A systematic review of the effectiveness of revascularization of the ulcerated foot in patients with diabetes and peripheral arterial disease

Authors

Errata

This article is corrected by:

  1. Errata: Erratum Volume 28, Issue 4, 376, Article first published online: 29 March 2012

  • Published on behalf of the International Working Group on the Diabetic Foot.

R. J. Hinchliffe, St George's Vascular Institute, 4th Floor, St James Wing, St George's Healthcare NHS Trust, Blackshaw Road, London SW17 0QT, UK.

E-mail: rhinchli@sgul.ac.uk

Summary

In several large recent observational studies, peripheral arterial disease (PAD) was present in up to 50% of the patients with a diabetic foot ulcer and was an independent risk factor for amputation. The International Working Group on the Diabetic Foot therefore established a multidisciplinary working group to evaluate the effectiveness of revascularization of the ulcerated foot in patients with diabetes and PAD. A systematic search was performed for therapies to revascularize the ulcerated foot in patients with diabetes and PAD from 1980–June 2010. Only clinically relevant outcomes were assessed. The research conformed to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, and the Scottish Intercollegiate Guidelines Network methodological scores were assigned. A total of 49 papers were eligible for full text review. There were no randomized controlled trials, but there were three nonrandomized studies with a control group. The major outcomes following endovascular or open bypass surgery were broadly similar among the studies. Following open surgery, the 1-year limb salvage rates were a median of 85% (interquartile range of 80–90%), and following endovascular revascularization, these rates were 78% (70.5–85.5%). At 1-year follow-up, 60% or more of ulcers had healed following revascularization with either open bypass surgery or endovascular revascularization. Studies appeared to demonstrate improved rates of limb salvage associated with revascularization compared with the results of medically treated patients in the literature. There were insufficient data to recommend one method of revascularization over another. There is a real need for standardized reporting of baseline demographic data, severity of disease and outcome reporting in this group of patients. Copyright © 2012 John Wiley & Sons, Ltd.

Abbreviations
ABI

ankle-brachial pressure index

AHA

Americal Heart Association

AKA

above-knee amputation

ARF

acute renal failure

AT

anterior tibial artery

BASIL

bypass versus angioplasty in severe ischaemia of the leg

BKA

below-knee amputation

CAD

coronary artery disease

CBA

controlled before-and-after (study)

CFA

common femoral artery

CIA

common iliac artery

CKD

chronic kidney disease

CLI

critical limb ischaemia

CVD

cerebrovascular disease

DFU

diabetic foot ulcer

DM

diabetes mellitus

DP

dorsalis pedis artery

IQR

interquartile range

ITS

interrupted time series (study)

ITT

intention to treat (analysis)

IWGDF

International Working Group on the Diabetic Foot

MI

myocardial infarction

MRA

magnetic resonance angiography

NA

not available

NPWT

negative pressure wound therapy

NR

not reported

NYHA

New York Heart Association

PAD

peripheral arterial disease

PT

posterior tibial artery

PTA

percutaneous transluminal angioplasty

RCT

randomized controlled trial

SD

standard deviation

SFA

superficial femoral artery

SIGN

Scottish Intercollegiate Guidelines Network

TASC

The Inter-Society Consensus for the Management of Peripheral Arterial Disease

TBI

toe-brachial pressure index

TcPO2

transcutaneous oxygen tension

UT

University of Texas (wound classification system).

Introduction

An amputation of the leg or foot is one of the most feared complications of diabetes mellitus, and it is has been calculated that at this moment, in every 20 s, a leg is lost in our world as a result of diabetes [1, 2]. Peripheral arterial disease (PAD) and infection are the major causes of lower leg amputation in diabetes, and >80% of these amputations are preceded by a foot ulcer [3, 4]. Diabetes is a risk factor for PAD, and depending on the definitions used, prevalence rates of 10–40% in the general population of patients with diabetes have been reported [5-7]. Moreover, in comparison with subjects without diabetes, PAD is more likely to progress in patients with diabetes [8]. A substantial number of individuals with a foot ulcer will therefore have PAD, ranging from relatively mild disease with limited effect on wound healing to severe limb ischaemia with delayed wound healing and a high risk of amputation. In several large observational studies, PAD was present in up to 50% of the patients with a diabetic foot ulcer (DFU) and was an independent risk factor for amputation [9-11]. The relatively poor outcome of ischaemic foot ulcers in diabetes is probably related to a combination of factors, such as the anatomic distribution of the vascular lesions rendering them more difficult to treat, the association with other abnormalities such as infection, neuropathy and renal failure and the presence of abnormalities in other vascular territories, such as the coronary or cerebral arteries [7, 9, 12-14] The mortality of patients with PAD and a DFU is high with 50% of patients dead at 5 years [15]; the results are worse after major amputation with a 50% mortality after 2 years. In addition, wound healing can be further disturbed by a complex interplay of several other factors such as poor glycaemic control, microvascular dysfunction, impaired collateral formation, abnormal mechanical loading of the ulcer and co-morbidities. The effect of PAD on wound healing in patients with diabetes and a foot ulcer will therefore relate in part to its severity and extent and also to these other factors [16].

Clearly, early recognition of PAD, an accurate estimation of its severity and prompt institution of effective treatment in cases of severe disease would seem to be a logical approach to reduce the high number of amputations. However, as shown in the Eurodiale cohort, less than 50% of the patients with diabetes and ankle-brachial pressure index (ABI) < 0.5 underwent adequate vascular evaluation and subsequent revascularization, suggesting that there is ample room for improvement in the delivery of care [14]. PAD in patients with diabetes has a number of important characteristics that renders it more difficult to treat. The atherosclerotic lesions are multilevel and particularly severe in tibial arteries, with a high prevalence of long occlusions [17]. The predilection for multiple crural vessel involvement combined with extensive arterial calcification increases the technical challenges associated with revascularization using either open bypass or endovascular techniques.

The term critical limb ischaemia (CLI) is frequently used in the PAD literature but may not be particularly relevant for patients with diabetes. The Inter-Society Consensus for the Management of Peripheral Arterial Disease (TASC II) document suggests that the term ‘should be used for all patients with chronic ischemic rest pain, ulcers or gangrene attributable to objectively proven arterial occlusive disease’ [18]. However, ulceration or gangrene are usually the result of several interacting factors in patients with diabetes, with neuropathy playing a central role in most patients in combination with varying degrees of PAD, and therefore, confusion may exist in reporting outcomes in the literature.

Updated guidelines on the diagnosis and treatment of PAD have been provided in 2007 by TASC II, but evidence-based guidelines on the treatment of PAD in patients with diabetes and ischaemic foot ulcers are currently lacking [18]. In recent decades, new techniques and technologies have been introduced in treating PAD, which might be relevant for the patient with diabetes and a poorly healing ischaemic foot ulcer. In particular, interesting results have been reported on endovascular approaches in the leg, and the field is rapidly evolving [19, 20]. The International Working Group on the Diabetic Foot (IWDGF) therefore established a multidisciplinary working group, including specialists in vascular surgery, interventional radiology, internal medicine and epidemiology to evaluate the effectiveness of revascularization of the ulcerated foot in patients with diabetes and PAD. The aim of this multidisciplinary working group was to produce a systematic review on the efficacy of (endovascular and surgical) revascularization procedures and medical therapies in diabetic patients with a foot ulcer and PAD. With this review, supplemented with expert opinion if necessary, practical guidelines were formulated. These documents were commented upon and subsequently approved by the members of the IWDGF. In this article, we report the results of the systematic review on revascularization techniques. The consensus practical guidelines are published elsewhere in this issue.

Materials and methods

The MEDLINE and Embase databases were searched for therapies to revascularize the ulcerated foot in patients with diabetes and PAD from 1980–June 2010, (Appendix A). Searches excluded studies before 1980 because of the changing nature of interventions for PAD and improving technology.

Peripheral arterial disease was defined for the purpose of this systematic review as any flow-limiting atherosclerotic lesion of the arteries below the inguinal ligament. All patients included had to have objective evidence of PAD (e.g. angiography or magnetic resonance angiography).

It was felt that studies on patients with less severe PAD (resulting in claudication, rest pain or unspecified ‘critical ischaemia’) were not really of great importance as the aim was to determine the effects of vascular intervention in patients with PAD, diabetes and a foot ulcer. We therefore restricted the studies to those with patients with tissue loss, and studies were only included if greater than 80% of patients had evidence of tissue loss (defined as any lesion of the skin breaching the epithelium/ulceration/gangrene). The diagnosis of diabetes was made according to the individual publication. Many studies in PAD are not exclusively on patients with diabetes, and the outcome of the patient with diabetes and PAD and a foot ulcer cannot be assumed to behave in a similar way in patients without diabetes. We reported studies where >80% of patients of the population studied had diabetes in studies of more than 40 patients. If a smaller proportion of patients had diabetes, we included them only if the outcomes of that cohort were specifically reported as a separate subgroup and the results of at least 30 patients were reported. We excluded studies solely reporting interventions on aortic and iliac arterial disease, studies that had only data on health-related quality of life or costs and studies examining the diagnosis and prognosis of PAD in diabetes. Studies reporting medical therapy or local/topical therapy to improve tissue perfusion or to increase oxygen delivery were excluded (for example, prostaglandin or hyperbaric oxygen therapy) as well as studies comparing one form of revascularization technology with another (for example, various atherectomy devices).

Only studies reporting ulcer healing, limb salvage, major amputation and survival as the primary outcome measures were included in the review. Early morbidity or mortality was considered within 30 days or within the first hospital admission. A major complication was defined as any that resulted in a systemic disturbance of the patient or prolonged hospitalization (or as defined by the reporting study). Target lesion revascularization was not considered.

Patient demographics that were assessed included patient age, gender, ethnicity and co-morbidities [cardiovascular, renal and cerebrovascular disease]. The specifics of the foot lesions were reported where possible, such as site on the foot, depth, presence of infection and stratified when possible according to any previously reported/validated DFU scoring system. The anatomical distribution of PAD was extracted according to the site of the disease; standard reporting systems were included where possible (e.g. TASC/Bollinger systems [18, 21]). Objective assessment of perfusion was reported when possible, which included ABI, toe pressure and transcutaneous oxygen concentration (TcPO2). We made no distinction between various endovascular techniques (e.g. angioplasty, stenting, subintimal angioplasty and atherectomy), all being referred to as ‘endovascular therapy’. Similarly, no distinction was made between various bypass techniques (e.g. in situ versus reversed venous bypass).

