M E D I C A L / S C I E N T I F I C S T A T E M E N T

Assessment of PeripheralVascular Disease In DiabetesReport and Recommendations of anInternational Workshop Sponsored by theAmerican Heart Association and the AmericanDiabetes Association18-20 September 1992, New Orleans,Louisiana

TREVOR J. ORCHARD, MD

D. EUGENE STRANDNESS, JR, MD

R ecognizing the considerable excessburden of both cerebrovasculardisease and LEAD suffered by in-

dividuals with IDDM or NIDDM (1,2), aworkshop was convened to 1) provide acurrent review of the knowledge pertain-ing to the prevalence, incidence, and riskfactor associations of cerebrovasculardisease and PVD in diabetes, and 2) re-view and make recommendations aboutthe methodology for identifying andquantifying LEAD in both clinical andresearch settings.

The workshop focused on the

specific problems of measurement ofPVD in diabetes and the need for a stan-dard approach to ensure appropriatecare and facilitate comparability of re-search findings across studies and overtime. Recommendations for the measure-ment of LEAD in a primary-care settingand for epidemiological studies were de-veloped, along with recommendationsfor more detailed assessment after refer-ral to a vascular clinic. Specific recom-mendations for the assessment of cere-brovascular disease were not addressed.The detection and appropriate measure-

FROM THE UNIVERSITY OF PITTSBURGH, GRADUATE SCHOOL OF PUBLIC HEALTH, PITTSBURGH, PENNSYL-

VANIA.

ADDRESS REPRINT REQUESTS TO THE OFFICE OF SCIENTIFIC AFFAIRS, AMERICAN HEART ASSOCIATION,

7272 GREENVILLE AVENUE, DALLAS, I X 75231-4596.

RECEIVED FOR PUBLICATION 16 FEBRUARY 1993 AND ACCEPTED 18 FEBRUARY 1993.

I D D M , INSULIN-DEPENDENT DIABETES MELLITUS; N I D D M , NON-INSULIN-DEPENDENT DIABETES MELLI-

TUS; LEAD, LOWER-EXTREMITY ARTERIAL DISEASE; PVD, PERIPHERAL VASCULAR DISEASE; ABI, ANKLE-

BRACHIAL INDEX; TSBP, TOE SYSTOLIC BLOOD PRESSURE; BP, BLOOD PRESSURE; HDL, HIGH-DENSITY

LIPOPROTEIN; LDL, LOW-DENSITY LIPOPROTEIN; CBF, CEREBRAL BLOOD FLOW; EDC STUDY, EPIDEMIOL-

OGY OF DIABETES COMPLICATION STUDY; SVA, SPECIALIST VASCULAR ASSESSMENT; CV, COEFFICIENT OF

VARIATION; NIDDK, NATIONAL INSTITUTE OF DIABETES AND DIGESTIVE AND KIDNEY DISEASES; NHLBI,

NATIONAL HEART, LUNG, AND BLOOD INSTITUTE; AHA, AMERICAN HEART ASSOCIATION; ADA, AMER-

ICAN DIABETES ASSOCIATION.

ment of LEAD was determined to be ofgreat importance, not only in terms ofLEAD, but also for the strong predictivepower that LEAD carries for subsequentcardiovascular mortality (3).

The workshop started with a se-ries of presentations, which are brieflysummarized below, followed by the de-velopment of specific recommendations.This report has been written on behalf ofall the participants who subsequently re-viewed, revised, and approved the finalversion.

SUMMARY OF PRESENTEDREVIEWS— Dr. J. David Curb re-viewed the current state of knowledgeconcerning the epidemiology of stroke indiabetes. Data from Framingham, theNational Health and Nutrition Examina-tion Survey, and the Systolic Hyperten-sion in the Elderly Program suggest atwofold increase in relative risk forstroke in subjects with diabetes, al-though in a further study (the RanchoBernardo Study) only a small increasewas seen for women with diabetes. Datafrom the Honolulu Heart Study sup-ported a relative risk of ~2 for peoplewith diabetes for developing stroke, andin common with some earlier reports,also suggest that the relative risk is notincreased for the subgroup with hemor-rhagic stroke. This is particularly trueafter adjustment for other risk factors.The reasons for this observation remainobscure and should be studied further.One possible explanation suggested wasthat the hypercoagulable state seen insome patients with diabetes may reducethe risk of bleeding.

Two presentations then focusedon the epidemiology of risk factors forLEAD. Dr. PJ. Palumbo reported datafrom the diabetes incidence cohort ofRochester, Minnesota (2,4-6) , thatshowed 8% of cases had LEAD at thetime of diagnosis of diabetes. The cumu-lative incidence of LEAD rose with ageand duration of diabetes to reach 45% by

DIABETES CARE, VOLUME 16, NUMBER 8, AUGUST 1993 1199

Assessment of PVD in diabetes

20 yr diabetes duration. In a multivariateanalysis of a cohort comprising bothnondiabetic and diabetic subjects withand without LEAD at baseline, progres-sion of LEAD was related to the presenceof both LEAD and diabetes as recordedat baseline, decreased postexercise ABI,increased systolic BP, and smoking. Dia-betes control (GHb) and lipoproteinprofile did not contribute further to theprediction of progression of LEAD.

In the Schwabing studies re-ported by Dr. Hans U. Janka, >600IDDM and N1DDM patients were pro-spectively followed for LEAD by ultra-sonic Doppler measurements (7,8). Atthe 5-yr, as well as a 9-yr follow-up, theincidence of mostly asymptomatic LEADwas consistently and significantly associ-ated with baseline systolic BP and thedyslipemic complex of high serum tri-glyceridesAow HDL cholesterol, but notwith total cholesterol, weight, or dia-stolic BP. The association with HbA^plasma C-pep tide, and daily insulin dose(in insulin-treated NIDDM) was weakerand at 9-yr follow-up only significant inunivariate analysis. Of diabetic patientsdying from cardiovascular causes withinthe 5-yr observation, 67% had LEAD(mostly asymptomatic) at baseline in con-trast with 15% in those who survived.

The clinical aspects of LEAD indiabetes were reviewed by Dr. Robert A.Graor. An increase in hospital mortalityand higher amputation rates amongthose with diabetes was stressed, as wellas the higher proportion of diabetic sub-jects with stenoses in the medium-sizedarteries below the knee. The importanceof determining the neuropathic as well asthe ischemic status of any diabetic sub-ject with foot ulcers was also stressed, aswas the need for protective care for anysubject with neuropathic lesions.

Attention then turned to method-ological issues. Dr. Niels A. Lassen pro-vided an overview of atraumatic methodsfor measuring CBF. The xenon133 inha-lation method showed normal levels ofCBF in long-term diabetes. This was thecase even in most patients with extracra-

nial atherosclerotic occlusive disease.However, a defect in CBF autoregulationmay be present, as manifested by thevariation in CBF that is observed with BPvariations. This defect, that also has beenshown by other techniques, probably re-flects increased stiffness of the hyalinizedarterioles. The ill-defined syndrome ofdiabetic encephalopathy, a form of vas-cular dementia in very-long-durationIDDM, may be related to these arteriolarchanges, that set the stage for subcorticallesions, particularly lacunar infarcts.

