic PAD patients [18]. In addition, combination of treadmill walking with functional electrical stimulation improved pain after 6MWT reflected by PPI, which was maintained after 3-month follow-up in PAD patients with diabetes [29]. Furthermore, combination of aerobic, strengthening and balance exercise for 3 months was also able to improve walk speed, distance and stair climbing in diabetic PAD patients [21] (Table 2).

Improvement in other functions in diabetic PAD

Though only two studies evaluated changes in functions following SET [19, 25], existing studies revealed that one month of moderate-to-high intensity interval treadmill walking was able to improve functional balance in diabetic PAD patients [19]. Three months of SET (treadmill walking) was able to improve SPPB but not sedentary behaviour variables or activities in mixed cohort of PAD patients with or without diabetes [25]. Whether similar effect would be achieved in PAD patients with diabetes remains unclear (Table 2).

Administration of M-HIIT for one month induced a more significant improvement in functional balance compared to LFPMT in diabetic PAD patients [19]. When administrating these two types of interventions sequentially, administration of M-HIIT first for one month, followed by LFPMT for one month, showed better improvement in functional balance compared to administration the methods in the reverse sequence [19]. Adding physical exercise to glucose management for 3 months was able to better improve motor function of lower limbs, reflected by higher rate of improvement in Fugl-Meyer motor function score, compared to glucose management alone in PAD patients with diabetes [22] (Table 2).

Improvement in diabetic control in diabetic PAD

FPG and 2hPG were improved, with value decreased, after 3 months of blood glucose management with physical exercise in diabetic PAD patients [22], though blood glucose management alone was also able to achieve such effects [22]. In addition, HbA1c was improved (value decreased) after 5 months of home-based exercise, and remained decreased at one month later in PAD patients with diabetes [17]. Similarly, HOMA-islet and HOMA-IR were also improved after 3 months of blood glucose management with physical exercise, though glucose management alone could also achieve such effects in PAD patients with diabetes [22] (Table 2).

Combination of multiple physical exercise may exert better effects on diabetic control in PAD patients with diabetes. Specifically, combination of aerobic, strengthening and balance exercise for 3 months was able to reduce FPG and HbA1c significantly [21]. In addition, combination of treadmill walking and resistance training for 3 months reduced HbA1c more significantly than administration of aerobic training alone [24] (Table 2). Treatment effects on neuropathy symptoms has also been evaluated by one study, showing improvement as reflected by MNSI following 4 days of BAE (27).

Other improvement in diabetic PAD

Only two studies evaluated metabolic changes and other changes in blood [17, 22]. The cholesterol level was improved after 5 months of home-based exercise, with effects on HDL maintained after 1 month in diabetic PAD patients [17]. The thrombelastogram indices was better in diabetes patients with arterial occlusive disease who finished 3 months of combination of glucose management with physical exercise than those who finished 3 months of glucose management alone [22] (Table 2).

Combination of multiple physical exercise or physical exercise with other interventions may improve quality of life. Administration of aerobic, strengthening and balance exercise for 3 months was able to improve WHOQOL-BREF in diabetic PAD patients [21]. In addition, combining treadmill walking with functional electrical stimulation for two months was also able to improve PADQOL and ICQ at 3 months follow up [29] (Table 2).

Effects of nutrition

In diabetic patients, adherence to Mediterranean diet was correlated with ABI, intermittent claudication and absence of pedal pulse [33], and MEDAS-14 score was correlated with intermittent claudication and pulse of posterior tibial artery [33]. In addition, higher FFQ score was associated with reduced risk of PAD [32]. Investigation on types of food revealed that nut consumption, white meat preference and intake of sauteed dish were associated with lower risk of PAD [33], while butter consumption was associated with increased risk of PAD [32]. In addition, diabetic patients who developed PAD had a lower intake of fish and shellfish [30] (Table 1). Due to a small number of studies reporting the effects of nutrition on diabetic patients with PAD, the evidence so far is still preliminary, requiring a further clinical validation.