The systematic search was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines [22]. Two reviewers assessed studies for inclusion on the basis of titles; two reviewers then excluded studies on the basis of review of the abstract; and then finally, a full text review was performed of selected articles. Studies were assessed for methodological robustness. To do this, the Scottish Intercollegiate Guidelines Network (SIGN) instrument was used as follows. Level 1 includes meta-analyses and randomized controlled trials (RCTs). Level 2 includes studies with case–control, cohort, controlled before and after or interrupted time series design. Studies were rated as: ++ (high quality with low risk of bias), + (well conducted with low risk of bias) and – (low quality with higher risk of bias), according to the SIGN methodological quality score [23]. Level 3 studies were studies without a control group (e.g. case series). These studies were not rated. Data were extracted into evidence tables by pairs of reviewers and then reviewed by the whole group. Pooling of data (and therefore weighting of studies) was not possible because of study heterogeneity and the generally low quality of evidence (see following discussion). When several studies reported on a specific item, we have summarized the data of these separate studies as interquartile ranges (IQRs) and median. It should be noted that these figures are not weighted means.

Results

After the identification and screening phase, 865 articles were assessed for eligibility and 49 papers were finally selected for full text review. These articles described revascularization of the ulcerated foot in 8290 patients with diabetes and PAD (Table 1). There were no RCTs, but there were three nonrandomized studies with an intervention and control group [31, 47, 57]. These were all of low quality and potentially subject to significant bias (SIGN 2−). The remaining 46 papers were case series (SIGN 3). Studies reported bypass surgery, endovascular therapy or both techniques used in combination. Although most reports adequately presented patient demographics and co-morbidities, a major limitation was that few studies adequately reported or categorized either baseline foot lesions or PAD severity. A number of studies were reported from the same institution, and it is likely that some patients were reported more than once.

Figure 1.

Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram

Co-morbidity and patient demographics

Overall, the proportion of men in each study varied from 37–96% with mean or median ages of individual studies varying from 36 to 74 years. Patient with diabetes, PAD and foot ulcers had significant co-morbidity with a relatively large proportion of patients having cardiovascular, CVD or renal disease. The prevalence of coronary artery disease was reported as 40–60% (IQRs) with a median of 50%, of CVD as 18–24% with a median of 20% and of end-stage renal disease (ESRD) as 11–48% with a median of 20% (although the definition varied from study to study and, in some studies, was only reported as renal impairment). Eight studies did not report any data on co-morbidity, and data on severity of co-morbidities was sparse (e.g. New York Heart Association classifications).

Early complications

Methods for reporting early complications were varied. Major systemic complications were frequent in both patients undergoing bypass surgery and endovascular procedures with the majority of studies reporting major systemic complications in the region of 10% with similar rates for endovascular and bypass surgery.

Perioperative mortality

Thirty-day or in-hospital mortality was described in 30 studies. The perioperative mortality following open surgery was reported in 20 studies and had an IQR of 0.8–3.7% with a median of 1.4% and was comparable in endovascular procedures: 0–4.3% with a median of 0.5%. In both open and endovascular series, there were several outlying studies with either no mortality or a mortality rate of 9% or greater. It was not clear why these results were so different. As the severity of co-morbidities frequently was not stated, it was difficult to infer the effect of co-morbidity on outcomes.

Mortality

Mortality at 1 year or longer following intervention was reported more frequently in studies describing open surgery. Morality at 1-year follow-up reported in these studies had an IQR of 11.3–21.8% with a median of 13.5% and at 5 years an IQR of 36–52.3% with a median of 46.5%. There was a paucity of long-term follow-up data in patients having undergone endovascular procedures. Three studies reported on 1-year follow-up of patients undergoing endovascular procedures with mortality rates of 10% in two studies and 29% in another; 5-year follow-up mortality rate was reported in two studies and varied widely (5 and 74%).

Limb salvage and wound healing

Limb salvage data were reported in the majority of studies; however, in almost all studies, it was not clearly defined. Following open surgery, the 1-year limb salvage rates had an IQR of 80–90% with a median of 85% in the 19 studies with 1-year data. Following endovascular revascularization, these rates were 70.5–85.5% (IQR) with a median of 78%. At 3 and 5 years following open surgery, these figures were 79.5–90% with a median of 82% and 74–78% with a median of 78%, respectively. At 3 years following endovascular procedures, the limb salvage rate was 72–78.5% (IQRs) with a median of 76% in four studies. At 5 years, limb salvage was 56 and 77% in the two studies that reported it. Wound healing was reported in seven studies [25, 30, 33, 35, 59, 65, 66]. Only one defined wound healing at a predefined time point of 12 months [59]. However, overall, the seven studies following endovascular and two following bypass surgery demonstrated an ulcer healing rate of 60% or more at 12 months follow-up.

Amputation

Major amputation was reported by 30 studies. The definition of major amputation was not always specified and sometimes differed between studies. Major amputations within 30 days were reported in three studies and varied from 2.1, 3.5 and 5%. Only two studies reported amputations at 12 months [29, 63]. The amputation rates within 24 months following open surgery had an IQR of 12.8–22.8% with a median of 17.3%, and following an endovascular procedure, these figures were 5.4–12.5% and 8.9%, respectively. The study by Malmstedt was an interpretation of the Swedish national vascular registry, Swedvasc, and therefore represents the results of a number of different vascular centres rather than those simply focused on distal bypass procedures [44]. A composite outcome of amputation and death (median follow-up 2.2 years) was given in the registry. The rate of ipsilateral amputation or death per 100 person years was 30.2 (95%CI 26.6–34.2). The median time to reach this endpoint in patients with diabetes and PAD undergoing bypass surgery (82% for ulceration) was 2.3 years.

Minor amputation rates varied widely in the 11 studies reporting on this complication [24, 28-30, 32, 39, 41, 43, 49, 65, 66]. Over the study periods, minor amputation was reported a median of 38% (IQR 23–59%). However, there was great variation with studies reporting a range of 12–91.7%. It was not clear whether patients received one or more minor amputations in any particular study. The rates of minor amputations varied between open surgery studies 26% (IQR 19–70%) and angioplasty studies 43% (IQR 38–53%); however, the number of studies was small and the demographics heterogeneous.

Infection

Only two studies specifically reported the outcomes of patients presenting with foot infection, PAD and diabetes [61, 62]. Mortality at 1 year was 5 and 19%. Limb outcomes were poorly described but limb salvage was 98% in one study at 1 year [61].

End-stage renal disease

Patients with ESRD were identified in six studies [40, 43, 47, 52, 58, 67]. The definition of ESRD varied and included patients both prior to and on dialysis and those with functioning renal transplants. Thirty-day mortality was 4.6% (IQR 2.6–8.8%), but 1-year mortality was high with 38% (IQR 25.5–41.5%) of patients perishing. One-year limb salvage rates were 70% (IQR 65–75%) in survivors. Long-term outcomes were also poor. Reported mortalities at 2 years were 48 [43] and 72% [40], at 3 years 56% [58] and at 5 years 91% [47].

Angioplasty-first strategy

Three studies with a mean follow-up of 20, 25 and 26 months reported on an angioplasty-first strategy, where angioplasty was the preferred fist-line option for revascularization (scoring of anatomical distribution was not given) [30, 39, 65]. In one of these studies, a large series of 993 consecutive patients with diabetes hospitalized with foot ulcer or ischaemic rest pain and PAD, percutaneous transluminal angioplasty (PTA) was technically not feasible in 16% of the patients because of the complete calcified occlusion of the vessel, which did not permit balloon catheter passage [30]. PTA did not establish in-line flow to the foot in only 1%. The second study was a consecutive series of 100 patients considered suitable for an infrainguinal PTA first approach, and 11% of the patients required bypass surgery for a failed PTA [39]. In the third study, a consecutive series of 534 patients were recruited from a tertiary referral hospital. Of these, data were available on 510. Angioplasty was attempted in 456 (89.4%). Angioplasty was a technical failure in 11% [65]. Mortality and limb salvage rates were comparable with the other series.

Crural vessel angioplasty

Crural PTA employed as a revascularization technique in isolation was reported in four studies [27, 32, 35, 67, 69]. Studies variously reported limb salvage outcomes, all of which exceeded 63% at 18 months (up to 93% at 35 months).

Pedal bypass grafts

Ten studies reported the results of pedal bypass grafting (one of which focused on outcomes in patients with ESRD). Studies reported limb salvage rates following pedal bypass grafting with an IQR of 85–98%, with a median of 86% at 1 year, 88.5% (81.3–82.3%) at 3 years and 78% (78–82.3%) at 5 years. However, the numbers available for follow-up at 3 and 5 years were low; the distribution/severity of PAD and the type of foot lesion were poorly reported.

Discussion

This systematic review examines the evidence to support the effectiveness of revascularization of the ulcerated foot in patients with diabetes and PAD. This is timely because the proportion of patients with diabetes and an ischaemic component to their ulcer is increasing. Recent reports suggest that up to 50% of the patients with diabetes with a foot ulcer have signs of PAD, which had a major impact on ulcer healing and the risk for lower leg amputation [3, 72, 73] . Early reports on the effectiveness of revascularization in patients with diabetes and PAD were not encouraging and led some to suggest that diabetes was ssociated with a characteristic occlusive small vessel arteriopathy, consequently leading to a nihilistic attitude toward revascularization. Subsequent laboratory studies and clinical results well summarized by the Beth Deaconess group suggested that revascularization was possible [74]. It has become increasingly recognized that patients with diabetes and an ischaemic foot ulcer represent a unique problem among patients with PAD. Consequently, the number of studies reporting a population or subgroup of patients with PAD, diabetes and ulceration is increasing, and more than 50% of the studies included in this systematic review were published after 2001. We specifically did not include studies reporting either outcomes that were not clinically relevant (e.g. target lesion revascularization) or that compared specific techniques (such as atherectomy versus transluminal versus subintimal angioplasty).

Although the quality of studies included in the review was frequently low, there were a surprising number reporting on the effectiveness of revascularization in patients with diabetes, PAD and tissue loss. The interpretation of the effectiveness of revascularization on outcomes in these studies is difficult as none of the studies included a matched control group receiving non-interventional therapy and the natural history of patients with PAD and an ulcerated foot remains poorly defined. Data on the natural history of patients with PAD, diabetes and CLI according to a standard definition (although not necessarily limited to those patients with tissue loss or ulceration) do, however, exist. In one study that reported the outcomes of patients with diabetes and CLI who were not revascularized, the limb salvage rate was 54% at 1 year [75]. This rate would appear much lower than in the series presented here where limb salvage rates in the majority of studies reported were between 78 and 85%. In a study by Marston in which patients with PAD and ulceration of the foot were treated without revascularization (70% diabetes), amputation was required in 23% at 12 months, but complete wound closure was achieved in 52% in the same time period. ABI at presentation (<0.5) predicted limb loss, and the only factor associated with healing was size of ulcer [76].