The role of the clinical exam forLEAD was reviewed by Dr. Michael H.Criqui (9). Using data from an elderly,population-based cohort of >500 sub-jects (5% of whom had diabetes), clinicalsigns and symptoms were correlatedwith LEAD defined by segmental BP ra-tio and flow velocity assessed by Dopplerultrasound. The recording of a pulse def-icit had a low predictive value for LEAD.The sensitivity of the Rose questionnairefor detecting LEAD was 9%, whereasonly 20% of those with LEAD had exer-cise-induced calf pain not present at rest.These low values were partly explainedby a high portion of LEAD being attrib-utable to isolated disease in the posteriortibial artery that would not lead to exer-cise-induced calf pain. A decreased orabsent posterior tibial pulse had a rela-tively high sensitivity (71%) for LEADand a fair (50%) positive predictivevalue. Criqui stressed that even isolatedposterior tibial disease carried a threefoldrisk of all-cause mortality and a fourfoldrisk of coronary heart disease mortality.

The usefulness of the ankle.armBP ratio (or ABI) was discussed in a pre-sentation by Dr. Donald E. McMillan,who described studies in nearly 400 di-abetic patients, 25% of whom hadIDDM. Patients with either low (<0.90)or high (> 1.30) ratios, suggestive of me-dial wall calcification, were studied. Ahigh prevalence of proteinuria was notedin diabetic patients <40 yr of age whowere found to have high ABIs. Little in-fluence of sex was noted. In those >40yr of age, the association of a high ABI

with proteinuria was lost. A high ABI wasobserved more often in men with littlerelation to either age or duration of dia-betes, as was a trend to higher serumcalcium levels. The milk intake of theaffected (high ABI) patients was notfound to be unusual. The frequency ofhigh and low ABIs together in the samepatient was low (—2% overall or 5% ofpatients with >1 abnormal ratio) andpredominantly in middle-aged men. Al-though concern exists about missing thediagnosis of occlusive disease because ofarterial calcification, the problem is lessthan had originally been expected. Theprevalence of low ABIs suggestive of oc-clusive disease increases with both pa-tient age and diabetes duration, and lowratios are almost as common in womenas in men.

Non-ABI methods of assessingLEAD were discussed by Dr. D. EugeneStrandness, Jr., who stressed the value ofmeasuring the TSBP (10) in patients re-ferred to a vascular clinic. This measure-ment overcomes the false elevation ofankle BPs from calcification and has asimilar repeatability to the ABI. In thevascular clinic, evaluation of velocity pat-terns Goss of normal triphasic or bipha-sic waveform, spectral broadening, andincreasing peak systolic velocity[>100%] between segments) are furtherfindings of significance. Recent improve-ments in duplex scanning techniques al-low such assessments from the aorta tothe ankle and permit localization of le-sions with good accuracy and precision.

The value of TSBP recording inthe management of diabetic foot lesionswas further demonstrated by Dr. PerHolstein (11,12), based on his 20 yearsof experience with this technique. Thespontaneous healing of ulcers was shownto be predicted by this measurement,such that a TSBP <20 mmHg was asso-ciated with only a 29% healing rate com-pared with a healing rate of 92% forsubjects with a TSBP >30 mmHg. Twothirds of limbs with a TSBP <30 mmHgeventually went on to amputation, as op-posed to none in those with a TSBP >30

1200 DIABETES CARE, VOLUME 16, NUMBER 8, AUGUST 1993

Orchard and Strandness

mmHg. TSBP also enables BP to be as-sessed in patients with leg ulcers wherethe use of an ankle cuff is not possible.

The issue of medial wall calcifica-tion was further addressed by two stud-ies that included X-ray data. Dr. Peter R.Cavanagh reviewed 94 patients with di-abetic sensory neuropathy who had un-dergone standardized weight-bearing ra-diographs and compared the findingswith those in 43 diabetic patients with-out neuropathy and 50 age-matchednondiabetic control subjects. The resultsshowed a strong and statistically signifi-cant association between the presence ofmedial arterial wall calcification and pe-ripheral sensory neuropathy. Arteries atthe level of the ankle were more fre-quently calcified than those at the toelevel. A history of foot ulceration wasalso associated with calcification, whichitself was more frequently seen in thearteries in the dorsum of the foot than inthe plantar arteries. Marked calcificationwas also associated with elevated serumcreatinine.

In a study of 147 participants inthe Pittsburgh EDC Study (based on anincident cohort of subjects with child-hood-onset IDDM), Dr. Trevor J. Or-chard reported a high prevalence of cal-cification in these subjects (mean age 34yr, mean duration of diabetes 26 yr);32% had calcification in the posteriortibial artery on ankle view X rays,whereas 28% had calcification in the an-terior tibial artery. Calcification was de-tected even more frequently in the foot;47% had calcification in arteries at themetatarsal level. Prevalence of calcifica-tion was, however, low in toe arteries,consistent with the data reported by Ca-vanagh. Arterial calcification was twice asprevalent in men as women and was un-related to the ABI in that it was present atall levels of ABI. A difference of 75mmHg between ankle and arm systolicBP gave a 100% positive predictive valuefor the presence of arterial calcificationby X ray, although the sensitivity waslow, between 6 and 9% (depending onwhich arteries were studied).

Data concerning the relationshipbetween resting and postexercise ABIwere also presented. Data from 657IDDM subjects from the first cycle of thePittsburgh EDC Study (13) showed that4% of subjects had an ABI <0.8 and15% had an ABI <0.9 at 1 min postex-ercise compared with only 1 and 5%,respectively, at rest. Though the postex-ercise test would appear to be more sen-sitive, it was noted that a resting ABI of0.80-0.89 had only a 26% sensitivity ofpredicting a postexercise ABI of <0.80.An analysis of risk factors and complica-tions suggested that subjects with an ABI<0.80 at rest or with an ABI <0.90 atrest and <0.80 postexercise had a greatlydisturbed lipoprotein profile, were morelikely than other subjects to have hyper-tension, and were more likely to havebeen smokers. This subgroup also had ahigh prevalence of neuropathy (75%),overt nephropathy (42%), and prolifera-tive retinopathy (75%). They also had ahigh 2-yr incidence of overt nephropathyand retinopathy of 40 and 33%, respec-tively. Such associations were not seenfor those positive (ABI <0.90) by eitherexercise or resting ABI alone. The 25subjects with an ankle systolic BP 75mmHg higher than the arm (who werepresumed, therefore, to have medial ar-terial wall calcification) showed a strik-ing association of disturbed lipoproteinprofiles, and an extremely high preva-lence of hypertension (60%), neuropathy(88%), overt nephropathy (88%), andproliferative retinopathy (92%).