Study quality

Quality assessment revealed that 100% of the nine RCTs presented some concerns, mainly in the domain of randomization process and deviations from the intended interventions, affecting certainties of evidence from these studies. Evaluation using ROBINS-I revealed that all three non-randomized interventional studies also showed some concerns. 50% of the four cohort studies presented high qualities, while 50% presented moderate qualities. The existing one cross-sectional study showed high quality, and the one case control study showed high quality in 75% of all domain-covered questions (Tables 3 and 4). The level of certainty assessed using GRADE framework in vascular functions revealed moderate to very low grade evidence on changes in ABI and flow velocity following interventions in DM +/- PAD or PAD +/- DM due to study types, low number of studies, small sample size, moderate risk of bias or conflicting results. Similarly, GRADE grading for functional changes including 6MWD, COT and PWT also revealed low to very low grade evidence due to study type, low number of studies, small sample size, moderate risk of bias or conflicting results. Analysis on effects and association of diet on risk of PAD also revealed low grade evidence due to study type and number of studies available (Supplementary Table S3 and S4).

Table 3.

Revised Cochrane risk-of-bias tool for randomized trials (RoB2) assessment for randomized clinical trials (RCTs)

Reference	Domain 1	Domain 2	Domain 3	Domain 4	Domain 5	Overall risk
Gardner, 2014 [16]	some concerns	some concerns	low risk	low risk	low risk	some concerns
Castro-Sanchez, 2013 [17]	some concerns	some concerns	low risk	low risk	low risk	some concerns
Ali, 2022 [18]	low risk	some concerns	low risk	low risk	low risk	some concerns
Abdelaal, 2024 [19]	low risk	some concerns	low risk	low risk	low risk	some concerns
Collins, 2011 [20]	some concerns	low risk	low risk	low risk	low risk	some concerns
Amin, 2024 [24]	low risk	some concerns	low risk	low risk	low risk	some concerns
Arora, 2024 [21]	low risk	some concerns	low risk	low risk	low risk	some concerns
Qiu, 2023 [22]	some concerns	some concerns	low risk	low risk	low risk	some concerns
Guirro, 2015 [23]	low risk	some concerns	low risk	low risk	low risk	some concerns

Domain 1 = risk of bias arising from the randomization process; Domain 2 = risk of bias due to deviations from the intended interventions (effect of assignment to intervention); Domain 3 = missing outcome data; Domain 4 = risk of bias in measurement of the outcome; Domain 5 = risk of bias in selection of the reported result

Table 4.

Quality assessment for non-RCTs

Reference	Domain 1	Domain 2	Domain 3	Domain 4	Domain 5	Domain 6	Domain 7	Overall
Patel, 2021 [26]	moderate risk	low risk	low risk	low risk	moderate risk	moderate risk	low risk	moderate risk
Radhika, 2020 [27]	moderate risk	moderate risk	low risk	low risk	moderate risk	moderate risk	low risk	moderate risk
Whipple, 2020 [25]	moderate risk	low risk	low risk	low risk	moderate risk	low risk	low risk	moderate risk

Domain 1 = bias due to confounding; Domain 2 = bias in selection of participants into the study; Domain 3 = bias in classification of interventions; Domain 4 = bias due to deviations from intended interventions; Domain 5 = bias due to missing data; Domain 6 = bias in measurement of outcomes; Domain 7 = bias in selection of the reported results

Discussion

DM has been identified to contribute to progression of PAD, highlighting the necessities of efficient intervention for this special cohort. Exercise has been suggested to be the best nonsurgical treatment option for pain relief and improvement of walking in patients with non-disabling intermittent claudication [35], and intensive walking has been previously shown to present better beneficial effects than usual care [36]. However, diabetic patients presented special features that may affect treatment effects, bringing uncertainties to treatment response from exercise.

DM participates in impairment of microvascular function in symptomatic PAD patients [37]. Given that exercise training was able to enhance endothelial function [38], this type of intervention may exert beneficial effects in improvement of peripheral microcirculation of lower limbs in PAD patients with diabetes. The present systematic review revealed that ABI was improved after SET or home-based exercise, and flow velocity was also improved after physical exercise or Buerger-Allen exercise in diabetic PAD patients. In contrast, limited study existed on endothelial function revealed no significant improvement. Of note, existing studies assessing ABI and vascular functions employed a variety of interventions, leading to challenges in synthesizing treatment response to a particular intervention across studies. In addition, the majority of these studies presented some concerns in study quality, affecting the confidence in these evidence. However, it could be indicated that physical exercise may improve blood flow but not necessarily endothelial function in diabetic PAD patients.