We defined periprocedural mortality and morbidity as any event occurring during a 30-day hospitalization period. Although perioperative mortality rates in this review were generally low given associated co-morbidities, perioperative major systemic complications were significant in the region of 10%. It is possible that part of these major complications was not related to the revascularization procedure per se but was more related to the poor general health status of the patients. These major systemic complications were usually poorly defined and are therefore not reported separately in this systematic review. However, our review does indicate that patients with diabetes and a foot ulcer undergoing revascularization for PAD should be optimized if possible prior to revascularization. There did not appear to be major differences in morbidity or mortality between open and endovascular techniques, although the studies are difficult to compare as discussed previously, and we cannot exclude that there were major differences in patient characteristics and severity of disease. Intermediate and long-term mortality rates during follow-up of studies in this systematic review were high. Over 10% of patients were dead at 1 year, and almost half were dead at 5 years. These results are similar to those reported in patients presenting with a foot ulcer of any origin, with 5-year mortality rates around 44% [15]. It was difficult to establish whether early aggressive and successful revascularization reduced mortality in the long term. Lepantalo also found that patients with diabetes and CLI appear to be at particularly high risk of death compared with those without diabetes [75]. However, patients in whom successful revascularization is performed appear to do better than those who undergo major amputation, half of whom are dead within 3 years [77, 78]. These findings underscore the importance of the severity and systemic nature of vascular disease in patients with diabetes. Patients with diabetes and an ischaemic foot ulcer should therefore receive aggressive and appropriate medical management of risk factors to reduce their high long-term mortality.

Ulceration of the foot in diabetes is often a complex interplay of many aetiologic factors, and the situation is compounded by the presence and severity of PAD. In any patient with diabetes and ulceration of the foot, the pathways to ulceration may differ (e.g. neuropathy and altered biomechanics) as well as the predominant factors affecting outcome (e.g. PAD, infection and co-morbidities). Although the current data indicate that revascularization should always be considered in a patient with diabetes, foot ulceration and severe ischaemia, it still remains unclear if such procedures have an added value in cases of mild to moderate perfusion deficits. There was little data to inform on the indications or timing for either diagnostic angiography or intervention among the studies, which should be one of the important topics of future studies. These studies should probably revolve around the influence of patient co-morbidity, the severity of PAD (anatomical distribution and degree of perfusion) and the characteristics of the foot wound itself.

End-stage renal disease is a strong risk factor for both foot ulceration and amputation in patients with diabetes [79]. These patients are frequently difficult to treat, and long-term mortality is high, which might negatively influence the decision to perform a revascularization procedure. However, our data indicate that even in these patients, favourable results can be obtained. The majority of studies reported 1-year limb salvage rates of 65–75% after revascularization.

Attempts have been made to categorize the distribution of PAD in patients with diabetes and correlate this with perfusion [17]. However, the severity of PAD was not well described in most studies. Results of ABI, toe pressure or TcPO2 measurements and the anatomical distribution pattern of the PAD were usually not reported, and our review indicates that all future reports on revascularization in patients with diabetes and an ischaemic ulcer should include objective measurements of the severity of PAD, including both anatomical and functional measurements, such as toe-pressure and/or TcPO2 measurements. Wound characteristics were also reported poorly, although prospective studies have shown the impact of factors such as size, depth or the presence of infection on healing and amputation rate. Clearly, a standardized wound classification system should be part of all future studies [80]. Moreover, there is a clear need for studies on the role of early revascularization in patients with diabetes, PAD and infected foot ulcers, as these patients are, in particular, at risk for a major amputation [3]. Standard reporting criteria exist for dealing with lower extremity ischaemia but are 15 years old and do not focus on factors that are specific to patients with diabetes [81]. Interestingly, outcome measures such as major amputation and wound healing were less frequently reported and, if reported, variously defined in the studies reviewed. Given the difficulty in strictly defining limb salvage, we suggest that major amputation and wound healing should be used as the major endpoints in future reports on revascularization in diabetic patients with PAD.

There are currently no RCTs directly comparing open versus endovascular revascularization techniques in patients with diabetes with an ischaemic foot ulcer. Therefore, there was insufficient data to demonstrate whether open bypass surgery or endovascular interventions were more effective in these patients. However, broadly speaking, the major outcomes appeared similar across all studies where revascularization of the foot was successful. Two meta-analyses on the outcomes of pedal bypass grafting and crural angioplasty have been performed by the same group, and the majority of patients in these two meta-analyses had diabetes [82, 83]. Although the inclusion criteria were different (many of the studies included did not specifically report on patients with diabetes or tissue loss) from our systematic review, limb salvage rates of pedal bypass grafting and crural angioplasty appeared to be equivalent to the results of our systematic review, and no major differences were reported between the two techniques. In contrast, primary and secondary patency rates were better after bypass surgery. In the two studies of consecutive patients with diabetes where angioplasty was the preferred first-line option for revascularization, favourable results were obtained and bypass surgery was only required in a minority of these patients [31, 40]. It is not possible to infer data from the BASIL trial that compared endovascular and bypass surgery in PAD because it only included 42% of patients with diabetes and no subgroup analysis was performed [84]. However, the results of both open and endovascular procedures will greatly depend upon the expertise in a given center. Clearly, further data is required to establish specifically which technique should be preferred taking patient characteristics, severity and distribution of PAD and wound characteristics into account.

Many of the studies reported herein were from well-recognized expert centres in revascularization techniques for patients with diabetes, which bias the results towards more favourable outcomes. Moreover, in some instances, there was probably significant overlap in the larger series of patients from certain centres. The data from the Swedvasc registry would suggest that it is possible to attain good outcomes when revascularization techniques are applied outside centres of expertise [44]. However, such a procedure should not be performed in isolation but should always be part of an integrated multifactorial approach that should include aggressive treatment of infection, debridement and offloading to protect the wound from repetitive biomechanical stress.

There were significant variations in the proportion of patients undergoing minor amputations. It is difficult to speculate why these rates may vary especially because few studies reported validated foot ulcer scoring systems and the indications for amputation were rarely reported. As discussed previously, healing of an ulcer with an intact foot should be one of the primary endpoints of future studies, but healing was, if reported, poorly defined in almost all studies. Where healing is only defined as intact skin, an apparently high proportion of patients with healed wounds can be reported if many patients undergo a minor amputation combined with primary wound closure. Clearly, this is also an area for future reporting standards.

Almost all studies were case series with high risk of selection and publication bias. Moreover, case series comparing bypass surgery and endovascular treatment are difficult to compare because of indication bias. Several studies included in this review were retrospective analyses containing a small number of patients. Because of heterogeneity, we could not pool the data. For ease of data presentation, we could only provide the median and (interquartile) ranges of the results of the studies we selected, but this did not correct for number of patients, severity of disease and co-morbidities. Because of these limitations, we cannot give reliable estimates of expected outcome. Clearly, there is an urgent need for properly controlled studies with a well-described population and outcomes that are relevant to patients with diabetes.

In conclusion, studies reported herein appear to demonstrate improved rates of limb salvage associated with revascularization compared with the results of medically treated patients with diabetes, PAD and ulceration previously reported in the literature. High perioperative morbidity and long-term mortality rates underline the importance of perioperative optimisation and long-term medical management of patients' diabetes and co-morbidities. Overall, there were insufficient data to recommend one method of revascularization over another. There is a real need for standardized reporting of baseline demographic data, severity of disease and outcome reporting in this group of patients. These standards should take into account both the specific characteristics of the PAD and of the wound in these patients. Further efforts are also required to standardize and improve outcome reporting, which should include wound healing, and it is important to move away from procedure-specific outcomes to disease-specific outcomes in this cohort of patients.

Conflict of interest

None declared.

Appendix A

MEDLINE and Embase search strings

MEDLINE diabetes treatments final search

Date of search: 2 June 2010

Search platform: OvidSP

File searched: Ovid MEDLINE® In-Process & Other Non-Indexed Citations and Ovid MEDLINE® 1948 to present

  1. diabet*.ti,ab.
  2. exp Diabetes Mellitus/
  3. 1 or 2
  4. (lower adj1 extremit*).ti,ab.
  5. (lower adj5 limb*).ti,ab.
  6. limb*.ti,ab.
  7. leg*.ti,ab.
  8. (foot or feet).ti,ab.
  9. toe*.ti,ab.
  10. Lower Extremity/
  11. Leg/
  12. Foot/
  13. Toes/
  14. Extremities/
  15. 4 or 5 or 6 or 7 or 8 or 9 or 10 or 11 or 12 or 13 or 14
  16. 3 and 15
  17. peripheral vascular disease*.ti,ab.
  18. peripheral arterial disease*.ti,ab.
  19. (pvd or povd).ti,ab.
  20. (pad or paod or poad).ti,ab.
  21. exp Peripheral Vascular Diseases/
  22. (claudication or claudicant*).ti,ab.
  23. exp Intermittent Claudication/
  24. exp Arterial Occlusive Diseases/
  25. exp Graft Occlusion, Vascular/
  26. exp Saphenous Vein/
  27. exp Femoral Artery/
  28. exp Popliteal Artery/
  29. 26 or 27 or 28
  30. occlus*.ti,ab.
  31. stenosis.ti,ab.
  32. 30 or 31
  33. 29 and 32
  34. 17 or 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 33
  35. 15 and 34
  36. 16 or 35
  37. perfusion.ti,ab.
  38. reperfusion.ti,ab.
  39. exp Reperfusion/
  40. (odema or edema or oedema).ti,ab.
  41. exp Edema/
  42. (swelling* or swollen).ti,ab.
  43. inflamed.ti,ab.
  44. inflammation.ti,ab.
  45. (flow or flux).ti,ab.
  46. exp Blood Flow Velocity/
  47. capillar*.ti,ab.
  48. exp Capillaries/
  49. (ischem* or ischaem*).ti,ab.
  50. exp Ischemia/
  51. (by-pass or by-pass).ti,ab.
  52. percutaneous.ti,ab.
  53. angioplast*.ti,ab.
  54. exp Angioplasty/
  55. (ballon adj1 dilation).ti,ab.
  56. (ballon adj1 dilatation).ti,ab.
  57. exp Balloon Dilatation/
  58. endotherapy.ti,ab.
  59. endovascular.ti,ab.
  60. evt.ti,ab.
  61. (revascularization or revascularisation).ti,ab.
  62. (endoscopic adj1 therapy).ti,ab.
  63. exp Endoscopy/
  64. atherectom*.ti,ab.
  65. endarterectom*.ti,ab.
  66. artherosclerosis.ti,ab.
  67. exp Atherectomy/
  68. stent*.ti,ab.
  69. exp Stents/
  70. patency.ti,ab.
  71. exp Vascular Patency/
  72. (limb adj1 salvage).ti,ab.
  73. exp Limb Salvage/
  74. subintimal.ti,ab.
  75. surg*.ti,ab.
  76. su.fs.
  77. pta.ti,ab.
  78. 37 or 38 or 39 or 40 or 41 or 42 or 43 or 44 or 45 or 46 or 47 or 48 or 49 or 50
  79. 51 or 52 or 53 or 54 or 55 or 56 or 57 or 58 or 59 or 60 or 61 or 62 or 63 or 64 or 65 or 66 or 67 or 68 or 69 or 70 or 71 or 72 or 73 or 74 or 75 or 76 or 77
  80. 36 and 78 and 79
  81. (letter or comment or editorial or case reports).pt.
  82. 80 not 81
  83. limit 82 to humans