In the final presentation of data,Dr. William R. Hiatt confirmed that aweak relationship existed between rest-ing and postexercise ABI in a population-based study of both diabetic and nondi-abetic subjects from the San Luis ValleyDiabetes Study (14). Subjects diagnosedwith peripheral arterial disease by theresting ABI did not necessarily have anabnormal postexercise ratio (the con-verse was also true). Additional findingsfrom this study concerned the develop-ment of criteria for an abnormal ABI.Normal ranges for the resting ABI were

determined from a healthy nondiabeticcontrol subset of the study population.This analysis revealed that women hadlower ABIs than men, and that the ABIderived from the dorsalis pedis arterywas lower than the ABI from the poste-rior tibial artery. These findings raise theneed for further investigation as towhether definitions of an abnormal ABIshould be both sex- and vessel-specific.

RECOMMENDATIONS— Based onthe data presented at the workshop, pre-vious published data, and extensive dis-cussion, the workshop participants de-veloped the following recommendationsfor the detection and follow-up of LEADin the primary-care setting and for moredetailed assessment after referral to aspecialized vascular clinic. In addition,further comments are provided to helpextend the recommendations made foruse in the detection of LEAD in theprimary-care setting to use in epidemio-logical studies.

Recommendations for the detectionand follow-up of LEAD in diabeticsubjects being followed in aprimary-care settingThe following four items can be ad-dressed as part of the generally recom-mended annual physical examination forpatients with diabetes.

1. ClaudicationOn an annual basis, diabetic patientsshould be asked about the presence ofexercise-induced calf leg pain notpresent at rest. Patients with lifestyle-limiting exercise-induced calf painshould be referred for SVA. SVA may bea vascular laboratory, clinic, or specialist,depending on local facilities and physi-cian preference. The SVA will allow con-firmation of LEAD, disease localization,and possible determination of the TSBP.TSBP will permit a more accurate assess-ment of vascular status independent ofmedial wall calcification. Measurementof an ABI or referral for SVA should also

DIABETES CARE, VOLUME 16, NUMBER 8, AUGUST 1993 1201

Assessment of PVD in diabetes

be considered for patients with any legpain not clearly ascribed to a nonvascularcause. The ABI, which should be as-sessed according to the protocol de-scribed later, will help determinewhether this pain may be caused by isch-emia. Because ischemic leg pain has anegligible prevalence in diabetic chil-dren, this recommendation is limited toIDDM or N1DDM patients > 18 yr of age.

2. Signs of critical ischemiaThe presence of any potential signs ofcritical ischemia, i.e.. foot or limb ulcer-ation. the presence of skin changes Cnailor skin atrophy, or dependent rubor). orthe detection of gangrene, should lead toa referral for SVA. Because of the poten-tial for medial wall calcification to raiseankle sBP above the normal range, evenin the presence of occlusive disease, theABI alone may not be sufficient to detectvascular insufficiency. Further testing,e.g., TSBP may be indicated (see below).

3. Palpation of peripheral (tibialisposterior and dorsalis pedis)pulsesPalpation of leg pulses should be per-formed on an annual basis for all adult(> 18 yr old) diabetic patients. An absentor decreased tibialis posterior pulse is anindication for performing an ABI (seebelow) or referral to a vascular laboratoryfor evaluation if an ABI cannot be deter-mined by the primary physician. It isfurther recommended that, wheneverpossible, the presence of decreased orabsent pulses be confirmed by a secondobserver or repeat examination before re-ferral.

Palpation of peripheral pulses hasa limited but useful place in the detectionof LEAD. Although repeatability and in-terobserver agreement are low, sensitiv-ity and positive predictive value aremoderate for the detection of LEAD;therefore, a significant number of caseswill be identified by detection of a reduc-tion or absence of these pulses. Further-more, the presence of these pulses inlow-risk diabetic subjects helps confirm

the absence of significant disease. Notethat the dorsalis pedis pulse is sometimescongenitally absent.

4. Femoral bruitsAuscultation for femoral bruits on an an-nual basis is recommended for all adultdiabetic patients. The detection of femo-ral bruits is an indication for performingan ABI (see below) or, if not available,referral to a vascular laboratory. Al-though it is recognized that auscultationfor femoral bruits has similar difficultiesto those described for pulse palpation, itnonetheless also has sufficient sensitivityto merit its performance on an annualbasis.

ABIIt is recommended that all physician of-fices providing routine care to adult dia-betic patients should be able to measureankle/brachial BP to detect LEAD. Theadditional equipment needed is minimal,e.g., a hand-held ultrasound doppler of5-10 MHz that costs ~$300-$600. Thesame BP cuffs used for the arm can, inmost cases, be used for the ankle BP (seeAPPENDIX 1). The justification for this rec-ommendation is the low sensitivity ofclinical history and examination to detectLEAD (9) and the high morbidity andmortality seen in patients with this con-dition (3). The objective is not only tooptimize care for the developing isch-emic limb, but also to identify a high-riskgroup of diabetic subjects for general car-diovascular disease in whom risk factormodification should be maximized. ABImeasurement is recommended for thefollowing situations:

• Any diabetic patient who has newlydetected decreased pulses, femoralbruits, or a foot ulcer.

• Any diabetic patient with leg pain ofunknown etiology.

• At baseline examination in all IDDMpatients >35 yr of age or with >20 yrduration of diabetes.

• At baseline examination in all NIDDMpatients >40 yr.

For a detailed protocol for performingthe ABI, see APPENDIX 1. Some majorpoints include the following:

• The patient should be strictly supinefor at least 5 min before testing.

• Arm as well as ankle BPs should betaken with the doppler.

• Ideally, both right and left arm pres-sures should be measured and thehigher one used.

• Ideally, both dorsalis pedis and tibialisposterior pulses should be measured ineach leg.

• The same sequence of recordings be-tween patients and over time should beused.

• The ABI should be calculated based onthe brachial BP closest in time to thespecific ankle BP being evaluated.

• The suggested sequence of measure-ment is pressure 1, right arm, left arm,taking highest reading; pressure 2,right dorsalis pedis; pressure 3, righttibialis posterior; pressure 4, arm withhighest reading as determined for pres-sure 1; pressure 5, left dorsalis pedis;pressure 6, left tibialis posterior; pres-sure 7, arm with highest reading asdetermined for pressure 1.

• Calculate ABI ratios as

right dorsalis pedis =

right tibialis posterior =

left dorsalis pedis =

left tibialis posterior =

pressure 2

pressure 1

pressure 3

pressure 4

pressure 5

pressure 4

pressure 6

pressure 7

• Ratios calculated in this way relate theankle and arm BPs taken closest to-gether in time and are therefore moremeaningful.

Recommended action to take after ABItesting:ABI <0.50 in any vessel. Prompt refer-ral for SVA is recommended becausethese patients almost certainly have se-vere PVD.