Exercise training may increase oxidative metabolism in skeletal muscle [39] that subsequently leads to improvement in muscle strength as well as endurance, and walking efficiency [40]. Analysis of studies included showed no significant improvement of 6MWD after SET in diabetic PAD patients. Though walking speed was improved, existing studies showed that maximal treadmill walking was not improved. In addition, pain free walking time and distance were not improved after exercise alone. This may be due to the fact that diabetic PAD patients with claudication reported higher severity of muscle deoxygenation [41], and DM impaired microvascular functions and oxygen kinetics during exercise [37, 42]. This suggests that lower intensity walking may be needed to reduce potential adverse reactions and therefore, enhance adherence. However, given that high intensity may be needed for exercises to achieve optimal goals, alternative methods or combination of different methods may be required. In addition, further validation is needed due to limited number of studies and moderate certainty of these findings.

Recent studies have suggested that introducing exercise other than walking exerted similar benefits as walking at moderate or higher levels of claudication pain, but did not introduce claudication during training. Some of the examples include recumbent stepping and arm ergometry or leg cranking [43–45]. The studies included showed that combination of different physical exercise was able to improve 6MWD, walking speed and distance in diabetic PAD, which was not improved when administrating SET alone. Similarly, existing study showed that combining physical exercise with FES was also able to improve 6MWD in diabetic PAD. FES has previously shown to reduce chronic pain, increase blood flow as well as cross-sectional area of muscles, and improve strength [46, 47], and therefore, potentially exerting additional benefits to symptom improvement. In addition, administration of shock wave therapy was able to improve pain free walking distance and time, as well as maximum walking distance and time more significantly compared to treadmill walking in diabetic PAD patients. Furthermore, administration of LFPMT or LFPMT combined with M-HIIT was able to improve functional balance. Together, these results indicate that alternative methods may be suitable to be administrated alone or in combination with physical exercise to improve walking and functional performance in diabetic PAD. Interestingly, administrating these methods in different sequences seems to exert different effects. Specifically, functional balance was more significantly improved when administrating M-HIIT prior to LFPMT, while ABI was improved more significantly when administrating LFPMT first, indicating that administration sequence should be considered for different treatment aims. It should be noted that only a limited number of studies were performed with moderate quality, and further validation is needed in future studies.

Several reasons may exist for the variation of response in diabetic PAD patients. The presence of DM may affect improvements in walking performance in PAD patients, and the intermittent claudication may reduce patients’ adherence to training programs. Indeed, repeated cycles of exercise-induced ischemia and reperfusion may be associated with muscular damage, which was featured by increase of inflammation and production of reactive oxygen species (ROS), changes of structures and metabolism of the muscles, and reduction in strength and function [48–50]. In addition, diabetic polyneuropathy (DPN) is associated with altered gait and introduces increased risk of fall [51]. Furthermore, women showed different responses to exercises when administrating the same training protocol compared to men, suggesting a sex difference in treatment response. Specifically, women presented more impairment of oxygen saturation in calf muscle while in ambulation [52] and shorter distances to onset of claudication [53]. Therefore, higher volume of exercise or alternative methods may be required to achieve optimized effects in female diabetic PAD patients, which requires further investigation.

Healthy diet has previously been related to reduction in risk of atherosclerosis [13]. Mediterranean diet is characterized by high consumption of foods from traditional agriculture, such as fruit, vegetables, whole grains and olive oil, and low consumption of red and processed meat and sweets [54, 55]. This type of diet plays an essential role in prevention of diabetes, reducing incidence of microvascular complications [56], and protect against atherosclerosis [57]. Less has been known on associations between dietary components and risk of PAD in diabetic patients [58]. Analysis of the limited studies included showed that adherence to Mediterranean diet was associated with improvement in blood flow and intermittent claudication [33]. In addition, consumption of appropriate food may also reduce risk of PAD [33]. These results highlight the importance of selecting and adhering to specific type of diet as an interventional component in management of diabetic PAD.