Embase diabetes treatments final search

Date of search: 30 June 2010

Platform: OvidSP

Database file searched: Embase 1980 to present

30 June

  1. diabet*.ti,ab.
  2. exp Diabetes Mellitus/
  3. exp Diabetic Foot/
  4. 1 or 3
  5. (lower adj1 extremit*).ti,ab.
  6. (lower adj1 limb*).ti,ab.
  7. limb*.ti,ab.
  8. leg.ti,ab.
  9. (foot or feet).ti,ab.
  10. exp Lower Extremity/
  11. Leg/
  12. Foot/
  13. Toes/
  14. toe*.ti,ab.
  15. Extremities/
  16. or/5-15
  17. 4 and 16
  18. peripheral vascular disease*.ti,ab.
  19. peripheral arterial disease*.ti,ab.
  20. (pvd or povd).ti,ab.
  21. (pad or paod or poad).ti,ab.
  22. exp peripheral vascular disease/
  23. (claudication or claudicant).ti,ab.
  24. exp intermittent claudication/
  25. exp peripheral occlusive artery disease/
  26. exp graft occlusion/
  27. exp saphenous vein/
  28. exp femoral artery/
  29. exp popliteal artery/
  30. 27 or 28 or 29
  31. occlu*.ti,ab.
  32. stenosis.ti,ab.
  33. 31 or 32
  34. 30 and 33
  35. 18 or 19 or 20 or 21 or 22 or 23 or 24 or 25 or 26 or 34
  36. 16 and 35
  37. 17 or 36
  38. perfusion.ti,ab.
  39. reperfusion.ti,ab.
  40. exp reperfusion/
  41. (odema or edema or oedema).ti,ab.
  42. exp edema/
  43. (swelling* or swollen).ti,ab.
  44. inflamed.ti,ab.
  45. inflammation.ti,ab.
  46. (flow or flux).ti,ab.
  47. exp blood flow velocity/
  48. capillar*.ti,ab.
  49. exp capillaries/
  50. (ischemi* or ischaemi*).ti,ab.
  51. exp ischemia/
  52. or/38-51
  53. (by-pass or bypass or by pass).ti,ab.
  54. percutaneous.ti,ab.
  55. angioplast*.ti,ab.
  56. exp angioplasty/
  57. (ballon adj1 dilation).ti,ab.
  58. (balllon adj1 dilatation).ti,ab.
  59. exp balloon dilatation/
  60. endotherapy.ti,ab.
  61. endovascular.ti,ab.
  62. revasculari#ation.ti,ab.
  63. (endoscopic adj1 therapy).ti,ab.
  64. exp endoscopy/
  65. artherosclerosis.ti,ab.
  66. exp atherectomy/
  67. stent*.ti,ab.
  68. patency/
  69. exp vascular patency/
  70. exp stents/
  71. patency.ti,ab.
  72. (limb adj1 salvage).ti,ab.
  73. exp limb salvage/
  74. subintimal.ti,ab.
  75. surg*.ti,ab.
  76. su.fs.
  77. pta.ti,ab.
  78. or/53-77
  79. 37 and 52 and 78
  80. (Letter or Editorial).pt.
  81. 79 not 80
  82. limit 81 to human

Appendix B

Evidence tables

ReferenceStudy designPopulation (age, sex, co-morbidity, number with diabetes)PAD (distribution and severity)Foot lesionCo-morbidityIntervention and control managementOutcomeCommentOpinion
AhChong [24]Case seriesDM patients 176Distribution: not reported (NR)Tissue loss 158 (90%) DMCoronary artery disease (CAD) 48%Bypass graft to DM patientsMedian follow-up 19 monthsChinese population may differ from Western worldEarly graft failure 6%
265 consecutive infrainguinal bypasses with outcomes described diabetes versus no diabetesNo DM 89Severity:No DM 70 (79%) tissue loss (P=0.014)CVD 26%Fem-pop 44%Mortality 30 days 8% DM versus 1% no DM (P = 0.04)Limited information about patient management65 grafts failed overall during total study
Age median 74 (45–94) years versus 75 (29–94) years no DMABI 0.43Ulcer score: NRESRD NRCrural 40%Cardiovascular complications 9 vs 4% (P = NS)
gender: 50% (88) male DM, 45 (51%) no DM (P = NS)Median toe pressure 26 mmHg (0–57)Infection: NRPedal 16%Overall graft patency 1 year 63%
No scoring system usedAutogenous vein 66%Graft patency 4 years 46% DM versus 34% no DM (P = 0.19)
No DMSurvival rate at 1,3 and 5 years 80, 57, and 33%, respectively
Fem-pop 56%Ulcer healing: NR
Crural 35%Limb salvage overall at 1 year 83% for both groups and 5 years 78% DM vs 81% no DM (P = 0.79)
Pedal 9% 
(P = NS)Major amputation: NR
Autogenous vein 63%Minor amputation: NR
Complications: 8% mortality perioperative
Alexandrescu [25]Case series161 DM patientsDistribution: majority multilevel diseaseWagner classification grade 2–4 in 104 limbs (59%) or as isolated calf ulcers in 42 subjects (24%). In 30 (17%) limbs, complex below-the-knee trophic lesions were noted.CVD 40 (22%)161 procedures majority multilevel with 124 subintimal PTA (26 had single subintimal PTA)Ulcer healing: 129 (73%) before end of study, mean follow-up 22 (1) months Level of intervention not described in all patients. Approximately 50% infrapopliteal or crural
A retrospective case series of subintimal PTA and PTA in 161 patients with diabetes and ischaemic wound, PTA-first approachAge: >70 years 41%Severity: NRInfection: NRCAD 122 (69%)Limb salvage: 12, 24, 36 and 48-month limb-salvage proportions: 89, 83, 80 and 80%, respectively.70% neuropathy
Gender: NRTASC classification reportedESRD 33 (18%) dialysisIn an intention-to-treat analysis, the cumulative primary and secondary patency at 12, 24, 36 and 48 months were 62, 45, 41 and 38%, together with 80, 69, 66 and 66%, respectively.
Major amputation: 24 (13%) during follow-up
Minor amputation: 67 (38%)
30-day mortality: 1%
2-year mortality: 19 %
Major complication: 4
Bargellini [26]Prospective case series of multilevel subintimal PTA in patients deemed not fit for surgical bypassDM patients: 60Distribution: NRFontaine: 100% Fontaine stage IVCAD 42%Subintimal PTA in patients not suitable for surgical bypass:Mean follow-up 23 months (range, 0–48 months)How follow-up was performed, was not defined 
Age: 69.4 years (SD 9.4)

