Reviews/Commentaries/Position Statements CONSENSUS STATEMENT

Peripheral Arterial Disease in People With Diabetes

AMERICAN DIABETES ASSOCIATION lower-extremity amputation, especially in patients with diabetes. Moreover, even for the asymptomatic patient, PAD is a marker for systemic in- eripheral arterial disease (PAD) is a 1) WHAT IS THE volving coronary, cerebral, and renal ves- condition characterized by athero- EPIDEMIOLOGY AND sels, leading to an elevated risk of events, P sclerotic occlusive disease of the IMPACT OF PERIPHERAL such as myocardial (MI), lower extremities. While PAD is a major ARTERIAL DISEASE IN , and . risk factor for lower-extremity amputa- PEOPLE WITH DIABETES? Diabetes and smoking are the stron- tion, it is also accompanied by a high PAD is a manifestation of gest risk factors for PAD. Other well- known risk factors are advanced age, likelihood for symptomatic cardiovas- characterized by atherosclerotic occlusive , and hyperlipidemia (3). cular and cerebrovascular disease. Al- disease of the lower extremities and is a Potential risk factors for PAD include though much is known regarding PAD in marker for atherothrombotic disease in other vascular beds. PAD affects ϳ12 mil- elevated levels of C-reactive protein the general population, the assessment (CRP), fibrinogen, , apoli- and management of PAD in those with lion people in the U.S.; it is uncertain how many of those have diabetes. Data from poprotein B, (a), and plasma diabetes is less clear and poses some viscosity. An inverse relationship has special issues. At present, there are no the Framingham Heart Study (1) revealed that 20% of symptomatic patients with been suggested between PAD and alcohol established guidelines regarding the consumption. care of patients with both diabetes and PAD had diabetes, but this probably greatly underestimates the prevalence, In people with diabetes, the risk of PAD. given that many more people with PAD PAD is increased by age, duration of dia- On the 7–8 of May 2003, a Con- are asymptomatic rather than symptom- betes, and presence of peripheral neurop- sensus Development Conference was atic. As well, it has been reported that of athy. African Americans and Hispanics held to review the current knowledge those with PAD, over one-half are asymp- with diabetes have a higher prevalence of regarding PAD in diabetes. After a series tomatic or have atypical symptoms, about PAD than non-Hispanic whites, even after of lectures by experts in the field of one-third have , and the re- adjustment for other known risk factors endocrinology, , vascular mainder have more severe forms of the and the excess prevalence of diabetes. It is surgery, orthopedic surgery, podia- disease (2). important to note that diabetes is most try, and nursing, a vascular medicine The most common symptom of PAD strongly associated with femoral- panel was asked to answer the following is intermittent claudication, defined as popliteal and tibial (below the knee) PAD, questions: pain, cramping, or aching in the calves, whereas other risk factors (e.g., smoking 1) What is the epidemiology and im- thighs, or buttocks that appears repro- and hypertension) are associated with pact of PAD in people with diabetes? ducibly with walking exercise and is more proximal disease in the aorto-ilio- 2) Is the biology of PAD different in relieved by rest. More extreme presenta- femoral vessels. people with diabetes? tions of PAD include rest pain, tissue loss, The true prevalence of PAD in people 3) How is PAD in diabetes best diag- or gangrene; these limb-threatening man- with diabetes has been difficult to deter- nosed and evaluated? ifestations of PAD are collectively termed mine, as most patients are asymptomatic, 4) What are the appropriate treat- critical limb ischemia (CLI). many do not report their symptoms, ments for PAD in people with diabetes? PAD is also a major risk factor for screening modalities have not been uni- formly agreed upon, and pain perception ●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●●● may be blunted by the presence of periph- eral neuropathy. For these reasons, a pa- From the American Diabetes Association, Alexandria, Virginia. tient with diabetes and PAD may be more Address correspondence to Nathaniel Clark, MD, MS, RD, American Diabetes Association, 1701 N. Beauregard St., Alexandria, VA 22311. E-mail: [email protected]. likely to present with an ischemic ulcer or Received and accepted for publication 8 September 2003. gangrene than a patient without diabetes. This consensus statement has been reviewed and endorsed by the Vascular Disease Foundation. While amputation has been used by some Abbreviations: ABI, ankle-brachial index; CABG, coronary bypass graft; CAPRIE, Clopidogrel as a measure for PAD prevalence, medical versus Aspirin in Patients At Risk of Ischemic Events; CLI, critical limb ischemia; CRP, C-reactive protein; eNOS, endothelial cell nitric oxide synthase; FDA, Food and Drug Administration; FFA, free fatty acid; MI, care and local indications for amputation ; MRA, magnetic resonance angiogram; NF-␬B, nuclear factor-␬B; PAD, peripheral versus revascularization of the patient arterial disease; PAI-1, plasminogen activator inhibitor-1; PI, phosphatidylinositol; PKC, protein kinase C; with critical limb ischemia widely vary. PVR, pulse volume recording; RAGE, receptor for advanced glycation end products; UKPDS, U.K. Prospec- The nationwide age-adjusted amputation tive Diabetes Study; VSMC, cell. rate in diabetes is ϳ8/1,000 patient years A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion ϳ factors for many substances. with a prevalence of 3%. However, re- © 2003 by the American Diabetes Association. gional patterns differ—there is nearly a

DIABETES CARE, VOLUME 26, NUMBER 12, DECEMBER 2003 3333 Peripheral arterial disease in people with diabetes ninefold variation of major amputations curring in ϳ4%. While the majority of have a slower walking speed (generally in people with diabetes across the U.S. patients remain stable in their lower-limb Ͻ2 mph) and a limited walking distance. Therefore, the incidence and prevalence symptomatology, there is a striking excess This may result in a “cycle of disability” of amputation may be an imprecise mea- cardiovascular event rate over the same with progressive deconditioning and loss sure of PAD. 5-year time period, with 20% sustaining of function. Finally, there are significant The reported prevalence of PAD is nonfatal events (MI and stroke) and a economic costs of health care, reduced also affected by the methods by which the 30% mortality rate (7). For those with productivity, and personal expenses asso- diagnosis is sought. Two commonly used CLI, the outcomes are worse: 30% will ciated with a chronic manifestation of ath- tests are the absence of peripheral pulses have amputations and 20% will die erosclerotic disease such as PAD. and the presence of claudication. Both, within 6 months (8). The natural history