1202 DIABETES CARE, VOLUME 16, NUMBER 8, AUGUST 1993

Orchard and Strandness

ABI >0.50 and <0.90 in any vesseltested. The ABI should be repeatedwithin 3 mo because these patients arelikely to have mild to moderate PVD. Ifrepeat ABI is <0.90, start intensive riskfactor modification and annual ABI fol-low-up. If repeat ABI is >0.90, repeatABI every 2 or 3 yr.ABI >0.90. Repeat testing every 2 to 3yr because these patients are unlikely tohave PVD.Any incompressible ankle artery (an-kle systolic BP >300 mmHg) or ankleBP >75 mmHg above arm BP. Thesevalues should be repeated within 3 mo; ifconfirmed, refer these patients for SVAbecause they almost certainly have sig-nificant medial wall calcification, andmeasurement of ankle BP to determineLEAD is therefore compromised. Thesepatients are also at high risk for both mac-rovascular and microvascular complica-tions and should be entered into an inten-sive risk-factor modification program.

Note that repeatability of the ABImeasure is moderately low (CVs between10 and 15% have been reported). Thus,like arm BP readings for hypertension orcholesterol measurements, it is impor-tant to confirm abnormal values beforereferral or intensive risk-factor interven-tion. Variability can be reduced by per-forming all BPs in duplicate and calcu-lating their mean.

Postexercise ABIThis measure is not recommended foruse in the primary-care setting.

Intensive risk-factor modificationAll patients with confirmed ABI <0.90r

and/or ankle systolic BP >300 mm Hg.and/or ankle BP > 75 mmHg above armBP. and/or exercise-induced calf pain notpresent at rest, should be carefullyscreened for cardiovascular risk factorsincluding, if not already available, twofasting lipoprotein profiles (total choles-terol. HDL cholesterol. triglyceridest andcalculated LDL cholesterol). In thosewith diabetes, triglyceride concentrationappears to have a greater cardiovascular

predictive power than in the generalpopulation and is also particularly re-lated to LEAD (1,15-17). Therefore,careful attention should be paid to cor-rection of an elevated triglyceride con-centration (by means of good glycemicand weight control) in addition to that ofa low HDL-cholesterol level and a raisedLDL-cholesterol level. The NationalCholesterol Education Program guide-lines (18) for the management of choles-terol in adults and the recommendationsin the consensus statements of the Amer-ican Diabetes Association (19,20) shouldbe followed. It is recommended that di-abetic patients with LEAD be treated asif they had coronary heart disease forthese purposes. Recent reviews concern-ing lipid management in diabetes are alsoavailable (21,22). BP should also be care-fully monitored and managed accordingto recommendations for people with di-abetes (23). Cessation of smoking in thisgroup is of paramount importance, andall available approaches should be con-sidered. Other interventions to reducecardiovascular risk should also be care-fully considered (weight control and ex-ercise may be of specific benefit in thosewith LEAD). Although no specific rec-ommendation is being made concerningaspirin use, evidence from secondaryprevention trials in nondiabetic subjectsindicates that aspirin has a protective ef-fect on subsequent cardiovascular mor-tality and morbidity (24). Recent datafrom the Early Treatment of Diabetic Ret-inopathy Study (25) suggest that aspirinis not harmful and may provide protec-tion against myocardial infarction in di-abetic subjects. Although it is suggestedthat risk factor modification be intensi-fied in those with an ABI <0.90, itshould be remembered that all diabeticsubjects are at increased risk for cardio-vascular disease, and close control of BPand serum lipids and smoking cessationis generally encouraged.

Other complications of diabetesLEAD and medial arterial wall calcifica-tion are often associated (especially in

IDDM) with an increased prevalence ofnephropathy, proliferative retinopathy,and neuropathy (13). Therefore, it isstrongly recommended that, at a mini-mum, existing guidelines for detection ofthese complications be followed. Guide-lines currently exist for neuropathy(26,27) and retinopathy (28). Distal sen-sory polyneuropathy is particularly im-portant to detect at an early stage so thatappropriate protective care of the feetcan be initiated. The detection and/orpresentation of any neuropathic foot le-sion with good perfusion (ABI >0.9 inall vessels) should lead to a referral to aspecialist foot-care team.

Detection, management, andfollow-up of LEAD in diabeticpatientsFor a summary of recommendations, seeTable 1.

Specific observations on the use ofprimary-care recommendations inepidemiological studiesClaudication. The current Rose ques-tionnaire (29), if applied strictly, misses asignificant number of patients withLEAD. This can be partly overcome byincluding as positive any exercise-relatedleg pain not present at rest. The currentRose questionnaire also does not permitassessment of whether one or both legsare affected.Pulses and bruits. It is strongly advisedthat the detection of reduced or absentpulses and the presence of femoral bruitsbe confirmed by a second trained ob-server. These findings should not beused alone as primary end points in ep-idemiological studies.ABI.• Full adoption of all recommendations

contained in this report, includingthose in APPENDIX 1, is strongly advised.

• All observers should be trained so thatthe CV over multiple measures forboth normal and diseased subjects is

• Interobserver variation should be as-sessed and minimized so that measure-

DIABETES CARE, VOLUME 16, NUMBER 8, AUGUST 1993 1203

Assessment of PVD in diabetes

Table 1—Summary of recommendations for detection, management, and follow-up of LEAD in diabetic patients

Test Diabetic patients Frequency Action

Claudication

Signs of critical ischemia

Peripheral pulses/femoralbruits

ABI

All adults (>18 yr of age)

All adults (>18 yr of age)

All adults (> 18 yr of age)

1DDM adults (>35 yr of ageor with >20 yr diabetesduration)

NIDDM adults (>40 yr of age)

Annually

Annually

Annually

Dependsonbaselineresult

If present, do ABI annually. If ABI <0.90, startIRFM. If present and life-style limiting, referfor SVA.

If present (i.e., gangrene, ulcer, skin changes, orischemic rest pain), refer for SVA and startIRFM.

If abnormal, do ABI annually. If ABI <0.90, startIRFM.

If ABI is <0.50, refer for SVA and start IRFM.If ABI is 0.50-0.89, repeat within 3 mo. If con

firmed to be <0.90, start IRFM and do ABIannually. If confirmed to be >0.90, repeatevery 2 -3 yr.

If ABI is >0.90, repeat every 2 -3 yr.If ankle BP is > 75 mmHg above arm BP, repeat

within 3 mo. If confirmed, refer for SVA andstart IRFM. If not confirmed, repeat every 2 -3yr. If ankle BP is >300 mmHg, refer for SVAand start IRFM.

merit of the same individual by all ob-servers is within 15% of all otherobservers' readings. Use of a double-headed stethoscope (in conjunctionwith the ultrasound device) and strictadherence to the protocol is recom-mended, as is the recording of all BPsin duplicate and using the mean of thetwo BPs.Although postexercise ABI is not rec-ommended in the primary-care setting,the validity of this modality and itsrelationship to resting ABI, risk factors,and complications and its predictivepower of future morbidity and mortal-ity need to be studied further. Consid-eration should therefore be given tousing this test with a standard workload (e.g., 5 min walking at 2 mph ona 12% incline), with ABI being mea-sured at 1, 3, and 5 min postexercise.