Gender: 41 male

Severity: NRInfection: NRCVD 25%Fem-pop level 56.7% (34)Ulcer healing: 75% (45/60)Long-term mortality low for a ‘high risk’ population medically unfit for bypass surgery
Infrapopliteal level 25% (15)Limb salvage: 93.3% (56/60)
Both levels combined 18.3% (11)Major amputation: 3 within 30 days and 4 within 16 months
Minor amputation:
Complications:
Periprocedural mortality was 5% (3/60)
Mortality at 1 and 3 years, 10 and 17%, respectively
Davidson [27]Retrospective case series54 DM patients (total population 70)Distribution: majority infrapoplitealGangrene 56%, ulcer 28% (of total population)CAD 55%, CVD 27%, haemodialysis 7% (total population)Vein graft below knee (57% to foot)Limb salvage: 90% at 12 months and 86% at 24 monthsFollow-up duration was variable, and after 1 year, 29 of 58 limbs were available for evaluation and, after 3 years, 6 of 58 limbs. 
Bypass below knee case seriesAge: 55–95 years; 38 men (total population)Severity: no informationInfection: NRMajor complications: 9/70
No score of distributionUlcer score: NREarly graft failure n = 3 (4.2%)
Patency 93% 1 year and 85% 2 years
Mortality: NR
Dosluoglu [28]Case series80 of 111 DM patientsDistribution: infrapoplitealAll tissue lossNRInfrapopliteal PTAUlcer healing: NR Strengths and weaknesses: No data on patient, leg or ulcer characteristics in DM patients provided.
A comparison of peroneal to other run-off vessels after PTAAge: NRSeverity: NRInfection: NRLimb salvage rate 75% in 24 months in diabetic patients with peroneal run-off and 76% in other run-off vesselsStudy with less than 80% with diabetes but limb salvage was reported separately for the diabetic patients in both groups
Gender: NRTASC classification providedNo other data on the diabetes sub-group divided
Dorweiler [29]Case series of pedal bypass graftsDM patients 46Distribution: crural occlusionsAll (100%) tissue lossCAD 46%Pedal bypass with vein graftFollow-up median 28 (1–70) monthsNo data on severity of PAD. 
Age: median 69 yearsSeverity: NRUlcer score: NRESRD 13%Ulcer healing: NRNo specific data on foot lesions
Gender: 36/46 (78%) maleInfection: NRLimb salvage: 30 days 98% and 87% at 2 yearsDrop out and loss to follow-up: NR
Major amputation: 4 (3 within 30 days)Well-defined study
Minor amputation: 32/46
Complications: perioperative mortality 2%
One patient failed graft within 30 days
Mortality at end of study 21/46 (47%)
Faglia [19]Case seriesAll DM patients 221Distribution:Wagner grade ulcerationCAD 55%PTA of stenoses greater than 50% diameter infrainguinalMedian follow-up 12 (5–30) months Probably significant (>80%??) of the data is a duplication of PDF 117, data reported in DE 117
Mixed series of PTAAge: NR11 patients ilio/femoral/popliteal axisI – 19%ESRD 4%Ulcer healing:221 had angioplasty but 2 had no significant stenoses, therefore 219 reported
Gender: NR81 exclusively infrapoplitealII – 25%Limb salvage:28 subjects PTA not possible (9 surgery and 19 no candidate for any revascularization)
127 femoropopliteal and infrapoplitealIII – 17%Major amputation: 10 (5%)Of the 190 patients, all ulcers healed with medical dressing – nothing more specific
Severity: TcPO2 21 (30 SD) mmHg in 180 cases.IV – 38%Minor amputation: 83 (38%)
ABI in 128 cases 0.53 (0.15)V – 1%Mortality 30 days: 0%
Mortality 5.3% at follow-up
Complications: n = 1 (transient renal failure)
Faglia [30]Retrospective case seriesDM patients 9937% iliofemoral88% tissue lossCAD 62%PTAMean follow-up 26 (15.1) monthsGood wound description at presentation, level of disease treated was well describedPossibly some patients reported elsewhere
Consecutive series of diabetic foot patients hospitalized.age: 65.5 (9.4)61% femoropopliteal /cruralTexas classificationESRD 5%68% procedures in crural arteriesUlcer healing: 862/868 wounds healedSome follow-up data were obtained by physician telephone interviewOf the 993 treated with PTA, only 10 did not manage to successfully get one vessel in-line flow to the foot
PTA as first choice revascularizationgender: 663 (67%) male32% crural0 – 12%Limb salvage: NR
Severity: TcPO2 17.0 (11.9)I – 16%
II – 19%21/993 major amputations
III – 53%Minor amputation: 478
Complications: NR
3.4%
Mortality 30-day 1/993
Primary patency at 5 years 88% (SD 9%)
Mortality at 1 year 5% (extrapolated from Kaplan–Meier curve)
Faglia [31]Cohort with follow-up 5.9 year (SD 1.28)PTA:Distribution:PTA: No lesion 62 (16%), rest Wagner score 1–4PTA: Dialysis 24 (5.7%) CAD 225 (54.8%), CVD 53 (19%)PTA, all stenoses >50% were treated (see PAD distribution)PTA:In addition, the authors analyzed their data using a case–control studyThe groups are the result of a stepwise treatment approach. Statistical analyses do not seem systematically performed, and analyses are missing. In particular, Kaplan–Meier data are incomplete: number of at risk at time points is missing. The study cannot be used as a cohort study comparing PTA and bypass. It does however give information about the results of PTA and information of the revascularized versus nonrevascularized patients
Follow-up study of 564 diabetic patients with ‘CLI’ referred for angiography, patients with obstruction more than 50% underwent PTA, when possible as first choice413 DM patientsPTA: Iliac–femoral–popliteal axis in 28 patients (6.8%)Infection: 65%Bypass: Dialysis 8 (7%); CAD 64 (59%), CVD 18 (15.8%)Iliac–femoral–popliteal axis in 28 patients (6.8%)Ulcer healing: NRBaseline characteristics of the groups are different. Therefore, confounding was induced.
Age: 69.7 years (SD 9.5)Infrapopliteal in 137 patients (32.2%)Bypass:No revascularization: Dialysis: NR, CAD 24 (88.9%), CVD 9 (33.3%)Infrapopliteal in 137 patients (32,2%)Limb salvage: NR
Gender: 146 (35.4%) femaleCombination of both in 248 patients (60%)No lesion 16 (14%), rest Wagner 1-4Combination of both in 248 patients (60%)Major amputation: 1 month 6 (2,3%); 34 (8%) at end of follow-up
Bypass group:Bypass: NRInfection 63%Bypass, femoropopliteal 58Minor amputation: NR
114 DM patientsNo revascularization: NRNo revascularization: no lesion 3 (11%), rest Wagner 1–4Fem-infrapopliteal 55Complications: NR
Age: 69.9 years (SD 9.4)Severity:Infection: 63%Other 1Bypass:
Gender: 35 (30.7%) femalePTA: TcPO2 15.3 (11.9)Ulcer healing: NR
No revascularization group:Bypass: TcPO2 10.2 (10.3)Limb salvage: NR
27 DM patientsNo revascularization: TcPO2: 7.0 (8.1)Major amputation: 1 month 3 (5,4%); 24 (21%) at end of follow-up
Age: 76.7 years (SD 10.4)Scoring: NRMinor amputation:
Gender: 13 (48,1%) femaleComplications: NR
36 (32%) primary bypass failures
No revascularization:
Ulcer healing: NR
Limb salvage: NR
Major amputation: 16 at end of follow-up
Minor amputation: NR
Complications: NR
PTA versus bypass p < 0.001
SIGN 2−
Ferraresi [32]Case series101 DM patients and 107 legsDistribution:34 ulcers, 74 gangreneCAD 28%PTA infrapoplitealUlcer healing: NRThis case series is a sub-analysis of a larger study.Strengths:
Long-term outcome of BK PTA in diabetesAge: 66 (SD 9.4) yearsInfrapoplitealRutherford classificationCVD 4%Limb salvage: 93% during mean follow-up 2.9 yearsTreated lesions clearly defined and standardized
Gender: 16 femaleSeverity:ESRD 3% (dialysis)Major amputation: 8/107 (7%) during mean follow-up 2.9 yearsPatients with marked tissue loss
TcPO2 18.1 (SD 11.2)Minor amputation: 64%Weaknesses:
Infection: NRComplications: NR1- and 3-year leg salvage and survival data are not provided, thus hindering interpretation.
Mortality 30 days: NR
Mortality during follow-up: 9%
Gargiulo [33]Prospective case series74 DM patients of 87 total populationDistribution: NR92% Fontaine stage IVCAD 53%Infrapopliteal PTAs (100%) combined with fem-pop angioplasty in 63% and in 3 (3.4%) patients combined with open revascularizationUlcer healing: 74.9% at 1 yearOnly technically successful PTA included in the analysisStrengths: well-performed prospective study with complete data set, provides relevant information on wound healing
Outcome of successful tibial PTA in ‘CLI’Age: 72 (SD 8.8) yearsSeverity: NRUlcer classification: University of TexasESRD 28%Limb salvage:

92.7% at 1 year

Gender: 44% femaleInfection: NRWeakness: the shortest follow-up data were 2 days. A Kaplan–Meier that includes duration of follow-up is missing, hampering interpretation
Major amputation: NR
Minor amputation: NR
Complications: No early perioperative complications
Gibbons [34]Retrospective case series259 DM patients (total population 318)Distribution:237/318 (74.8%) ulcer or gangreneNo informationInfrainguinal open revascularizationUlcer healing: NAWalking devices used at start of study 63% and at end of 6 months 74%Primarily health-related quality of life study
Infrainguinal bypass seriesAge: mean 66 yearsMultilevel diseaseInfection: NRFem-pop 84 (26.4%)Limb salvage: 97% at 6 months38% more active at follow-up 32.5% about the same and 29.5% worse. Less than half back to normal at 6 months (47.4%)
Gender: 62.3% maleSeverity: no informationUlcer score: NRFem-tibial/peroneal 132 (41.5%)Major amputation: NA
No score of anatomical distributionFem-pedal/plantar 100 (31.4%)Minor amputation: NA
93% primary graft patency at 6 months and secondary 97%
Complications: perioperative morbidity 21%
Hering [35]Prospective case seriesAll diabetics 44Distribution: NRWagner gradeCAD 77%Peroneal PTAMean follow-up 23 (5–45) months Really a prognostic study of Doppler waveform patterns predicting outcome of peroneal PTA
Case series of crural PTAGender: 33 maleSeverity: NRI – 0CVD 52% Ulcer healing: 26/44 Overall, 50% had a restenosis or occlusion of peroneal artery
Age: 72 (42–88) yearsII – 6 (14%)ESRD 16%Limb salvage: 81% at 6, 71% at 12 and 63% at 18 months
III – (30 (68%)Mortality 30 days: 9.1%
IV – 8 (18%)Major amputation: 5 (11%)
Infection: NRMinor amputation: NR
Complications: 1 renal failure
Hertzer [36]Case series312 DM patients of 650 (48%)Distribution: NR71% ulceration or gangreneNRInfrainguinal bypass grafts for occlusive diseaseMedian follow-up 4 years Strengths and weaknesses: very long follow-up time-limited specific data on diabetic patients
Mixed case series of bypass graftsAge: NRSeverity: NRUlcer score: NRUlcer healing: NR
Gender: 62 maleInfection: NRLimb salvage: 73% (95%CI 67–78) at 5 years, 51% (38–64) at 15 years
Major amputation: 29 amputations in 201 diabetic patients
Minor amputation: NR
Mortality 6.7% 30 days
Mortality at end of study 83% at median 4 years
Complications: not reported separately for diabetes
Hughes [37]Retrospective case seriesDM patients 82 (84%), total 98Distribution: crural93 (95%) tissue lossCAD 40%Bypass to plantar and tarsal arteries with vein graftDuration of follow-up median 9 (1–112) monthsPrimary patency 41% and secondary patency of 50% at 5 yearsNo differences in outcome between tarsal/plantar and dorsalis pedis (DP) bypass
Series of pedal bypassesAge: 68 (SD 12) yearsSeverity: NRInfection: NRESRD 4%(one prosthetic)Ulcer healing: NRConsecutive series of all revascularizations
Gender: 81/98 (83%) maleNo scoringUlcer classification: NRPopliteal inflow 72%Mortality 30 days 1/98 totalExcluded lost to follow-up cases from analysis (n = 26)
Complications: 124 perioperative complications
Mortality at 1 year 9%, 5 years 37%
Limb salvage 75% at 1 year, 69% 5 years
Secondary patency 70% at 1 year
Isaksson [38]Retrospective case seriesDM patients 43 (48 legs)Distribution: NR7 (15%) rest painPrevious myocardial infarction (MI) 11 (26%), angina 6 (14%)Pedal bypass with vein

(proximal anastomosis femoral artery 20 (42%) and popliteal artery or below 28 (58%))

Follow-up up to 1 yearShort follow-up – early results only 
Pedal bypass graft case seriesAge: 74 (40–84) yearsSeverity: ABI median 0.47 (0–2.14)All others (85%) tissue lossUlcer healing:
Gender: females 27 (63%)Score: NRUlcer score: NRLimb salvage: 1 year 85%
Infection: NRMajor amputation: NR
Minor amputation: NR
Complications at 30 days: two died (4%), one patient MI
Mortality rate at 1 year 14%
Patency at 1 year 83%
Jamsen [39]Retrospective case series100 patients (116 limbs)Distribution: NRRest pain 23 (20%), ulcer 50 (43%), gangrene 43 (37%)CAD 47%,AngioplastyIntention-to-treat analysis 1 year 67%, 3 years 63%, 5 years 56%, 8 years 45% limb salvage11 required bypass for PTA failure 
Outcome of consecutive series of 100 infrainguinal PTA considered suitable for PTA-first approach76 (76%) DM patientsSeverity: ankle systolic pressure <50 mmHgWound classification: NRCVD 28%Femoropopliteal 54%Ulcer healing: NRMedian follow-up 25 months. Validity of 5 and 10 years questionable – very small numbers available after 3 years
Age: 72 (38–90) years total populationScoring system not usedInfection: NRCrural 17%Limb salvage for endovascular treatments at 3, 5 and 8 years was 65, 60 and 60%, respectivelyPreselected to PTA-first approach
Gender: 40 (40%) male total populationMultilevel 29%Major amputation: 37 (32%) during total follow-up
Minor amputation: 14 (12%)
Major complications: 11%
The survival rates of the present study are 72, 26 and 14% at 1, 5 and 10 years.
Johnson [40]Retrospective case series43 DM patientsDistribution: NR69 limbs43 ESRD (kidney transplant 10)Total population 69 venous bypasses:Ulcer healing: NR59% ‘foot amputations’ performed with patent graftAmputation can be related not only to occlusion but also to other factors such as infection.
Retrospective review of popliteal distal bypass grafts in patients with ESRDIn total population 53Severity: in general, toe pressure < 40, ABI < 0.5 (or incompressible)(53 with tissue loss)CAD 38%Fem-pop 19Limb salvage: 1 year 65% and 62% at 18 months
Age: 59 years (total population)Scoring: NRUlcer score: NRCVD 15%Crural 50Major amputation: 22 (foot amputations)
Gender: 27 male (total population)Minor amputation: NR
Major complications: NR
Perioperative mortality 10%