however, suffer from insensitivity. A more of PAD in patients with diabetes has not 2) IS THE BIOLOGY OF PAD accurate estimation of the prevalence of specifically been studied longitudinally, DIFFERENT IN PEOPLE PAD in diabetes should rely upon a vali- but it is known from prospective clinical WITH DIABETES? dated and reproducible test. Such a test is trials of risk interventions that the cardio- Diabetes affects nearly every vascular bed; the ankle-brachial index (ABI), which in- vascular event rates in patients with PAD however, the pervasive influence of dia- volves measuring the systolic blood pres- and diabetes are higher than those of their betes on the atherothrombotic milieu of sures in the ankles (dorsalis pedis and nondiabetic counterparts. the peripheral vasculature is unique. The posterior tibial ) and arms (bra- abnormal metabolic state accompanying chial artery) using a hand-held Doppler Diagnosis of PAD diabetes results in changes in the state of and then calculating a ratio. Simple to Diagnosing PAD is of clinical importance arterial structure and function. The onset perform, it is a noninvasive, quantitative for two reasons. The first is to identify a of these changes may even predate the measurement of the patency of the lower patient who has a high risk of subsequent clinical diagnosis of diabetes. Relatively extremity arterial system. Compared with MI or stroke regardless of whether symp- little is known about the biology of PAD an assessment of pulses or a medical his- toms of PAD are present. The second is to in individuals with diabetes in particular. tory, the ABI has been found to be more elicit and treat symptoms of PAD, which However, it is felt that the atherogenic accurate. It has been validated against an- may be associated with functional disabil- changes observed with other manifesta- giographically confirmed disease and ity and limb loss. PAD is often more subtle tions of atherosclerotic disease, such as found to be 95% sensitive and almost in its presentation in patients with diabe- coronary and carotid artery disease, are 100% specific (4). There are some limita- tes than in those without diabetes. In generally applicable to patients with both tions, however, in using the ABI. Calci- contrast to the focal and proximal athero- PAD and diabetes. fied, poorly compressible vessels in the sclerotic lesions of PAD found typically in The proatherogenic changes associ- elderly and some patients with diabetes other high-risk patients, in diabetic pa- ated with diabetes include increases in may artificially elevate values. The ABI tients the lesions are more likely to be vascular inflammation and derangements may also be falsely negative in symptom- more diffuse and distal. Importantly, PAD in the cellular components of the vascu- atic patients with moderate aortoiliac ste- in individuals with diabetes is usually ac- lature, as well as alterations in blood cells noses. These issues complicate the companied by peripheral neuropathy and hemostatic factors. These changes are evaluation of an individual patient but are with impaired sensory feedback. Thus, a associated with an increased risk for ac- not prevalent enough to detract from the classic history of claudication may be less celerated atherogenesis as well as poor usefulness of the ABI as an effective test to common. However, a patient may elicit outcomes. Given the large size of the pe- screen for and to diagnose PAD in patients more subtle symptoms, such as leg fatigue ripheral vascular bed, the potential im- with diabetes. Using the ABI, one recent and slow walking velocity, and simply at- pact of these abnormalities is great. survey (5) found a prevalence of PAD in tribute it to getting older. It has been people with diabetes Ͼ40 years of age to reported that patients with PAD and dia- Diabetes, inflammation, and risk for be 20%, a prevalence greater than antici- betes experience worse lower-extremity PAD pated using less reliable measures, such as function than those with PAD alone (9). Inflammation has been established as symptoms or absent pulses. Moreover, Also, diabetic patients who have been both a risk marker and perhaps a risk fac- another survey of patients with diabetes identified with PAD are more prone to the tor for atherothrombotic disease states, Ͼ50 years of age showed a prevalence of sudden ischemia of arterial including PAD (11). Elevated levels of PAD of 29% (6). Thus, the prevalence of (10) or may have a pivotal event leading CRP are strongly associated with the de- PAD in diabetes appears higher than pre- to neuroischemic ulceration or infection velopment of PAD (12). In addition, lev- viously estimated. that rapidly results in an acute presenta- els of CRP are abnormally elevated in tion with critical limb ischemia and risk of patients with impaired glucose regulation Impact of PAD amputation. By identifying a patient with syndromes, including impaired glucose The impact of PAD can be assessed by its subclinical disease and instituting preventa- tolerance and diabetes. progression, the presence of symptoms, tive measures, it may be possible to avoid In addition to being a marker of dis- and the excess cardiovascular events asso- acute, limb-threatening ischemia. ease presence, elevation of CRP may also ciated with systemic atherosclerosis. Ap- PAD in diabetes also adversely affects be a culprit in the causation or exacerba- proximately 27% of patients with PAD quality of life, contributing to long-term tion of PAD. CRP has been found to bind demonstrate progression of symptoms disability and functional impairment that to endothelial cell receptors promoting over a 5-year period, with limb loss oc- is often severe. Patients with claudication apoptosis and has been shown to colocal-

3334 DIABETES CARE, VOLUME 26, NUMBER 12, DECEMBER 2003 American Diabetes Association ize with oxidized LDL in atherosclerotic (PKC), inhibition of phosphatidylinositol thrombosis via their role in adhesion and plaques. CRP also stimulates endothelial (PI)-3 kinase (an eNOS agonist pathway), aggregation. production of procoagulant tissue factor, and production of reactive oxygen species. leukocyte adhesion molecules, and che- The sum effect of all these leads to the loss of Diabetes, , and rheology motactic substances and inhibits endo- NO homeostasis. Diabetes leads to a hypercoagulable state thelial cell nitric oxide (NO) synthase The effects of endothelial cell dys- (17). It is associated with the increased (eNOS), resulting in abnormalities in the function, along with activation of the re- production of tissue factor by endothelial regulation of vascular tone. Finally, CRP ceptor for advanced glycation end cells and VSMCs, as well as increased may increase the local production of com- products (RAGE), increase the local in- plasma concentrations of factor VII. Hy- pounds impairing fibrinolysis, such as flammatory state of the vascular wall, me- perglycemia is also associated with a