There are several limitations to the studies included. Firstly, various interventions and associated protocols were employed in the studies included, bringing heterogeneities of the study and reducing the strength and certainty of evidence. Secondly, variations were present in the study cohorts, and subgroup analysis may be desired for future studies with sufficient sample size. More specifically, some studies did not include PAD patients without diabetes, which weakens the conclusions for a direct comparison. Multi-center studies on larger cohorts with more standardized interventions will be required in the future to further validate these findings. Thirdly, though treatment response was reported in diabetic and non-diabetic PAD patients, only a limited number of studies compared response directly among groups, which potentially reduced the certainty of evidence. Fourthly, majority of studies presented moderate risk of bias, which affected certainty of evidence. More carefully designed clinical trials are needed for further validation and investigation. There are also limitations for the present systematic review. Firstly, only studies published in English were included, and therefore, research published in other languages were excluded from evaluation. Secondly, although the review was performed according to PRISMA guidelines, applying predefined eligibility criteria, comprehensive literature search and reporting, the absence of a prospective registration may reduce methodological transparency. Prospective protocol registration will be established in future systematic reviews.

Conclusions

Diabetic PAD patients may present reduced treatment response to physical exercise in walking performance and other functions when a positive response was observed in non-diabetic PAD patients, though direct comparison was performed in a limited number of studies of moderate risk with inconsistent findings. These characteristics should be taken into consideration for strategy making in patient management following further validation. In addition, though limited comparative evidence was revealed reconciling discordant results, existing studies suggested that combination of different physical exercises, utilization of other types of interventions alone or in combination with physical exercise may show better improvement in different domains, which requires further validation in more carefully designed clinical trials. Females may present different treatment response, and modifications of interventional protocols may be needed. Healthy diet, specifically Mediterranean diet, was associated with beneficial effects on diabetic PAD patients. However, the limited number of studies and inconsistency in findings requires further investigations for a more accurate component/dose correlation with clinical outcomes. The present systematic review provided valuable information to strategy optimization for management of diabetic PAD patients in clinical practice.

Supplementary information

Below is the link to the electronic supplementary material.

Abbreviations

PAD

peripheral artery disease

CLTI

chronic limb-threatening ischemia

DM

diabetes mellitus

SET

supervised exercise therapy

PRISMA

Preferred Reporting Items for Systematic Reviews and Meta-analysis

RoB2

revised Cochrane risk-of-bias tool for randomized trials

RCTs

randomized controlled trials

ROBINS-I

risk of bias in non-randomized studies of interventions

JBI

Joanna Briggs Institute

NOS

Newcastle-Ottawa Scale

ABI

ankle-brachial index

6MWD

six minute walk distance

PWT

peak walking time

COT

claudication onset time

LFPMT

low-frequency pulsed electromagnetic field therapy

FES

functional electrical stimulation

1-RM

one-rep max

BAE

Buerger-Allen exercise

BAFMD

brachial artery flow-mediated dilation

TBI

toe brachial index

PADQOL

PAD quality of life

ICQ

intermittent claudication questionnaire

MOS

medical outcomes study

SF-36

36-Item short form survey

WHOQOL-BREF

World Health Organization quality of life-BREF

TUG

timed up and go

SPPB

short physical performance battery

WIQ

walking impairment questionnaire

PROMIS

patient-reported outcomes measurement information system physical function

BBS

Berg balance scale

FPG

fasting plasma glucose

2hPG

2 h postprandial blood glucose

HbA1c

glycosylated hemoglobin A1c

BMI

body mass index

HDL

high density lipoprotein

LDL

low density lipoprotein

MNS

Michigan neuropathy screening instrument

PPI

perceived pain intensity

FES-I

falls efficacy scale-international

SEE

self-efficacy for exercise

MEDAS-14

14-item Mediterranean Diet Adherence Screener

FFQ

food-frequency questionnaire

Author contributions

QW: Writing – original draft, Data curation, Formal analysis; NQ: Writing – original draft, Data curation, Formal analysis; YL: Writing – original draft, Data curation, Formal analysis; MZ: Writing – review and editing, Conceptualization, Supervision, Methodology. All authors read and approved the final manuscript.

Funding

This research received no specific grant from any funding agency in the public, commercial or non-for-profit sectors.

Data availability

No datasets were generated or analysed during the current study.

Declarations

Competing interests

The authors declare no competing interests.