1-year mortality 42%, 2-year mortality 72%

Kalra [41]Retrospective case seriesDM patients 191 (75%), total population 256; 280 proceduresDistribution: NR90% tissue loss total populationCAD 132 (52%), CVD 54 (21%), ESRD 19 (7%)All vein bypass grafts to pedal vesselsUlcer healing:Survival rate was 65% if had patent graft at 5 years versus 26% if leg amputation 
Series of pedal bypass grafting using veinAge: median 70 (30–91) years total populationSeverity: TcPO2 < 20 mmHg in 88% and ABI = 0.44 (38% incompressible) in 150 limbsInfection: NRCumulative limb salvage rates at 1, 3 and 5 years were 85% (95%CI, 80.3–89.5), 79% (95%CI, 73.9–85.1) and 78% (95%CI, 71.7–83.7), respectively.57% of patients had one or more secondary interventions for pedal graft
Gender: 174 (68%) male total populationScoring system: NRWound classification: NRUlcer healing: NR
Long grafts (proximal anastomosis above popliteal): 130 (46%) of total populationMajor amputation: 15% at 2.7 years mean follow-up
Short grafts (proximal anastomosis at or below popliteal): 150 (54%) of total populationMinor amputation: 12.4%
Complications: 1.6% perioperative mortality
Long-term mortality at 1, 3 and 5 years: 13, 24 and 40%, respectively
Secondary patency at 1 year 78%, 3 years 72% and 5 years 71%
Kandzari [42]Case series52 DM patients out of total population of 69Distribution: 154/160 lesions infrainguinal; 43% crural93% Ruth 5 and 7% Ruth 6 (total population)CAD 57%

CVD 23%

Endovascular plaque excisionUlcer healing: NR Data reported comparing patients with and without diabetes, however very little information given
Endovascular revascularization using catheter-based plaque excisionAge: 70 years (SD 12) (total population)Severity: ankle pressure < 50 mmHgNo ulcer classificationInfection: NRLimb salvage: NR
Gender: 49% maleMajor amputation: 20% with diabetes versus 18% with no diabetes (p = 0.86) at 6 months
Minor amputation: NR
Complications: major adverse events (26.7% with diabetes versus 22.2% with no diabetes, p = 0.72).
Leers [43]Retrospective case seriesDM patients 31 (91%) 34 totalDistribution: infrapopliteal in 23 legs and infrainguinal in 13 legs of the total populationProbably, >90% had tissue loss. Although, this was not explicitly stated in the article.CAD: 28 (82%)Pedal venous bypass 88% of the total populationAverage follow-up 13.5 (1–84) monthsRetrospective, some data were obtained from family or dialysis institutions 
Pedal bypass grafts in ESRDAge: 64 (39–85) years total populationSeverity (only in 16 patients): ABI 0.48 (0–0.95) mean, toe pressure 18 (0–78)Wound classification: NRESRD: 100% (29 haemodialysis and 2 transplants)Ulcer healing: NRData difficult to interpret – self-reported data
Gender: male 59% total populationInfection: NRCumulative assisted primary patency at 1 and 2 years: 62 and 62%, respectively
Limb salvage: 56% at 1 year of the total population and at 2 years 50%
Major amputation: 16 (39%) at 13.5 months average follow-up
Minor amputation: 51 (26%) of the total population at 1 year
Complications: survival 64% at 1 year

one perioperative death (2%)

Mortality 36% at 1 year and 48% at 2 years
Malmstedt [44]Part of country-wide observational data base (Swedvasc)742 DM patientsDistribution: NR82% tissue lossCAD 65%261 femoral–popliteal bypassesUlcer healing: NR Composite primary endpoint was amputation or death
Outcome after bypass surgery in diabeticsAge: 74 years (SD 9.8)Severity: NRUlcer classification: NRCVD 19%481 infrapopliteal bypassesLimb salvage: NRWith a median follow-up of 2.2 years, the rate of ipsilateral amputation or death per 100 person years was 30.2 (95%CI 26.6–34.2)
gender: 42% femaleInfection: NRESRD defined as creatinine 150 umol/L 20%Major amputation: NRMedian time to life or limb loss was 2.3 years (CI 1.9–2.8)
Minor amputation: NRThe use of the composite endpoint renders interpretation very difficult.
Complications: NR
Mills [45]Retrospective case series of patients with popliteal distal vein bypass grafts46 DM patients (total population 53)Distribution: infrapopliteal52 tissue lossESRD 28%; CAD 57%Infrapopliteal vein bypassUlcer healing: Strength: well-defined cohort
Age: 62.4 years (total population)Severity: NRInfection: NRAll crural bypassLimb salvage: 85% after 1 year (22 of 56 legs available at 1 year).Weaknesses: high withdrawal rate, probably combination of short duration and lost to follow-up (not reported separately)
Gender: 37 men (total population)Scoring distribution: NRUlcer score: NRMajor amputation: NRThis paper is an example of the confusion between the total population, number of diabetic patients, number of extremities and number of procedures.
Minor amputation: NR
Complications: Perioperative mortality: 2 of 53 (3.6%); within 30 days, two graft occlusions with subsequent two major amputations
Mortality 1 year, 13%, 2 years
Mohan [46]Case seriesAll DM patients 32Distribution: popliteal artery inflowNR

18 (51%) ulcers

15 (43%) gangrene

2 (6%) patients rest pain

CAD 47%Popliteal to distal artery bypassMean follow-up 24 (1–72) monthsSmall study population and no information regarding drop-out rate 
Pedal bypass graft case seriesMean age: 60 (range 42–84) yearsAK popliteal 9Ulcer score: NRChronic renal failure 28%Posterior tibial (PT) 9Ulcer healing: NR
BK popliteal 26
Gender: 50% maleSeverity: NRInfection: NRAnterior tibial (AT) 830-day mortality 0%
DP 10Limb salvage: 90% at 1 year and 82% at 3 years
Peroneal 8Major amputation: 5 within 20 months
All vein graftsMinor amputation: NR
Patency at 1 and 3 years: 95 and 89%, respectively
Complications: four failing graft surgeries revised. Three bypass occlusions of which two resulted in major amputation
Three additional amputation because of infection
Mortality (long term): NR
Owen [47]Cohort studyChronic kidney disease (CKD) 4 [epidermal growth factor receptor (eGRF) 15–29]:Distribution:CKD 4 (eGRF 15–29):CKD 4 (eGRF 15–29):Infrainguinal bypassCKD 4 (eGRF 15–29): A study that provides relevant data on CKD in severe forms as a prognostic factor.
According to four different levels of kidney disease25 DM patients of 32 (total cohort)Infrainguinal, no further data given84% foot lesionsCAD: 23 (71.9%) Ulcer healing: NRInfrainguinal bypass, outflow data not provided
Age: 67.5 years (SD 11.5)Severity: NRCKD 5 (eGRF < 15 and HD):CKD 5 (eGRF < 15 and HD):Limb salvage: at 5 years 77 (SD 14)This study was reported as a case series
Gender: 19 men (59%)90% foot lesionsCAD: 44 (61.1%)Major amputation: NRProbably only sufficient data on CKD 5 patients
CKD 5 (eGRF < 15 and HD):Ulcer score: NRMinor amputation: NRDifficult to use patency data because mortality is very high
60 DM patients of 72 (total cohort)Infection: NRComplications: 30-day mortality 3.1%
Age: 65 years (SD 11)CKD 5 (eGRF < 15 and HD):
Gender: 38 men (53%)Ulcer healing: NR
Limb salvage: at 5 years 50 (SD 12)
Major amputation: NR
Minor amputation: NR
Complications: 30-day mortality 4.2%
CKD 5 mortality at 1 year 46% and 91% at 5 years
SIGN 2−
Panneton [48]Retrospective case seriesDM patients 157Distribution: NR93% tissue lossCAD 80 (51%), ESRD 41 (26%)Pedal bypass graft with veinMean follow-up 2.7 yearsA subgroup of a series comparing diabetic patients and nondiabetic patients in which no differences were observed between the two groups.Comparison of diabetes and no diabetes
Pedal bypass graft seriesAge: 66 (30–78) yearsSeverity: NR53% gangreneUlcer healing: NR
Gender: 111 menScoring system: NRWound classification: NRLimb salvage: at 1 year 86% and at 5 years 78%
Infection: 27%Major amputation: NR