de- plasminogen activator inhibitor (PAI)-1. diated in part by increased production of creased concentration of antithrombin the transcription factors, nuclear fac- and protein C, impaired fibrinolytic func- Diabetes and endothelial cell tor-␬B (NF-␬B), and activator protein 1. tion, and excess production of PAI-1. dysfunction Local increases in these proinflammatory Finally, abnormalities in rheology are The endothelial cell lining of the arterial factors, together with the loss of normal seen in diabetic patients as an elevation in vasculature is a biologically active organ. NO function is associated with increased blood viscosity and fibrinogen. Elevated It modulates the relationship between the leukocyte chemotaxis, adhesion, transmi- viscosity and fibrinogen are both correla- cellular elements of the blood and the vas- gration, and transformation into foam tive with abnormalities in ABI among pa- cular wall, mediating the normal balance cells. This latter process is further aug- tients with PAD, and elevated fibrinogen between thrombosis and fibrinolysis, and mented by increased local (or its degradation products) has been as- plays an integral role in leukocyte/cell (15). transformation is the ear- sociated with the development, presence, wall interactions. Abnormalities of endo- liest precursor of atheroma formation. and complications of PAD. thelial function can render the arterial In summary, diabetes increases the system susceptible to atherosclerosis and risk for atherogenesis via deleterious ef- Diabetes and the VSMC its associated adverse outcomes. fects on the vessel wall, as well as effects The presence of diabetes is also associated Most patients with diabetes, including on blood cells and rheology. The vascular with significant abnormalities in VSMC those with PAD, demonstrate abnormalities abnormalities leading to atherosclerosis function. Diabetes stimulates pro- of endothelial function and vascular regula- in patients with diabetes may be evident atherogenic activity in VSMC via mecha- tion (13). The mediators of endothelial cell before the diagnosis of diabetes, and they nisms similar to that in endothelial cells, dysfunction in diabetes are numerous, but increase with duration of diabetes and including reductions in PI-3 kinase, as well an important final common pathway is de- worsening blood glucose control. Further as local increases in oxidative stress and up- rangement of NO bioavailability. NO is a studies of the diabetes-specific mecha- regulation of PKC, RAGE, and NF-␬B. The potent stimulus for vasodilatation and lim- nisms responsible for the development of sum total of these changes might be ex- its inflammation via its modulation of leu- atherosclerosis, as well as the specific pected to promote the formation of athero- kocyte-vascular wall interaction. pathways responsible for PAD in this pop- sclerotic lesions. These effects may also Furthermore, NO inhibits vascular smooth ulation, are needed. increase VSMC apoptosis and tissue factor muscle cell (VSMC) migration and prolifer- production, while reducing de novo syn- ation and limits activation. There- 3) HOW IS PAD IN thesis of plaque stabilizing compounds, fore, the loss of normal NO homeostasis can DIABETES BEST such as . Thus, the above events ac- result in a cascade of events in the vascula- DIAGNOSED AND celerate atherosclerosis and are also associ- ture leading to atherosclerosis and its con- EVALUATED? ated with plaque destabilization and sequent complications. precipitation of clinical events (16). Several mechanisms contribute to the Clinical evaluation: history and loss of NO homeostasis, including hyper- physical glycemia, insulin resistance, and free fatty Diabetes and the platelet The initial assessment of PAD in patients acid (FFA) production. Hyperglycemia play an integral role in the con- with diabetes should begin with a thor- blocks the function of endothelial eNOS nection between vascular function and ough medical history and physical exam- and boosts the production of reactive ox- thrombosis. Abnormalities in platelet bi- ination to help identify those patients ygen species, which impairs the vasodila- ology may not only promote the progres- with PAD risk factors, symptoms of clau- tor homeostasis fostered by . sion of atherosclerosis, but also influence dication, rest pain, and/or functional im- This oxidative stress is amplified because, the consequence of plaque disruption and pairment. Alternative causes of leg pain in endothelial cells, glucose transport is atherothrombosis. As in the endothelial on exercise are many, including spinal not downregulated by hyperglycemia. cell, platelet uptake of glucose is un- , and should be excluded. PAD In addition to hyperglycemia, insulin checked in the setting of hyperglycemia patients present along a spectrum of se- resistance plays a role in the loss of normal and results in increased oxidative stress. verity ranging from no symptoms, inter- NO homeostasis (14). One consequence of Consequently, platelet aggregation is mittent claudication, rest pain, and finally insulin resistance is excess liberation of enhanced in patients with diabetes. Plate- to nonhealing wounds and gangrene. FFAs. FFAs may have numerous deleteri- lets in diabetic patients also have in- A thorough walking history will elicit ous effects on normal vascular homeostasis, creased expression of glycoprotein Ib and classic claudication symptoms and varia- including activation of protein kinase C IIb/IIIa receptors, which are important in tions thereof. As these symptoms are of-

DIABETES CARE, VOLUME 26, NUMBER 12, DECEMBER 2003 3335 Peripheral arterial disease in people with diabetes ten not reported, patients should be asked An ABI value Ͼ1.3 suggests poorly com- toe pressure is also useful in the evalua- specifically about them. Two important pressible arteries at the ankle level due to tion of the patient with medial arterial cal- components of the physical examination the presence of medial arterial calcifica- cification, where the ABI is less accurate. are visual inspection of the foot and pal- tion. This renders the diagnosis of PAD by Another method of predicting healing is pation of peripheral pulses. Dependent ABI alone less reliable. the measurement of the transcutaneous rubor, pallor on elevation, absence of hair Due to the high estimated prevalence of partial pressure of oxygen (TcPO2). A growth, dystrophic toenails, and cool, PAD in patients with diabetes, a screening value Ͻ30 mmHg is associated with poor dry, fissured skin are signs of vascular in- ABI should be performed in patients Ͼ50 healing of wounds or amputations. sufficiency and should be noted. The in- years of age who have diabetes. If normal, terdigital spaces should be inspected for the test should be repeated every 5 years. A Anatomic studies: duplex fissures, ulcerations, and infections (18). screening ABI should be considered in dia- sonography, magnetic resonance Palpation of peripheral