NONINVASIVE VASCULARDIAGNOSTIC LABORATORY

Referral categoriesWhen the physician finds one or more ofthe following conditions, as discussed in

the primary-care recommendations, re-ferral to a vascular laboratory or clinic forfurther study may occur. The testingsubsequently to be done will depend onpatient presentation, clinical findings,and the results of preliminary screeningdone in the physician's office.Intermittent claudication. It is recog-nized that exercise-induced leg pain is acommon sequela of arterial narrowingand occlusion at one or more levels ofthe arterial blood supply to the leg. Theextent to which further testing needs tobe done will depend on the severity ofthe problem, particularly as it relates todaily activity. In general terms, those pa-tients whose exercise-induced pain af-fects their life-style will deserve the mostcomplete evaluation. The most commonsite for exercise-induced pain is in thecalf, but it can also develop in the thigh,hip, or buttock when the disease is lo-calized above the inguinal ligament. Of-ten the pain will start in the calf and thenprogress to the thigh and/or buttock ifexercise is continued despite the onset ofpain. It must be remembered that severeclaudication is most often the result ofmultilevel arterial disease, which can be

best evaluated in the noninvasive labora-tory.Critical ischemia. The clinical defini-tion of critical ischemia is a clinical pre-sentation that, if not reversed, is likely toresult in an amputation either at the levelof the foot or below the knee. In somerare cases, amputation may be requiredat the above-knee level. Because therapyis so critical for limb preservation, it isessential that patients with critical isch-emia have a complete noninvasive eval-uation before arteriography (unless life-threatening infection is present, in whichcase immediate arteriography may be in-dicated). The categories of presentationthat are relevant to this subset of patientsare as follows:

• Ischemic rest pain: In this setting, thepain will be in the toes and forefoot. Itwill, during its early phases, be re-lieved by dependency. If it does notimprove with development of collat-eral circulation, amputation will be in-evitable unless some form of interven-tion (surgical or endovascular) isconducted.

• Ulceration: When a break in the skinoccurs at any location of the foot or

1204 DIABETES CARE, VOLUME 16, NUMBER 8, AUGUST 1993

Orchard and Strandness

lower leg, healing might not occur un-less some form of intervention is con-ducted. The single exception may befor a patient who has an ulcer over apressure point or site of direct injurythat is secondary to neuropathy. How-ever, it must be remembered that isch-emia and neuropathy may coexist,making patients in this subset candi-dates for amputation unless the areasof occlusion are either bypassed oreliminated by transluminal angio-plasty.

• Gangrene: Tissue death, when it in-volves one or more toes or the forefoot,will require amputation that may belimited to the involved areas if directintervention can bring more blood tothe ischemic area.

• Skin changes: Although not as defini-tive as the preceding categories, skinatrophy, nail changes, and dependentrubor occur in some patients who mayrequire further evaluation. This is par-ticularly true if the ABI is found to beabnormal.

Abnormal ABI. In the absence of any ofthe above conditions, a referral to a vas-cular laboratory is recommended forsubjects whose ABI is confirmed to be<0.50 in one or more vessels. This willpermit confirmation, localization, andpossible assessment of TSBP index andwill provide a more accurate picture ofvascular status independent of medial ar-terial wall calcification.

The finding of incompressible ar-teries (i.e., systolic BP >300 mmHg) atthe level of the ankle makes it impossibleto estimate the systolic BP. In some cases,the arteries may be compressible but maygive falsely high levels of recorded pres-sure. If the ankle systolic BP is 75 mmHgor higher than the arm systolic BP and/orthe ABI is >1.30, medial arterial wallcalcification is virtually certain and mea-surement of the TSBP index in the vas-cular laboratory should be performed.

Recommended vascular laboratorystudiesStudies done in the resting state. It ispossible to determine the severity of the

occlusive disease by a combination ofmeasurement of the ABI and a toe sBPindex. It is recommended that both testsbe done to provide confirmation of thevascular status. There are various cutofflevels that must be considered in con-ducting these tests. These are as follows:

• ABI: If the ABI is <0.90, occlusive ar-terial disease may be present. If a valueof <0.80 is noted, it is highly likelythat vascular disease will be found.Levels between 0.50 and 0.80 arelikely to be found in patients with sin-gle-segment occlusions, whereas ABIs<0.50 are commonly found in patientswith multisegment disease.

• TSBP index: When measurements aremade at the level of the toes, there aretwo cutoff values that are of clinicalvalue. First, for screening purposes aTSBP index of >0.60 ± 17% is nor-mal. Second, the absolute levels of sys-tolic BP are of great value in estimatinghealing potential when an ulceration isfound. If the absolute pressure is <30mmHg, healing is unlikely to occurunless some form of direct interventionis conducted.

• Alternative forms of study: The ABI isan overall indicator of obstructive dis-ease in the legs, the aorta, or both.Thus, other supplemental modalitiesare desirable. Measurement of segmen-tal systolic BP permits, to some degree,the localization of sites of arterial oc-clusion in the limbs and may provide amore sensitive indicator of progres-sion. Similarly, pulse volume recordingtaken at several levels of the limb islargely a qualitative testing procedurethat provides data on the pattern ofvolume changes in response to pres-sure changes. The waveforms recordedwill reflect underlying occlusive dis-ease but are rarely specific as to exactsites of involvement.

Another widely used method, ve-locimetry, uses continuous wave Dop-pler techniques. By using systems thatpermit an accurate display of the ve-locity patterns (fast fourier transformspectral analysis), it is possible to con-firm the presence of arterial disease ator proximal to the recording sites. Theprocedure can be used at the femoral,

popliteal, and tibial arteries at the an-kle. Loss of the normal triphasic wave-forms normally seen can be taken ascertain evidence of arterial disease.

Because the ABI and TSBP indexprovide quantitative indexes, they areto be considered the definitive diag-nostic studies. More qualitative testscan be added to provide additional in-formation.

Exercise testing. Not all patients withexercise-induced leg pain will have arte-rial occlusive disease as the cause. It iswell known that neurospinal disease andmusculoskeletal disorders can also leadto leg pain and can be confused with truevascular intermittent claudication. Whenthe etiology is uncertain, the cause canbe elucidated by exercise testing.

The separation becomes possiblebecause intermittent claudication sec-ondary to arterial disease is always ac-companied by a decrease in ankle sys-tolic BP after exercise. This is a stress test,but the levels of exercise required tomake this distinction are minimal.

Although several methods of ex-ercise can be used to bring out the ab-normality associated with arterial dis-ease, it is recommended that a treadmillbe used with a standard elevation andspeed. A speed of 2 mph at a 12% ele-vation is most commonly used. Thisspeed and elevation can be tolerated bynearly all patients but can be decreased ifthe user desires. Regardless of the speedand elevation, it is important that thepatient walk to the point of pain andpreferably to the point of being unable tocontinue or for a full 5 min. It has beenfound that periods of walking for >5min provide little further information.