Minor amputation: NR

Complications: 30-day mortality 1.3%, MI 11 (7%), ARF 5 (3.2%), major amputation 3 (1.8%)
Pomposelli [49]Case series350 DM patients, total population 367Distribution: NR219 (72%) with ulcer; 47 (12%) with gangrene; 16% other indicationsPrior MI 29%, CVD 12%, ESRD 5% (dialysis)DP arterial bypassUlcer healing: NRCo-morbidity subdivided in various kinds of cardiovascular disease.Large case series, long follow-up period (5 years). Outcome is rather thoroughly described.
Retrospective review of 367 consecutive patients undergoing 384 distal bypassesAge: 58 years meanSeverity: NRInfection: 222 (55%)of the total population Limb salvage: cumulative limb salvage rate 87% at 5 years. At 1 and 2 years estimated from K-M 90 and 85%, respectively. Retrospective evaluation; not based on predefined problem; there is no drop-out rate reported. Outcome limb salvage was not defined any further.
Gender: 352 male; 114 femaleScoring: NRUlcer classification: NRSecondary patency rates 82% at 5 years
Major amputation: 13 (3.5%) within 30 days. Total number of major amputations 30 (8.1%) within the 5-year follow-up.
Minor amputation: 75 (19%)
Complications: 30-day mortality 1.8% MI 5.4%. Graft failures 7.5% at 30 days
Mortality 43% after 5 years
Pomposelli [50]Retrospective case series865 total populationDistribution: inflow vessel78% ulcerCAD 47%; ESRD 11% of the total populationSubgroup analysis of 1032 DP artery bypassUlcer healing: NR43 (4.2%) failed within 30 daysSubgroup analysis of a large 3731 bypasses to 1032 to DP arteries of which some were diabetic patients (865)
Pedal bypass graft series92% diabetes41% BK poplitealInfection: NRAll except two with veinLimb salvage: 78% at 5 years and 58% at 10 years
Age: 67 years29% CFAUlcer score: NRGraft patency 85% at 1 year
Gender: 69% male12% AK popSecondary patency at 5 years 66% DM versus 56% no DM
11% superficial femoral artery (SFA)51 and 76% mortality at 5 and 10 years, respectively
Severity: NRMajor amputation: NR
No scoringMinor amputation: NR
Complications: 10 (1%) deaths within 30 days; 3% MI
Pua [51]Case series91% DM patients out of 46 totalDistribution: NR37/46 patients with foot lesions33% CADMixed 25Ulcer healing: at 13 months 66% patients with gangrene healedFive technical failuresLimited information regarding patient characteristics, co-morbidity and selection procedures.
Consecutive patients receiving PTA for limb salvageAge: NRSeverity: NRUlcer score: NR20% CVDFive cruralLimb salvage: 78% at 1 year.Foot ulcer/gangrene is not specified any further.
Gender: NRNo scoreInfection: NR16 fem-popMortality: NR
Three aortoiliacMajor amputation: NR
Minor amputation: NR
Ramdey [52]Prospective case series (registry)DM patients; 92% of a total population of 146Distribution: NRTissue loss: 91% (total population)CAD 115 (65%)Artery inflowFollow-up: no data providedFollow-up not specified 
Infrainguinal revascularizationAge: 63 years (SD 13) (total population)Severity: NRUlcer score: NRMI 64 (36%)Iliac or femoral 123 (70%); suprageniculate popliteal 20 (11%)Complications:

30-day morbidity 23%; 30-day mortality 5%

Gender: 65% (total population)Infection: 48%CVD 27 (15%)Infrageniculate popliteal 34 (19%)Ulcer healing: NR
ESRD: all patientsOutflowPatency at 1 and 3 years: 85 and 68%, respectively
Iliac/femoral 1 (0.6%)Limb salvage: 1 and 3 years 80% both
Suprageniculate popliteal 17 (10%)Major amputation: 21
Infrageniculate poplitealMinor amputation: NR
28 (16%)Survival 60% 1 year, 3 years 18% and only 5% alive at 5 years
Tibial 50 (28%)
DP 80 (45%)
Tarsal 1 (0.6%)
Reed [53]Retrospective case seriesDM patients 140, total population 217, procedures 249Distribution: NRNecrosis 127 (80%), rest pain 27 (17%)CAD 95 (60%), ESRD 53 (33%) with 35 (23%) on dialysisInfrainguinal vein bypass graft30-day mortality 0.6%Data extracted from a cohort study comparing diabetic with nondiabetic patientsMajor amputation was required in nine patients with a patent graft.
Case series of bypass grafts originating distal to the groinAge: 65 (30–90) yearsSeverity: NRInfection: NRPedal (35%),Complications: major post-operative morbidity 16 (10%)21% secondary procedures
Gender: 79 femaleScoring system: NRWound classification: NRCrural (60%), femoropopliteal (4%)Ulcer healing: NRMean 27 months follow-up
Limb salvage rate was 84% (SD ± 4) at 5 years
Major amputation:
Minor amputation:
5-year patient survival was 44 (±5)%
Rosenbaum [54]Retrospective case seriesDM patients: 39Distribution: NR100% tissue lossNRPeripheral bypass: 79% infrapoplitealFollow-up: Mean 21.2 (2–64) monthsData of this study may be included in other reports of this groupNo life-table analysis, no information about healing time, small series; follow-up procedures unclear
Case series of infrapopliteal bypass graftsAge: 62.3 (45–78) yearsSeverity: NRUlcer score: Gibbons classificationPopliteal 19%Ulcer healing: 40 limbs (of total 42 limbs) with or without foot surgery
Gender: 33 (85%) maleScore: NRInfection: NRTibial/peroneal: 48%Limb salvage:
DP/plantar artery: 31%Major amputation: 1
Aotobifemoral 2%Major amputation: 1
Complications:
Saltzberg [55]Retrospective case seriesDM patients: 96% of total population in 51 patients all younger than 40 yearsDistribution: 76 bypass procedures with inflow:86% tissue lossCAD 37%Venous (95%) or prosthetic (5%) bypass with outflow:How follow-up was performed not described; no data on follow-up reportedUnspecified follow-up 
Case series of mixed bypass graftsAge: 36 (27–40 years)Common iliac 2.6%Ulcer score: NRESRD (creatinine > 2 mg/dL, dialysis or transplant) 53% (29% of which dialysis)DP: 30.3%NR
Gender: 49% maleFemoral 67%Infection: NRTibial artery: 18.4%Patency at 1 and 5 years: 82 and 63%, respectively
All data in this table as reported on total populationAbove-knee popliteal 7,9%Peroneal artery 3.9%Ulcer healing: NR
Below knee popliteal 21.1 %Below knee popliteal: 23,7%Limb salvage: 87% at 1 year and 77% at 5 years
Tibial artery 1.3%Above-knee popliteal: 11.8%Major amputation: 23.5% required amputation level unspecified (12/51)
Severity: NRFemoral artery: 3.9%