pulses should betic patients Ͻ50 years of age who have angiogram, and contrast be a routine component of the physical other PAD risk factors (e.g., smoking, hy- exam and should include assessment of pertension, hyperlipidemia, or duration of For those patients in whom revasculariza- the femoral, popliteal, and pedal vessels. diabetes Ͼ10 years). A diagnostic ABI tion is considered and anatomical local- It should be noted that pulse assessment should be performed in any patient with ization of stenoses or occlusions is is a learned skill and has a high degree of symptoms of PAD. It should be noted that important, an evaluation with a duplex interobserver variability, with high false- in the evaluation of the individual patient ultrasound or a magnetic resonance an- positive and false-negative rates. The dor- there may be errors and that the reliability of giogram (MRA) may be valuable. Duplex salis pedis pulse is reported to be absent any diagnostic test is dependent on the prior ultrasound can directly visualize vessels in 8.1% of healthy individuals, and the probability of disease (Bayes’ Theorem). and is also useful in the surveillance of posterior tibial pulse is absent in 2.0%. postprocedure patients for graft or Nevertheless, the absence of both pedal Vascular lab evaluation: segmental patency. MRA is noninvasive with mini- pulses, when assessed by a person expe- pressures and pulse volume mal risk of renal insult. It may give images rienced in this technique, strongly sug- recordings that are comparable with conventional X- gests the presence of vascular disease. In the patient with a confirmed diagnosis of ray angiography, especially in occult PAD in whom an assessment of the location pedal vessels, and may be used for ana- and severity is desired, the next step would tomical diagnosis. Noninvasive evaluation for PAD: ABI be a vascular laboratory evaluation for seg- While MRA is a safe and promising new In contrast to the variability of pulse assess- mental pressures and pulse volume record- technology, the gold standard for vascular ment and the often nonspecific nature of ings (PVRs). These tests should also be imaging is X-ray angiography, and it is in- information obtained via history and other considered for patients with poorly com- dicated primarily for the anatomical evalu- components of the physical exam, the ABI is pressible vessels or those with a normal ABI ation of the patient in whom a a reproducible and reasonably accurate, where there is high suspicion of PAD. Seg- revascularization procedure is intended. noninvasive measurement for the detection mental pressures and PVRs are determined Because it is an invasive test with a small risk of PAD and the determination of disease se- at the toe, ankle, calf, low thigh, and high of contrast-induced nephrotoxicity, “ex- verity (19). The ABI is defined, as noted pre- thigh. Segmental pressures help with lesion ploratory” angiography should not be per- viously, as the ratio of the systolic blood localization, while PVRs provide segmental formed for diagnosing PAD. For patients pressure in the ankle divided by the systolic waveform analysis, a qualitative assessment with suspected pedal ischemia, the angiog- at the arm. The tools re- of blood flow. raphy should include an aortogram with se- quired to perform the ABI measurement in- lective unilateral runoff and a magnified clude a hand-held 5–10 MHz Doppler Treadmill functional testing lateral view of the foot. It should be noted probe and a blood pressure cuff. For patients with atypical symptoms or a that the decision to perform an angiogram is The ABI is measured by placing the pa- normal ABI with typical symptoms of clau- made on a clinical basis and the need for tient in a supine position for 5 min. Systolic dication, functional testing with a graded revascularization, sometimes independent blood pressure is measured in both arms, treadmill may help with diagnosis. Patients of any prior noninvasive tests. and the higher value is used as the denom- with claudication will typically exhibit a inator of the ABI. Systolic blood pressure is Ͼ20-mmHg drop in ankle pressure after 4) WHAT ARE THE then measured in the dorsalis pedis and exercise. Treadmill testing may also be used APPROPRIATE MEDICAL posterior tibial arteries by placing the cuff as an evaluation of treatment efficacy and as TREATMENTS FOR PAD IN just above the ankle. The higher value is the an assessment of physical function. PEOPLE WITH DIABETES? numerator of the ABI in each limb. The diagnostic criteria for PAD based Additional evaluation Treatment of systemic on the ABI are interpreted as follows: In patients with possible CLI, further atherosclerosis associated with PAD noninvasive studies may help with clini- Most cardiovascular risk factors for indi- ● Normal if 0.91–1.30 cal decision making regarding revascular- viduals with PAD are similar to those for ● Mild obstruction if 0.70–0.90 ization. A toe pressure Ͻ40 mmHg or a people with diabetes alone. Although ● Moderate obstruction if 0.40–0.69 toe waveform Ͻ4 mm may predict im- there is little prospective data showing ● Severe obstruction if Ͻ0.40 paired wound healing and is often used in that treating these risk factors will im- ● Poorly compressible if Ͼ1.30 the evaluation of ischemic ulcers. Systolic prove cardiovascular outcomes in people

3336 DIABETES CARE, VOLUME 26, NUMBER 12, DECEMBER 2003 American Diabetes Association with both PAD and diabetes specifically, showed that ramipril, an ACE inhibitor, the efficacy of prolonged treatment with consensus strongly supports such inter- significantly reduced the rate of cardio- antiplatelet agents (in most cases, aspirin) ventions, given that both PAD and diabe- vascular death, MI, and stroke in a broad (29). tes are associated with significantly range of high-risk patients without hyper- This meta-analysis combined data increased risks of cardiovascular events. tension (24). Of the 9,297 patients in this from Ͼ100,000 patients, including Cigarette smoking. Cigarette smoking study, 4,051 had PAD. Patients with PAD ϳ70,000 high-risk patients with evi- is the single most important modifiable had a similar reduction in the cardiovas- dence of . A 27% risk factor for the development and exac- cular endpoints when compared with reduction in odds ratio (OR) in the com- erbation of PAD. In patients with PAD, those without PAD, thus demonstrating posite primary endpoint (MI, stroke, and tobacco use is associated with increased that ramipril was effective in lowering the vascular death) was found for high-risk progression of atherosclerosis as well as risk of fatal and nonfatal ischemic events patients compared with control subjects. increased risk of amputation (20). Thus, among all patients. Nonetheless, the po- However, when a subset of Ͼ3,000 pa- tobacco cessation counseling and avoid- tential benefit of ACE inhibitors has not tients with claudication was analyzed, ef- ance of all tobacco