The patient with true claudica-tion will sustain a fall of >20% in theankle systolic BP with a recovery time>3 min. It is also possible to express thepostexercise results in terms of ABI. Ei-ther approach is acceptable as long as thevascular laboratory has experience withthis testing procedure. At the recom-mended low work load, patients withnormal limbs or those with pain second-

DIABETES CARE, VOLUME 16, NUMBER 8, AUGUST 1993 1205

Assessment of PVD in diabetes

ary to nonvascular causes will not have afall in ankle pressure as noted above.

Concern must be given to the pa-tients with known coronary artery dis-ease who have angina pectoris. No firmguidelines have been established for thisgroup of patients. However, adverse car-diac events at this workload are ex-tremely rare. Some laboratories may wishto combine the testing procedure withECG monitoring, but this is not abso-lutely necessary.Ultrasonic duplex scanning. The intro-duction of duplex scanning has added anew dimension to the field of vasculartesting. While the technology is evolving,it clearly will play an increasingly impor-tant role in vascular testing. Based oncurrent evidence, the following recom-mendations are made.

1. There is no justification for its useas a routine screening procedure. Itis not necessary to establish thepresence or absence of arterial oc-clusive disease.

2. If information is desired as to theexact location of arterial disease,duplex scanning can provide thisinformation. However, it cannot beused as an index of severity, whichis best determined by clinical pre-sentation, the ABI, and TSBP index.

3. In some centers, duplex scanning isbeing used as a method of selectingthe appropriate interventional ther-apy. If available, it can be so used,but this is not widely practiced atpresent, even though there are in-dications that this use will becomemore important in the future.

Tissue PO2. The likelihood of healing ofischemic skin lesions may be assessed bythe measurement of tissue PO2 levels. Itis generally believed that values >20 ton-indicate questionable capacity for heal-ing and those >10 torr are not compat-ible with healing. Thus, the need for ad-ditional medical or surgical referral willbecome apparent when this test indicatesnonviability of injured tissue. PO2 is alsoa sensitive indicator of improvement in

blood flow, supplemented by either col-lateral flow development or surgicaltreatment. Another possible use of thismeasurement is determination of thelevel of amputation when this becomesnecessary.Arteriography. Although invasive, arte-riography remains the definitive diagnos-tic procedure before any form of surgicalintervention. However, it should not beused as a diagnostic procedure to estab-lish the presence of arterial disease. Sev-eral considerations are relative to its useand must be kept in mind.

1. Patients with diabetes mellitus mayalso have renal disease, whichplaces them at high risk for renalfailure secondary to the use of con-trast material. This risk must betaken into account whenever arte-riography is contemplated. Thevolume of contrast material usedmust be limited as much as possi-ble.

2. It is strongly recommended that ar-teriography be done before an am-putation to assess the exact statusof the arterial supply. Although it isrecognized that some patients withneuropathy and irreversible tissuedamage do not have arterial dis-ease, confirmation of the status ofthe arterial supply is necessary,particularly when the ABI and toesBP index indicate that arterial dis-ease is present.

Vascular laboratory accreditationBecause the quality of vascular diagnosticlaboratory studies is key to the evalua-tion process, it is strongly recommendedthat vascular laboratories participate inthe voluntary accreditation process cur-rently in place. The Intersocietal Com-mission for the Voluntary Accreditationof Vascular Laboratories, which is spon-sored by several societies, has establishedrigid criteria and procedures for certifi-cation. This will provide assurance to theuser that procedures for optimal perfor-mance are followed. Further details canbe obtained by writing to the Intersoci-

etal Commission for Accreditation ofVascular Laboratories, 11200 RockvillePike, Suite 205, Rockville, MD 20852-3139.

FUTURE RESEARCHNEEDS— During the course of delib-erations, numerous research needs wereidentified. These are specified below inno specific order of priority except thefirst. The recommendations that inten-sive risk-factor intervention be instigatedin those with LEAD, as outlined above,are based on the strength of risk-factorassociations and extrapolation of resultsfrom clinical trials in the general popu-lation that have generally excluded dia-betic subjects. The absence of clinicaltrial data (especially in terms of correc-tion of disturbed lipoprotein profiles) indiabetes is most worrisome, and theNIDDK, NHLBI, AHA, and ADA arestrongly urged to mount appropriate re-search initiatives in this area.

Research recommendations

1. Clinical trial evidence of the benefit(e.g., cardiovascular disease pre-vention, limb preservation) of risk-factor modification for the preven-tion of macrovascular disease indiabetic patients, especially focus-ing on lipids and lipoproteins,should be conducted. To date,most prospective trials of cardio-vascular disease prevention haveexcluded diabetic patients. Bothprimary and secondary preventionstudies should be considered.

2. The Rose questionnaire needs to berevised for epidemiological use tominimize false-negatives and ob-tain data on laterality.

3. Further study is needed on theprognosis of arterial calcificationand, in particular, its associationwith nephropathy. Could it beused as an early marker?

4. Further study is needed to deter-mine whether a sex difference ex-ists in ABI measures.

1206 DIABETES CARE, VOLUME 16, NUMBER 8, AUGUST 1993

Orchard and Strandness

5. Further research is needed to eval-uate why there is such a poor as-sociation between abnormal restingABI (e.g., <0.9) and abnormal pos-texercise ABI (e.g., < 0.8).

6. An evaluation of the efficacy andcost-effectiveness of establishing afoot clinic/foot-care team shouldbe performed.

7. The interrelation of various nonin-vasive testing approaches and theirrelationships to arteriographicallydefined LEAD needs to be formallyevaluated. It is recommended thatduplex scanning, Doppler flow ve-locity, magnetic resonance arte-riography, exercise and restingABIs, claudication, palpation ofpulses, and detection of femoralbruits all be studied in terms oftheir relationships to each otherand to the gold standard of arterio-graphically defined LEAD.

Acknowledgments—This workshop wassponsored by the Council on Epidemiologyand Statistics of the ADA and the Council onEpidemiology and Prevention of the AHA. Itwas cosponsored by the AH As Stroke Coun-cil and Council on Arteriosclerosis and theADA's Council on Foot Care and Council onComplications.

APPENDIX 1— PROTOCOL FORMEASUREMENT OF THE ABIINSTRUMENTATION— The mostpractical device for the assessment ofboth ankle and arm systolic BP is thecontinuous-wave Doppler system. Ahand-held device that has either a smallspeaker for the output or a stethoscopeheadset that can be used to listen to theblood velocity patterns from the brachial,radial, dorsalis pedis, and posterior tibialarteries at the ankle.

Cuffs.The same cuffs used for measurement ofarm BP should be used at the ankle.Similar concerns with regard to the needfor appropriate-sized cuffs for obese arms

apply also to the ankle, and an appropri-ate cuff the width of which is —40% ofthe limb in circumference should beused at both sites (30). In most cases, aregular arm BP cuff will be adequate.

Method of measurement.Several aspects of BP measurement mustbe observed if the examiner is to obtainthe best and most reproducible results.The recommendations are as follows:

1. The BPs should be measured inboth arms; use the higher of thetwo.

2. It is preferable to have the patientrest quietly and supine for at least 5min before the BPs are measured.This ensures that any changes in BPthat might have occurred becauseof previous walking have had achance to stabilize.