Other: 7.9%

Minor amputation: see previous entry
Complications: 30-day mortality rate, 0%; post-operative heart failure, 1.32%.
Overall mortality 88 and 73% at 1 and 5 years, respectively
Schneider 1993 [56]Case series of pedal bypass extracted from an initial cohort study of diabetic versus nondiabetic patients, but this study compares tibial with pedalDM patients 45 of total population n = 53Distribution: NR77% tissue lossCVD NRAll pedal bypass graft with veinFollow-up: 22.5 months (SD 3.4)Major amputation defined as amputation proximal to metatarsals 
Age: 67 (42–78) years total populationSeverity: ABI 0.53 total populationUlcer score: NRCAD: NRUlcer healing: NRLow numbers of patients (6) at 5 years
Gender: 33 men of total populationInfection: NRESRD: NRPatency at 1, 3 and 5 years: 70, 58 and 58% years
Limb salvage at 1, 3 and 5 years: 98, 98 and 95%, respectively
Major amputation: NR
Minor amputation: NR
Complications: perioperative mortality 9%
Mortality 27, 39 and 50% at 1, 3 and 5 years, respectively
Schneider 2001 [57]Retrospective cohortSFA PTA plus short distal bypassDistribution: Combined: Below knee disease plus focal SFA disease (<3 cm length)All gangreneCombinedDistal target vesselsMean follow-up 23 monthsSmall sample 
Revascularization using either fem-distal bypass, combined SFA PTA and distal bypass grafting or short distal bypass graftDM patients 12Severity ABI 0.52 (0.19)Infection: NRCAD 33%, ESRD 58%CombinedUlcer healing: NRHeterogeneous populations – different distribution of PAD
Age: 70 (13) yearsLong distal bypass: Extensive infrainguinal disease involving fem-pop and infrageniculate arteriesWound classification: NRLong distalTibial 25%Limb salvage at 2 yearsConfounding by indication
Gender: 83% maleSeverity ABI 0.42 (0.17)CAD 38%Pedal 75%Combined 90 (9)%Drop-out and loss to follow-up not reported
Long distal bypassShort distal bypass: Severe infrageniculate occlusive disease and patent fem-pop arteriesESRD 74%Long distalLong distal 78 (9)%
DM patients 46Severity: ABI 0.46 (SD 0.15)Short distalTibial 57%Short distal 98 (2)%
Age: 68 (11) yearsScoring system: NRCAD 49%Pedal 43%Patency all procedures 78 (±5)% at 2 years, 63 (8)% at 5 years
Gender: 50% maleESRD 67%Short distalMajor amputation: NR
Short distal bypassTibial 35%Minor amputation: NR
DM patients 52Pedal 65%Complications: NR
Age: 69 (11) yearsMortality: NR
Gender: 65% maleNo differences between groups
SIGN 2−
Sigala [58]Case seriesAll diabetic patients 97 with 121 proceduresDistribution: Large variation49 necrosis 32 gangrene, 24 ulcers, 16 rest pain,CAD 78%Infrainguinal revascularizationsUlcer healing: NRHeterogeneous population of patients with wide variation of PAD distribution and revascularization proceduresSigala [58]
Mixed bypass graft plus 50 PTA66% maleSeverity: NRUlcer score: NRCVD 20%Endovascular – 36% onlyLimb salvage: 86% at 6 months, 75% at 12 months and 56% at 3 yearsAll patients had ESRDPDF 845
Mean age 68 (range 41–85) yearsInfection: NR100% ESRD5% combination endovascular and openMajor amputation: NRNumber of infections not stated in study but outcomes reported in Kaplan–Meier relative to infection
Bypass only 59%Minor amputation: NR
Crural artery 55%Complications 12/97 patients
10% crural artery onlyMortality 30 days 10%, 1 year 22% and 3 years 56%
28% femoropopliteal
18% external iliac to femoropopliteal
Soderstrom [59]Prospective case series74 DM patients of 148 total populationDistribution: NRClassification provided: All Fontaine stage IV ulcersNRInfrainguinal bypass in all subjects, with 13 PTA inflow procedures (total population)Ulcer healing: 63% in 12 months in the diabetic patientsMedian time to achieve healing 213 daysArterial run-off for patients with diabetes not specified.
Healing of ischaemic ulcers after infrainguinal bypass surgeryAge: NRSeverity: ABI < 0.5, systolic toe pressure < 30 mmHg. Fontaine stage IVInfection: NRLimb salvage: NRDiabetes was the only risk factor that delayed tissue healing (HR 0.5 95%CI 0.3–0.8 in multivariate analysis)No specific data on diabetic patients reported other than healing.
Gender: NRClassification: NRMajor amputation: NR
Minor amputation: NR
Complications: NR
Mortality: NR
Stonebridge [60]Case seriesAll DM patients (117)Distribution: tibialNonhealing 65 (52%), gangrene 20 (16%)CAD 37/117Pop-distal bypass graft (129 procedures)Ulcer healing: NR No data about inclusion criteria according to PAD severity.
Retrospective review of 117 diabetic patients with a popliteal artery (or below) to distal bypassAge: 64 (27–92) yearsSeverity: NRInfection: 40 (32%)ESRD 17/117Limb salvage: NR
Gender: Male : female 5 : 1Scoring: NRFoot abscess: 2 (1.6%)Major amputation: 8 during mean follow-up 13 months: minor amputation: 34
Osteomyelitis: 6 (5%)Complications: operative mortality 0.8 %,
Ulcer score: NR1 and 3 years secondary patency rates 92 and 89%, respectively
Tannenbaum [61]Retrospective case seriesDM patients 53Distribution: NR73% ulcersNRDP bypass with veinFollow-up: average 25 (SD 14) monthsStudy of acute sepsis in ischaemic diabetic feet 
Case series of pedal bypassAge: NRSeverity: NRInfection: 45% cellulitis, 29% osteomyelitis, 20% gangrene and 2% abscess10 patients lost to follow-upExcellent limb survival and patient survival and healing
Gender: 34 maleScore: NR11 minor amputations performed prebypassLimb salvage at 1, 2 and 3 years, 98, 98 and 95%, respectivelyNo report on severity of PAD
Major amputation: NRWound infection 13%
Minor amputation: NR
Patency at 1–3 years, all 95%
Complications: NR
Mortality rate at 1, 2 and 3 years, 5, 16 and 16%, respectively
In the patients who did not die or had a major amputation, all ulcers healed during follow-up
Taylor [62]Retrospective case seriesDM patients 114 patients with a foot infection (138 limbs):Distribution: NRAll infected ulcersNRPeripheral bypass undefinedMean follow-up 3 (1–11) years17 lost to follow-up 
Case series of 114 patients with infection, 43 of whom revascularized43 (48 limbs) with ischaemia and 71 without ischaemiaSeverity: NRUlcer score: NRUlcer healing: NRNo data on lost to follow-up on revascularized patients
Age: NRBut ischemia was defined as absent pulses +ABI < 0.6 or TBI < 0.4 and abnormal wave formsLimb salvage: 2 and 4 years, 87 and 73%, respectivelyMuch important data missing
Gender: NRScore: NRMajor amputation: 9 (19%) at 3 years mean follow-up4/9 amputations because of infection
Minor amputation: NR
Complications: NR
Mortality rate at 1, 3 and 5 years 19, 62 and 84%, respectively
Toursarkissian [63]Primarily a prognostic study of the use of duplex as a predictor of bypass graft failure in diabetic patientsDM patients 65Distribution: NR61 (94%) tissue lossCAD 38%,68 limbsDuration of follow-up 12 months (SD 6 months) 86% Hispanic population
Age: 61 yearsSeverity: TBI 0.2Infection: NRESRD 16%Femoral to distal bypass 42Ulcer healing: NR
Gender: 40/64 (63%) maleScore: NRUlcer score: NRBK pop distal 16Limb salvage: whole group 80% at 1 year
Fem-pop 10Major amputation: 8/68 limbs at 12 months (SD 6 months)
All vein bypassMinor amputation: NR
Graft patency assisted primary 75% at 1 year (estimate of Kaplan–Meier)
Complications: nil
Mortality: NR
Toursarkissian [64]Case series of pedal bypass135 patients 144 procedures all DM patientsDistribution: NR96% tissue lossCAD 62%DP bypass graftsMedian follow-up 8 (1–62) months82% Hispanic 
Age: 62 (SD 11) yearsSeverity: NRUlcer score: NRESRD 20% Ulcer healing: NRStudy comparing outcome in various ethnic groups (Hispanic versus non-Hispanic). Higher amputation rate in Hispanic
Gender: 78% maleInfection: NR Limb salvage: 83% at 30 months
Major amputation: 19% at total follow-up (mean 8 months)
Minor amputation: 36%
Patency: 70% at 1 year and 68% at 30 months
Complications: 25 perioperative complications  
30-day mortality 1.5%
Mortality at end of study 10%
Uccioli [65]Retrospective case series of patients with diabetes and CLI and tissue loss treated using an endovascular-first approach in a tertiary care clinicDM patients: 510 (100%) (total population 534 but 24 lost)Distribution: NR100% tissue lossCAD 42%456/510 (89%) underwent attempted PTAMean follow-up 20 (13) monthsGood description of cohort 
Age: 70 (0.8) yearsSeverity: TcPO2 16 mmHg ±0.8Ulcer score:CVD 23%34, 35 and, 31% respectively for AK, BK, AK + BK PTAUlcer healing: 61% at 9.4 (0.5) months and 7% at 23 monthsOutcomes better reported using Kaplan–Meier analysis.
Gender: 64% maleScore: NRAll class C/D and grades 2 and 3 of the Texas wound classificationESRD 13%1.8 (0.04) vessels treated per limb (total number of arterial stenoses 2.6 (0.06) per limb)Major amputation 16% at 5 monthsOf the 89.4% of consecutive patients who were able to be treated using a PTA-first approach, 11% had technical failure.
Infection: 79%Minor amputation: 54%23% of PTA subintimal
Complications: NR
30-day mortality: NR
Mortality: 16% at 9 months
Verhelst [66]Retrospective case seriesDM patients: 33 (92% of total population n = 36)Distribution: NR89% tissue lossCAD 44%Popliteal-to-distal venous bypass grafts (n = 44):Meal follow-up 27 months (1–65)Confusion between patients/extremities. Small study. Mixture of vascular interventions. 
Case study of pedal and crural bypass graftAge: 62 (29–78) yearsSeverity: TcPO2 18 mmHg ±7Ulcer score: NRDialysis 4PT: 13Ulcer healing:Started treating 33 patients – No standard error in curve and therefore high likelihood of significantly small numbers during follow-up
Gender: 29 maleScore: NRInfection: NRAT: 10In 33/36 patients, complete healing of skin lesions and that includes minor amputations
All data in this table as reported on total populationPeroneal: 6Limb salvage: 90, 82 and 77% at 1, 3 and 5 years, respectively.
DP/plantar: 23Major amputation: occurred in follow-up period of 27 months
Minor amputation: 35
Patency at 1 and 3 years 87 and 74%, respectively
Complications: MI 1; heart failure 1; post-operative bypass occlusion and major amputation 3
30-day mortality 0%
Deaths: 4 during following follow-up
Werneck [67]Case series40 DM patients (total population 49)Distribution: All had ‘severe’ tibial disease, ‘some also had femoropopliteal PAD’Classification: 20% Ruth 4CAD 69%Tibial angioplasty in all and in 45% multilevel (fem-pop segment)Ulcer healing: NRAngiographic success rate was 84%.Number of patients with surgery versus PTA not given.
Tibial PTA in patients with ‘CLI’ at ‘high risk’ retrospective case seriesAge: 70 yearsSeverity: NR80% Ruth 5*ESRD 73% of the total populationLimb salvage: 76% after mean follow-up 8 months. Cumulative limb salvage rate in tibial PTA only after 1 year: approximately 70% estimated from Kaplan–MeierThere are ten amputations in diabetic patients. However, it is unclear how many cases were in the PTA group.
Gender: 71% maleTASC reported:Infection: NRMajor amputation: NR
Minor amputation: NR
Complications: major complications occurred in 6.1%
30-day mortality 2%
Mortality after 1 year 10%
Woelfle [68]Retrospective case seriesDM patients: 72Distribution: Isolated tibioperoneal vessel; occlusive diseaseAll with minor tissue lossCAD 41Distal vein graft reconstruction:Follow-up: no information provided or data reported on how this was performedUlcer healing not reported; total number of BK amputations not reported. 
Case study of mixed bypass graftsAge: 70.5 yearsSeverity: NRUlcer score: NRESRD (creatinine > 2 mg/dL): 18Proximal anastomosis:Ulcer healing: NRNo data on follow-up
Gender: NRInfection: NRSymptomatic carotid disease: 15Below-knee popliteal: 56Limb salvage at 30 days 93%, at 1 year 81% and 5 years 72%.
AT: 18Patency: 30 days 97%, 1 year 86% and 5 years 75%
Distal anastomosis:Major amputation: NR
ATA 10Minor amputation: NR
DPA 37Complications:
PTA 13Mortality within 30 days 1.3%
Peroneal: 1223 patients died during follow-up (including post-operative mortality)
Plantar 3 
Wolfle [69]Retrospective case series of two different proceduresBypassBypassBypassBypassVein to DP in 63 or AT artery in 20 and PT in 28 and peroneal in 19Average follow-up probably 24 monthsPoor information on loss to follow-up and drop out. 
Bypass crural versus PTA cruralDM patients 125 (130 grafts)Distribution: crural127 tissue lossCAD 57%AngioplastyBypassRetrospective case series of two different procedures and not a controlled study
Age: 70 (50–87) yearsSeverity: NRPTACVD 18%Crural arteriesLimb salvage 80% at 1 year, 73% at 3 years and 69% at 6 years
Gender: NRPTA84 tissue lossESRD 25%AHA classification (1994)2.3% 30-day mortality
Distal PTADistribution: cruralUlcer score: NRPTA1–8Patency at 1, 3 and 5 years, 76, 70 and 60%, respectively
DM patients 74 (89 limbs), 84 totalSeverity: TcPO2 6.7 (0–29)Infection: NRCAD 48%2–2830 major amputation at 24 months
Age: 68 (48–89) yearsScore: AHACVD 17%3–26Minor amputations: NR
Gender: NRESRD 42%4–2764 died during follow-up
PTA
Limb salvage 82% at 1 year, 77% at 3 years and 77% at 5 years
30-day mortality 6%
17 major amputations during 24 months
Minor amputation: NR
26 died during follow-up
Woefle [70]Retrospective case seriesDM patients: 135 (143 procedures)Distribution: extensive intrapopliteal occlusionsTissue loss in 140 limbsCAD 82 (61%)All venous bypass with proximal anastomosis:Follow-up duration not reportedNo data on mean duration of follow-up or on severity of PAD 
Case studies infrapopliteal bypass graftAge: 70 (50–89) yearsSeverity: NRUlcer score: NRESRD 43 (16%)BK popliteal 113Ulcer healing: NR
Gender: NRInfection: NRCVD 29 (20%)ATA 29Patency at 1 year 83%, 5 years 60% and 7 years 51%
PA 1Limb salvage: at 30 days 94%, 1 year 80%, 5 years 74% and 7 years 64%
Distal anastomosis:Major amputation: 35 during follow-up
ATA 21Minor amputation: NR
DPA 71Complications:
PTA 2930-day mortality 8%
Peroneal 22Mortality at 1 year 27%, 5 years 70% and 7 years 82%
PTA of SFA prior to surgery in 37
Zayed [71]Retrospective seriesDM patients: 312Distribution: NR93% tissue lossCAD 107 (34%)257 (82%) PTA,Follow-up not defined and no data reportedNo data on follow-up 
Series of combined PTA and bypass surgeryAge: 188 males (40%)Severity: NRUlcer score: NRDialysis: 33 (10.5%)55 (18%) surgical bypass open surgeryUlcer healing: NRPTA not specified, severity of PAD not described
Gender: 188 males (60%)Classification: NRInfection: NR20 had combination of both proceduresLimb salvage: NRAll amputations above or through knee
Major amputation: 13 cases (4.1%); of these, seven had PTA and six had reconstructive vascular surgery
Minor amputation: NR
Complications: NR

Ancillary