products is absolutely been studied in prospective, randomized fects of antiplatelet therapy were not essential. trials in patients with PAD. Such trials are significant. Thus, the use of aspirin to pre- Glycemic control. Hyperglycemia may needed before making definite treatment vent cardiovascular events and death in be a cardiovascular risk factor in individ- recommendations regarding the use of an patients with PAD is considered equivo- uals with PAD; however, evidence for the ACE inhibitor as a unique pharmacologic cal; however, aspirin therapy for people benefit of tight glycemic control in ame- agent in the treatment of PAD. with diabetes is recommended (30). liorating PAD is lacking. In the U.K. Pro- Dyslipidemia. Although treating dyslip- The Clopidogrel Versus Aspirin in Pa- spective Diabetes Study (UKPDS), idemia decreases cardiovascular morbid- tients At Risk of Ischemic Events (CAP- intensive glycemic control reduced diabe- ity and mortality in general, no studies RIE) Study evaluated aspirin versus tes-related endpoints and diabetes- have directly studied the treatment of clopidogrel in Ͼ19,000 patients with re- related (21). However, it was not disorders in patients with PAD. In a cent stroke, MI, or stable PAD (31). The associated with a significant reduction in meta-analysis of randomized trials in pa- study results showed that 75 mg of clopi- the risk of amputation due to PAD. In fact, tients with PAD and dyslipidemia who dogrel per day was associated with a rel- the major reduction in adverse end points were treated by a variety of therapies, ative risk reduction of 8.7% compared was due to improved microvascular Leng et al. (25) reported a nonsignificant with the benefits of 325 mg of aspirin per rather than macrovascular end points. An reduction in mortality and no change in day for a composite endpoint (MI, isch- additional caveat is that, although it is nonfatal cardiovascular events. However, emic stroke, and vascular death). More likely that many patients with PAD were the severity of claudication was reduced striking, in a subgroup analysis of Ͼ6,000 included in the UKPDS study, the preva- by lipid-lowering treatment. Similarly, in patients with PAD, clopidogrel was asso- lence of PAD was not defined, therefore a subgroup analysis of the Scandinavian ciated with a risk reduction of 24% com- conclusions from this study may not di- Simvastatin Survival Study (4S), the re- pared with aspirin. Clopidogrel was rectly relate to patients with diabetes and duction in level by simvastatin shown to be as well tolerated as aspirin. PAD. Nevertheless, good glycemic con- was associated with a 38% reduction in Based on these results, clopidogrel was trol (A1C Ͻ7.0%) should be a goal of the risk of new or worsening symptoms of approved by the Food and Drug Admin- therapy in all patients with PAD and dia- intermittent claudication (26,27). In the istration (FDA) for the reduction of isch- betes in order to prevent microvascular Heart Protection Study, adults with coro- emic events in all patients with PAD. In complications. nary disease, other occlusive arterial dis- the CAPRIE study, about one-third of the Hypertension. Hypertension is associ- ease, or diabetes were randomly allocated patients in the PAD group had diabetes. ated with the development of atheroscle- to receive simvastatin or placebo (28). A In those patients, clopidogrel was also su- rosis as well as with a two- to threefold significant reduction in coronary death perior to aspirin therapy. increased risk of claudication (22). In the rate was observed in people with PAD, In summary, patients with diabetes UKPDS, diabetes endpoints and risks of but the reduction was no greater than the should be on an antiplatelet agent (e.g., were significantly reduced and effect of the drug on other subgroups. aspirin or clopidogrel) according to cur- risk of MI was nonsignificantly reduced Thus, although there are no data showing rent guidelines (30). Those with diabetes by tight blood pressure control (23). Risk direct benefits of treating dyslipidemia in and PAD may benefit more by taking for amputation due to PAD was not re- individuals with both PAD and diabetes, clopidogrel. duced. In general, the effects of treating dyslipidemia in diabetic patients should hypertension on atherosclerotic disease be treated according to published guide- Treatment of symptomatic PAD or on cardiovascular events have not been lines, which recommend a target LDL Medical therapy for intermittent claudica- directly evaluated in patients with both cholesterol level Ͻ100 mg/dl. Following tion currently suggests exercise rehabili- PAD and diabetes. Nevertheless, consen- this guideline, it is our belief that lipid- tation as the cornerstone therapy, as well sus still strongly supports aggressive lowering treatment may not only decrease as the potential use of pharmacologic blood pressure control (Ͻ130/80 mmHg) cardiovascular deaths, but may also slow agents. in patients with PAD and diabetes in or- the progression of PAD in diabetes. Exercise rehabilitation. Since 1966, der to reduce cardiovascular risk. Antiplatelet therapy. The Antiplatelet many randomized controlled trials have Results of the Heart Outcomes Pre- Trialists’ Collaboration reviewed 145 ran- demonstrated the benefit of supervised vention Evaluation (HOPE) study domized studies in an effort to evaluate exercise training in individuals with PAD

DIABETES CARE, VOLUME 26, NUMBER 12, DECEMBER 2003 3337 Peripheral arterial disease in people with diabetes

(32,33). These programs call for at least 3 are more likely to present with advanced Treatment of infection months of intermittent treadmill walking disease compared with nondiabetic pa- Although ulcers often become infected, three times per week. Exercise therapy tients (36). the of foot infection has minimal associated morbidity and is The “neuroischemic” foot—with PAD are diminished in diabetic patients. The likely to improve the cardiovascular risk and neuropathy—is more prone to trau- early warning signs of infection may be factor profile. Of note, however, in nearly matic ulceration, infection, and gangrene. subtle because of an impaired neuroin- all studies, unsupervised exercise regi- Each requires specific man- flammatory response. Furthermore, it mens have shown lack of efficacy in im- agement as well as treatment of the under- may be difficult to differentiate between proving functional capacity. lying ischemia. the erythema of cellulitis and the rubor of Pharmacologic therapies. Pentoxifyl- In contrast to the plantar location of ischemia. The redness of ischemia, which line, a hemorheologic agent, was ap- neuropathic ulcers, ischemic ulcers are is most marked on dependency, will dis- proved by the FDA in 1984 for treating commonly seen around the edges of the appear upon elevation of the limb, claudication. The results of postapproval foot, including the apices of the toes and whereas that of cellulitis will remain