3. The Doppler transducer should beplaced at an angle of 60° to theartery being tested so that the bestvelocity signals will be obtained. Ifthe transducer is at right angles tothe artery, the detected velocity(frequency shift) will be at a mini-mum and often difficult to hear,particularly if the velocities are re-duced secondary to arterial diseaseproximal to the recording site. Inaddition, care must be taken tomaximize the Doppler signal bymoving the probe back and forthover the artery to obtain the loud-est signal. Failure to do so mayresult in underestimation of the ar-terial BP measurement.

4. To obtain the most reproducibleand accurate measurements the fol-lowing rules must be followed:

• BP must be taken at the point at whichthe Doppler signal first appears duringdeflation of the cuff (i.e., BP is nevertaken during cuff inflation).

• The cuff should be inflated to at least20 mmHg above arm systolic BP levelsto ensure complete collapse of the dor-salis pedis and posterior tibial arteries.

• The BP that is recorded is taken at thepoint at which flow returns as detectedby the Doppler system.

• Cuff deflation must be slow (e.g., 2mmHg/s) to accurately determine thepoint at which flow is restored.

• It is mandatory that the Doppler sys-tem be used to measure the arm sBP.To determine the arm BP, the examinermay use either the brachial artery distalto the cuff or the radial artery at thewrist.

When the arterial signals at the level ofthe ankle cannot be obliterated by cuffinflation (e.g., BP is >300 mmHg), thisis conclusive evidence that medial calci-fication is present. Similarly, when theankle BP is >75 mmHg above the armBP or the ABI is >1.3, partial incom-pressibility because of medial calcifica-tion is likely to be present, giving rise tofalsely high ankle BP.

APPENDIX 2 — PROTOCOL FORMEASUREMENT OF TSBPINSTRUMENTATION— A s t r a i ngauge or a photoplethysmograph can beused as a sensor. (The strain gauge also ispractical for measurement of ankle BPs).The cuffs used to measure toe BPs shouldbe 2.4 cm wide by 10 cm long (encir-cling the first toe).

Method of measurement.To obtain useful and reproducible re-sults, the following aspects must be ob-served:

1. It is preferable to have the patientrest quietly and supine for at least 5min before the BPs are measured.

2. Room temperature should be com-fortable (more than —22°C) so thepatient does not feel cold.

3. The skin temperature of the toesshould be at least 25°C. If neces-sary, the feet should be warmedgently for 10 min in tempered wa-ter (37-39°C). In case of a skinlesion, a protective plastic bag canbe drawn over the foot.

4. Before inflation of the miniaturecuff, the toe should be squeezedgently between the examiner's fin-

DIABETES CARE, VOLUME 16, NUMBER 8, AUGUST 1993 1207

Assessment of PVD in diabetes

gers to empty the vessels of blood.This produces a more clean-cutcurve. At very low BPs, a stepwisedeflation of the cuff will often facil-itate reading of the tracings.

APPENDIX 3 — WORKSHOPPARTICIPANTS— Peter R Cavanagh, PhD, Penn State University, Uni-versity Park, PA; John A. Colwell, MD,PhD, Medical University of South Caro-lina, Charleston, SC; Michael H. Criqui,MPH, MD, University of California, SanDiego, CA; J. David Curb, MPH, MD,University of Hawaii, Honolulu, HI;Robert A. Graor, MD, Cleveland Clinic,Cleveland, OH; William R. Hiatt, MD,University of Colorado, Health ScienceCenter, Denver, CO; Per Holstein, MD,Bispebjerg Hospital, Copenhagen, Den-mark; Hans U. Janka, MD, Zentral Kran-kenhaus Bremen-Nord, Bremen, Ger-many; Niels A. Lassen, MD, BispebjergHospital, Copenhagen, Denmark;Donald E. McMillan, MD, Diabetes Cen-ter, University of South Florida, Tampa,FL; Trevor J. Orchard, MD, University ofPittsburgh, Graduate School of PublicHealth, Pittsburgh, PA; P. J. Palumbo,MD, Mayo Clinic, Rochester, MN; Ro-nald J. Prineas, MD, PhD, University ofMiami School of Medicine, Miami, FL(nonattending consultant); SolomonRosenblatt, MD, West Bloomfield, MI; J.Michael Sprafka, MPH, PhD, Universityof Minnesota, Minneapolis, MN; GeorgeSteiner, MD, Toronto General Hospital,Toronto, Canada (nonattending consult-ant); D. Eugene Strandness, Jr, MD, Uni-versity of Washington School of Medi-cine, Seattle, WA; Sidney K. Wolfson,MD, University of Pittsburgh School ofMedicine, Pittsburgh, PA; Bruce Zim-merman, MD, Mayo Clinic, Rochester,MN.

References1. Orchard TJ, Dorman JS, Maser RE,

Becker DJ, Drash AL, Ellis D, LaPorte RE,Kuller LH: The prevalence of complica-

tions in insulin-dependent-diabetes mel-litus by sex and duration: Pittsburgh Ep-idemiology of Diabetes ComplicationsStudy. II. Diabetes 39:1116-24, 1990

2. Melton LJ III, Macken KM, Palumbo PJ,Elveback LR: Incidence and prevalenceof clinical peripheral vascular disease in apopulation-based cohort of diabetic pa-tients. Diabetes Care. 3:650-54, 1980

3. Criqui MH, Langer RD, Fronek A, Fei-gelson HS, Klauber MR, McCann TJ,Browner D: Mortality over 10 years inpatients with peripheral arterial disease.N Engl J Med 326:381-86, 1992

4. Osmundson PJ, O'Fallon WM, ClementsIP, Kazmier FJ, Zimmerman BR,Palumbo PJ: Reproducibility of noninva-sive tests of peripheral occlusive arterialdisease. J Vase Surg 2:678-83, 1985

5. Osmundson PJ, O'Fallon WM, Zimmer-man BR, Kazmier FJ, Langworthy AL,Palumbo PJ: Course of peripheral occlu-sive arterial disease in diabetes: vascularlaboratory assessment. Diabetes Care 13:143-52, 1990

6. Palumbo PJ, O'Fallon WM, OsmundsonPJ, Zimmerman BR, Langworthy AL, Ka-zmier FJ: Progression of peripheral oc-clusive arterial disease in diabetes melli-tus: what factors are predictive? ArchIntern Med 151:717-21, 1991

7. Janka HU, Standl E, Mehnert H: Periph-eral vascular disease in diabetes mellitusand its relation to cardiovascular risk fac-tors: screening with the Doppler ultra-sonic technique. Diabetes Care 3:207-13,1980

8. Janka HU, Ziegler AG, Standl E, MehnertH: Daily insulin dose as a predictor ofmacrovascular disease in insulin-treatednon-insulin-dependent diabetics. Diabe-tes Metab 13:359-64, 1987