irre- spective of foot position. Infections in the trials, however, suggest that it does not the back of the heel. They are generally diabetic foot are often polymicrobial; increase walking distance to a clinically associated with a pivotal event: trauma or broad spectrum antibiotics are initially in- meaningful extent. wearing unsuitable shoes. Important as- Cilostazol, an oral phosphodiesterase dicated. Severe infections require intrave- pects of conservative management in- nous antibiotic therapy and urgent type III inhibitor, was the second drug to clude debridement, offloading the ulcer, gain FDA approval for treating intermit- assessment of the need for surgical drain- appropriate dressings, and adjunctive age and debridement. tent claudication. Significant benefit has wound healing techniques (37). been demonstrated in increasing maximal Both wet and dry gangrene can occur Prompt and timely referral of the pa- in the neuroischemic foot. Wet gangrene walking time in six of eight randomized tient to appropriate foot care and vascular controlled trials, in addition to improving is caused by a septic , secondary to specialists is critical. soft-tissue infection or ulceration. Gas in functional status and health-related qual- Debridement. Debridement should re- ity of life (34). The use of this drug is the soft tissues is a serious finding requir- move all debris and necrotic material to ing an immediate trip to the operating contraindicated if any degree of heart fail- render infection less likely. The preferred room for open drainage of all infected ure is present due to concerns about ar- method is frequent sharp debridement spaces and intravenous broad-spectrum rhythmias. In a single trial, pentoxifylline with a scalpel, normally undertaken at the antibiotics. It is important to emphasize was inferior when compared with treat- hospital bedside or in the outpatient set- that medical treatment of infection with ment with cilostazol (35). Based on the ting. Indications for surgical debridement antibiotics alone is insufficient to resolve above, cilostazol is the drug of choice if include the presence of necrotic tissue, the majority of diabetic foot infections. pharmacologic therapy is necessary for localized fluctuance, and drainage of pus Incision and drainage is the basic te- the management of PAD in patients with or crepitus with gas in the soft tissues on net of treatment for nearly all infections of diabetes. X-ray. the diabetic foot. Sometimes amputation Preventative foot care. All patients with Footwear. With the neuroischemic foot, of a toe, toes, or ray(s) may be necessary to diabetes and PAD should receive preven- establish drainage. Salvage of the diabetic tive foot care with regular supervision to the chief aim is to protect the foot from pressure and shear. Ulcers may be pre- foot is usually possible but may require minimize the risks of developing foot aggressive debridement and revascular- complications and limb loss (18). vented from healing if the patient wears tight shoes or slip-on styles. It is most im- ization. Postoperatively there may be con- portant that the shoe does no harm. A siderable tissue deficit or exposure of bone or tendon. In such circumstances Treatment of the ischemic foot shoe that is sufficiently long, broad and deep, and fastens with a lace or strap high the foot should be revascularized as indi- CLI manifested by rest pain, ulceration, or cated and soft tissue deficits may be re- on the foot may be all that is needed to gangrene in the foot of a person with dia- paired by reconstructive surgery at a latter protect the margins of the foot and allow betes portends limb loss and requires ur- stage. A vacuum-assisted wound closure gent treatment. The frequent presence of healing of the ulcers. It may be necessary, device provides topical subatmospheric neuropathy strongly influences the clini- however, to provide special footwear, pressure that is most helpful in staged cal presentation. The presence of neurop- such as sandals or braces. procedures. athy blunts pain perception, allowing a Dressings. Nonadherent dressings Dry gangrene is secondary to a severe later presentation with more severe le- should cover diabetic foot ulcers at all reduction in arterial perfusion and occurs sions than in the nondiabetic patient. In a times. No single ideal dressing exists, and in chronic critical ischemia. Revascular- vicious cycle, the presence of PAD in- there is no evidence that any one dressing ization should be initially carried out fol- creases nerve ischemia, resulting in wors- is better for the diabetic foot than any lowed by surgical debridement. If ened neuropathy. In addition, such other. However, the following properties revascularization is not possible, surgical arterial lesions may progress undetected are desirable: ease of removal from the debridement or amputation should be for long intervals due to the distal distri- foot and ability to accommodate pres- considered if the necrotic toe or any other bution, making the severity of the under- sures of walking without disintegrating. area of necrosis is painful or if the circu- lying PAD often underestimated. Occlusive dressings may lower the risk of lation is not severely impaired. Otherwise Accordingly, diabetic patients with PAD infection. the necrosis should be allowed to autoam-

3338 DIABETES CARE, VOLUME 26, NUMBER 12, DECEMBER 2003 American Diabetes Association putate as a surgical procedure may result Stenoses of the superficial femoral ar- sions that may progress to sudden throm- in further necrosis and a higher level of tery may be treated with an endovascular boses if uncorrected. These lesions are amputation. approach, but restenosis is common. best detected by duplex ultrasonography. More durable results appear obtainable In addition, ϳ50% of patients with CLI in Indications for revascularization with open bypass to the popliteal artery, one limb will develop threatened limb The indications for limb revascularization particularly using saphenous . loss in the contralateral limb, underscor- are disabling claudication or CLI (rest Whether newer endovascular techniques, ing the need for ongoing risk factor reduc- pain or tissue loss) refractive to conserva- such as to prevent restenosis, will tion and close monitoring of lower-limb tive therapy. Disabling claudication is a affect the longer-term outcome of endo- circulation. relative, not absolute indication, and re- vascular management of superficial Major amputation in the neuroisch- quires significant patient consultation. femoral artery occlusions remains emic foot is necessary and indicated only One must weigh existing symptoms speculative. when there is overwhelming infection against the risk of the procedure and its Bypass to the tibial or pedal vessels that threatens the patient’s life, when rest expected effect and durability. Although with autogenous vein has a long track pain cannot be controlled, or when exten- most ischemic limbs can be revascular- record in limb salvage and remains the sive necrosis secondary to a major arterial ized, some cannot. Lack of a target vessel, most predictable method of improving occlusion has destroyed the foot. Using unavailability of autogenous vein, or irre- blood flow to the threatened limb. The these criteria, the number of major limb versible gangrene beyond the midfoot procedure is safe, durable, and effective. amputations should be limited. may preclude revascularization. In such Below the knee bypass accounts for 75% Most amputations can be prevented patients a choice must be made between of infrainguinal procedures in patients and limbs salvaged through a multiarmed prolonged medical therapy and primary with diabetes, with the anterior tibial/ treatment of antibiotics, debridement, re- amputation. dorsalis pedis artery the most common vascularization, and staged wound clo- Two general techniques of revascular- target vessel. Indeed, surgical bypass with sure. On the other hand, amputation may ization exist: open surgical procedures greater saphenous vein has become the offer an expedient return to a useful qual- and endovascular interventions. The two procedure of choice for patients with di- ity of life, especially if a prolonged course approaches are not mutually exclusive abetes and tibial disease. of treatment is anticipated with little like- and may be combined, such as iliac an- Advances in endovascular therapy, lihood of healing. Diabetic patients gioplasty combined with infrainguinal sa- particularly smaller instrumentation and should have full and active rehabilitation phenous vein bypass. The risks, expected standardization of thrombolytic therapy following amputation. Decisions should benefit, and durability of each must be for periprocedural thromboses, have al- be made on an individual basis with reha- considered. In either approach, meticu- lowed more aggressive use of tibial angio- bilitative and quality-of-life issues consid- lous technique, flexibility and resource- plasty. Despite this increased use, ered highly. fulness of judgement, and contingency however, the efficacy of tibial plans are important. Appropriate patient remains uncertain. Nonetheless, it may CONCLUSIONS — In summary, preparation, intra-procedure monitoring, provide a means to “buy time” to allow a PAD is a common cardiovascular compli- and postprocedure care will minimize patient to heal and recover from a limb- cation in patients with diabetes. In con- complications. threatening situation. trast to PAD in nondiabetic individuals, it Endovascular intervention is more The morbidity and mortality of vas- is more prevalent and, because of the dis- appropriate in patients with focal disease, cular surgical procedures in patients with tal territory of vessel involvement and its especially stenosis of larger more proxi- diabetes has improved significantly with a association with peripheral neuropathy, it mal vessels, and when the procedure is protocol of preoperative risk assessment is more commonly asymptomatic. performed for claudication. Open proce- and perioperative risk management, espe- Patients with PAD and diabetes thus dures have been successfully carried out cially with the use of ␤-blockers. The out- may present later with more severe dis- for all lesions and tend to have greater comes are now comparable with those of ease and have a greater risk of amputa- durability. However, open procedures are nondiabetic vascular patients. The choice tion. Moreover, the presence of PAD is a associated with a small but consistent of preoperative coronary artery bypass marker of excess cardiovascular risk. morbidity and mortality. The choice be- grafts (CABGs) is not encouraged, as the It is important to diagnose PAD in pa- tween the two modalities in an individual risk of two procedures (CABG and leg by- tients with diabetes to elicit symptoms, patient is a complex decision and requires pass) exceeds the risk of leg bypass alone. prevent disability and limb loss, and iden- team consultation. The decision for CABG should be based tify a patient at high risk of MI, stroke, and Aortoiliac disease is traditionally and on the same indications as for the nonop- death. The diagnosis is made with a de- effectively treated with prosthetic aorto- erative patient. termination of the ABI. It is recom- femoral bypass but is increasingly amena- Regular postoperative follow-up is mended that patients with diabetes who ble to endovascular angioplasty and mandatory because most late revascular- are Ͼ50 years of age have an ABI per- stenting. Although percutaneous angio- ization failures involve progression of in- formed. An ABI is also useful in patients plasty and stenting have achieved their timal hyperplasia at areas of anastomosis, with other PAD risk factors and in those best results in the aortoiliac vessels, open vein injury, valve sites, or angioplasty. with symptoms. revascularization probably offers results History, clinical exam, and the ABI are Treatment of the patient with diabe- that are more durable when diffuse aor- simple and effective methods of detecting tes and PAD should be twofold: 1) pri- toiliac disease or occlusion is present. major restenosis but may miss silent le- mary and secondary CVD risk factor

DIABETES CARE, VOLUME 26, NUMBER 12, DECEMBER 2003 3339 Peripheral arterial disease in people with diabetes modification and 2) treatment of PAD idemiological aspects. Vascular Medicine 6 creation. Arterioscler Thromb Vasc Biol 22: symptoms (claudication and critical limb (Suppl. 1):3–7, 2001 1370–1380, 2002 ischemia) and limiting progression of dis- 4. Bernstein EF, Fronek A: Current status of 17. Schneider DL, Sobel BE: Diabetes and ease. non-invasive tests in the diagnosis of pe- thrombosis. In Diabetes and Cardiovascu- It is the hope of this panel that by ripheral arterial disease. Surg Clin North lar Disease. Johnstone MT, Veves A, Eds. Am 62:473–487, 1982 arriving at a consensus of the fundamen- Totowa, NJ, Humana Press, 2001 5. Elhadd TA, Robb R, Jung RT, Stonebridge 18. American Diabetes Association: Preven- tals of assessment and management of this PA, Belch JJF: Pilot study of prevalence of tive foot care in people with diabetes (Po- devastating complication of diabetes, we asymptomatic peripheral arterial occlu- sition Statement). Diabetes Care 26 may effect more uniformity of care and sive disease in patients with diabetes at- (Suppl. 1):S78–S79, 2003 achieve better outcomes for our patients tending a hospital clinic. Practical Diabetes 19. Strandness DE Jr, Bell JW: Peripheral vas- with diabetes. We also strongly encourage Int 16:163–166, 1999 cular disease: diagnosis and evaluation clinicians to function cooperatively and 6. Hirsch AT, Criqui MH, Treat-Jacobson D, using a mercury strain gauge. Ann Surg effectively as teams of specialists in the Regensteiner JG, Creager MA, Olin JW, 161 (Suppl. 4):1–35, 1965 management of this complex patient pop- Krook SH, Hunninghake DB, Comerota 20. Lassila R, Lepantalo M: Cigarette smoking ulation, with the common goal of reduc- AJ, Walsh ME, McDermott MM, Hiatt and the outcome after lower limb arterial WR: Peripheral arterial disease detection, ing vascular events—MI, stroke, and surgery. Acta Chir Scand 154:635–640, awareness, and treatment in primary care. 1988 amputation—that too often result in dis- JAMA 286:1317–1324, 2001 ability, social decline, and death. 21. UK Prospective Diabetes Study (UKPDS) 7. Weitz JI, Byrne J, Clagett GP, Farkouh Group: Intensive blood-glucose control ME, Porter JM, Sackett DL, Strandness DE with sulphonylureas or insulin compared Jr, Taylor LM: Diagnosis and treatment of Acknowledgments— This conference was with conventional treatment and risk of chronic arterial insufficiency of the lower complications in patients with type 2 di- sponsored in part by educational grants from extremities: a critical review. Circulation AstraZeneca, Aventis Pharmaceuticals, Bristol- abetes (UKPDS 33). Lancet 352:837–853, 94:3026–3049, 1996 1998 Myers Squibb, Eli Lilly, GlaxoSmithKline, 8. Dormandy JA, Rutherford RB: Manage- Merck and Merck/Schering-Plough Pharma- 22. Stokes J, Kannel WB, Wolf PA, Cupples ment of peripheral arterial disease (PAD): LA, D’Agostino RB: The relative impor- ceuticals, Monarch Pharmaceuticals, Wyeth TASC Working Group: TransAtlantic In- Pharmaceuticals, Novartis, and Pfizer. tance of selected risk factors for various ter-Society Concensus (TASC). J Vasc Surg manifestations of cardiovascular disease 31:S1–S296, 2000 among men and women from 35 to 64 APPENDIX 9. Dolan NC, Liu K, Criqui MH, Greenland years old: 30 years of follow-up in the Fra- P, Guralnik JM, Chan C, Schneider JR, mingham Study. Circulation 75:V65– Mandapat AL, Martin G, McDermott MM: Consensus panel V73, 1987 Peripheral artery disease, diabetes, and re- Peter Sheehan, MD, Chair; Michael Ed- 23. UK Prospective Diabetes Study (UKPDS) duced lower extremity functioning. Dia- monds, MD; James L. Januzzi, Jr., MD, Group: Tight blood pressure control and betes Care 25:113–120, 2002 FACC (American College of Cardiology); risk of macrovascular and microvascular Judith Regensteiner, PhD; Lee Sanders, 10. McDaniel MD, Cronenwett JL: Basic data related to the natural history of intermit- complications in type 2 diabetes (UKPDS DPM; and Mellick Sykes, MD. tent claudication. Ann Vasc Surg 3:273– 38). BMJ 317:703–713, 1998 277, 1989 24. Yusuf S, Sleight P, Pogue J, Bosch J, Davies Presenters at the conference 11. Beckman JA, Creager MA, Libby P: Diabe- R, Dagenais G: Effects of an angiotensin- Christopher E. Attinger, MD; Joshua tes and atherosclerosis: epidemiology, converting-enzyme inhibitor, ramipril, Beckman, MD, FACC; Michael Criqui, pathophysiology, and management. on cardiovascular events in high-risk pa- MD, MPH; James B. Froelich, MD, MPH; JAMA 287:2570–2581, 2002 tients: the Heart Outcomes Prevention John M. Giurini, DPM; Linda Haas, PhC, 12. Ridker PM, Cushman M, Stampfer MJ, Evaluation Study Investigators. N Engl J Med 342:145–153, 2000 RN, CDE; Allen Hamdan, MD; Larry Har- Tracy RP, Hennekens CH: Plasma con- centration of C-reactive protein and risk 25. Leng GC, Price JF, Jepson RG: Lipid-low- kless, DPM; Michael R. Jaff, DO; Ishwarlal ering for lower limb atherosclerosis (Co- Jialal, MD, PhD; Frank LoGerfo, MD; of developing peripheral vascular disease. Circulation 97:425–428, 1998 chrane Review). Cochrane Database Syst Emile R. Mohler, MD; Jeffrey Olin, DO, 13. Veves A, Akbari CM, Primavera J, Dona- Rev 2:CD000123, 2000 FACP, FACC; Michael Pinzur, MD; Gayle ghue VM, Zacharoulis D, Chrzan JS, De- 26. Randomised trial of cholesterol lowering E. Reiber, MPH, PhD; Kenneth Rosen- Girolami U, LoGerfo FW, Freeman R: in 4444 patients with coronary heart dis- field, MD; Solomon Tesfaye, MD, FRCP; Endothelial dysfunction and the expres- ease: the Scandinavian Simvastatin Sur- Aristidis Veves, MD; and E. Kent Yucel, sion of endothelial nitric oxide synthetase vival Study (4S). Lancet 344:1383–1389, MD. in diabetic neuropathy, vascular disease, 1994 and foot ulceration. Diabetes 47:457–463, 27. Kjekshus J, Pedersen TR: Reducing the 1998 risk of coronary events: evidence from the References 14. Steinberg HO, Baron AD: Vascular func- Scandinavian Simvastatin Survival Study 1. Murabito JM, D’Agostino RB, Silbershatz tion, insulin resistance and fatty acids. (4S). Am J Cardiol 76:64C–68C, 1995 H, Wilson WF: Intermittent claudication: Diabetologia 45:623–634, 2002 28. MRC/BHF Heart Protection Study of cho- a risk profile from the Framingham Heart 15. Tsao PS, Wang B, Buitrago R, Shyy JY, lesterol lowering with simvastatin in 20, Study. Circulation 96:44–49, 1997 Cooke JP: Nitric oxide regulates mono- 536 high-risk individuals: a randomised 2. Hiatt WR: Medical treatment of periph- cyte chemotactic protein-1. Circulation placebo-controlled trial. Lancet 360: eral arterial disease and claudication. 96:934–940, 1997 7–22, 2002 N Engl J Med 344:1608–1621, 2001 16. Geng YJ, Libby P: Progression of ather- 29. Antiplatelet Trialists’ Collaboration: Col- 3. Criqui MH: Peripheral arterial disease: ep- oma: a struggle between death and pro- laborative overview of randomised trials

3340 DIABETES CARE, VOLUME 26, NUMBER 12, DECEMBER 2003 American Diabetes Association

of antiplatelet therapy. I: prevention of muscular exercise in patients with inter- trolled trials. J Am Geriatr Soc 50:1939– death, myocardial infarction, and stroke mittent claudication. Lancet 2:1093– 1946, 2002 by prolonged antiplatelet therapy in vari- 1096, 1966 35. Dawson DL, Cutler BS, Hiatt WR, Hobson ous categories of patients. Br Med J 308: 33. Leng GC, Fowler B, Ernst E. Exercise for RW 2nd, Martin JD, Bortey EB, Forbes 81–106, 1994 intermittent claudication (Cochrane Re- WP, Strandness DE: A comparison of 30. American Diabetes Association: Aspirin view). Cochrane Database Syst Rev 2: cilostazol and pentoxifylline for treating therapy in diabetes (Position Statement). CD000990, 2000 intermittent claudication. Am J Med 109: Diabetes Care 26 (Suppl. 1):S87–S88, 34. Regensteiner JG, Ware JE Jr, McCarthy 523–530, 2000 2003 WJ, Zhang P, Forbes WP, Heckman J, 36. Kannel WB, Risk factors for atheroscle- 31. CAPRIE Steering Committee: A random- Hiatt WR: Effect of cilostazol on tread- rotic cardiovascular outcomes in different ized, blinded, trial of clopidogrel versus mill walking, community-based walk- arterial territories. J Cardiovasc Risk aspirin in patients at risk of ischaemic ing ability, and health-related quality of 1:3333–3339, 1994 events (CAPRIE). Lancet 348:1329– life in patients with intermittent claudi- 37. American Diabetes Association: Diabetic 1339, 1996 cation due to peripheral arterial disease: foot wound care (Consensus Statement). 32. Larsen OA, Lassen NA: Effect of daily meta-analysis of six randomized con- Diabetes Care 21:1354–1360, 1999

DIABETES CARE, VOLUME 26, NUMBER 12, DECEMBER 2003 3341