9. Criqui MH, Fronek A, Klauber MR, Bar-rett-Connor E, Gabriel S: The sensitivity,specificity, and predictive value of tradi-tional clinical evaluation of peripheral ar-terial disease: results from noninvasivetesting in a defined population. Patho-physiology and Natural History-PeripheralVascular Disease. 71:516-22, 1985

10. Strandness DE, Bell JW: Peripheral vas-cular disease: diagnosis and objectiveevaluation using a mercury strain gauge.Ann Surg 161:3-35, 1965

11. Holstein P, Lassen NA: Healing of ulcerson the feet correlated with distal bloodpressure measurements in occlusive arte-rial disease. Ada Orthop Scand 51:995-1006, 1980

12. Holstein P: The distal blood pressurepredicts healing of amputations on thefeet. Ada Orthop Scand 55:227-33, 1984

13. Maser RE, Wolfson SK, Stein EA, DrashAL, Becker DJ, Dorman JS, Ellis D, Or-chard TJ: Cardiovascular disease and ar-terial calcification in insulin-dependentdiabetes mellitus: interrelationships andrisk factor profiles. Pittsburgh Epidemi-ology of Diabetes Complications Study.V.ArteriosderThromb 11:958-65, 1991

14. Hiatt WR, Marshall JA, Baxter J, Sando-val R, Hildebrandt W, Kahn LR, Ham-man RF: Diagnostic methods for periph-eral arterial disease in the San Luis ValleyDiabetes study. J Clin Epidemiol 43:597-606, 1990

15. West KM, Ahuja MMS, Bennett PH,Czyzyk A, DeAcosta OM, Fuller JH, GrabB, Grabauskas V, Jarrett RJ, Kosaka K,Keen H, Krolewski AS, Miki E, SchliackV, Teuscher A, Watkins PJ, Stober JA:The role of circulating glucose and trig-lyceride concentrations and their interac-tions with other risk factors as determi-nants of arterial disease in nine diabeticpopulation samples from the WHO Mul-tinational Study. Diabetes Care 6:361-69, 1983

16. Janka HU: Five-year incidence of majormicrovascular complications in diabetesmellitus. Horm Metab Res Suppl 15:15-19, 1985

17. Fontbonne A, Eschwege E, Cambien F,Richard JL, Ducimetiere P, Thibult N,Warnet JM, Claude J, Rosselin GE: Hy-pertriglyceridemia as a risk factor of cor-onary heart disease mortality in subjectswith impaired glucose tolerance or dia-betes: results from the 11-year follow-upof the Paris Prospective Study. Diabeto-logia 32:300-304, 1989

18. The Expert Panel: National CholesterolEducation Program: report of the expertpanel on detection, evaluation, and treat-ment of high blood cholesterol in adults.Arch Intern Med 148:36-69, 1988

19. Consensus Statement: Role of cardiovas-cular risk factors in prevention and treat-

1208 DIABETES CARE, VOLUME 16, NUMBER 8, AUGUST 1993

Orchard and Strandness

ment of macrovascular disease in diabe-tes. Diabetes Care 12:573-79, 1989

20. Consensus Statement: Detection andManagement of Lipid Disorders in Dia-betes. Diabetes Care 16:106-12, 1993

21. Orchard TJ: Dyslipoproteinemia and di-abetes. In Lipid Disorders: Endocrinol

Metab Clin North Am LaRosa J, Ed. 19:

361-80, 199122. Garber A], Vinik Al, Crespin SR: Detec-

tion and management of lipid disordersin diabetic patients. Diabetes Care 15:1068, 1992

23. The Working Group on Hypertension inDiabetes: Statement on hypertension indiabetes mellitus. Arch Intern Med 147:830-42, 1987

24. Antiplatelet Trialists' Collaboration: Sec-ondary prevention of vascular disease byprolonged antiplatelet treatment. Br MedJ 296:320-31, 1988

25. ETDRS Investigators: Aspirin effects onmortality and morbidity in patients withdiabetes mellitus: early treatment dia-betic retinopathy study report 14. JAMA268:1292-1300, 1992

26. Consensus Statement: Report and rec-ommendations of the San Antonio con-ference on diabetic neuropathy. DiabetesCare 11:592-92, 1988

27. Proceedings of a Consensus Develop-ment Conference on Standardized Mea-sures in Diabetic Neuropathy. DiabetesCare 15:1080, 1992

28. National Diabetes Advisory Board: The

prevention and treatment of five compli-cations of diabetes. USDHHS/PHS/CDC,1983 (HHS83-8392)

29. Rose GA, Blackburn H: CardiovascularSurvey Methods. Geneva, World HealthOrg., 1968, p. 172-75 (monogr. no. 56)

30. Prineas RJ: Measurement of blood pres-sure in the obese. Ann Epidemiol 1:321-36, 1991

Other selected readings—Prineas RJ, Harland WR, Janzon L, Kannel

W: Recommendations for use of noninvasivemethods to detect atherosclerotic peripheralarterial disease: in population studies. Circu-lation 65:1561A-66A, 1982

Marinelli MR, Beach KW, Glass MJ, Pri-mozichJF, Strandness DE: Noninvasive test-ing versus clinical evaluation of arterial dis-ease. JAMA 241:2031-34, 1979

Kreines K, Johnson E, Albrink M, Knat-terud GL, Levin ME: The course of peripheralvascular disease in non-insulin-dependentdiabetes. Diabetes Care 8:235-43, 1985

Bild DE, SelbyJV, Sinnock P, Browner WS,Braveman P: Lower extremity amputation inpeople with diabetes epidemiology and pre-vention. Diabetes Care 12:24-31, 1989

McMillan DE: Monitoring the appearanceand progress of blood and vascular abnormal-ities. Clinical Diabetes Mellitus. 2nd ed. Dav-

idson JK, Ed. Leipzig, Germany, Thieme,1991, p. 414-26

Beach KW, Bedford GR, Bergelin RO, Mar-

tin DC, Strandness DE: Progression of lowerextremity arterial occlusive disease in type IIdiabetes mellitus. Diabetes Care 11:464-72,1988

Criqui MH: Peripheral arterial disease andsubsequent cardiovascular mortality: a strongand consistent association. Circulation 82:2246-47, 1990

Colwell JA: Antiplatelet drugs and preven-tion of macrovascular disease in diabetes mel-litus. Metabolism 41:7-10, 1992

Raines JK: Mechanics of air plethysmogra-phy in arterial disease: the pulse volume re-corder. In Noninvasive Diagnostic Techniques in

Vascular Disease. 3rd ed. Berstein EF, Ed. StLouis, MO, Mosby, 1985, p. 160-63

Rutherford RB, Lowenstein DH, Klein MF:Combining segmental systolic pressures andplethysmography to diagnose arterial occlu-sive disease of the legs. Am] Surg 138:211-17, 1979

Kram HB, Appel PI, Shomaker WC: Mul-tisensor transcutaneous oximetric mapping topredict below-knee amputation would heal-ing: use of a critical PO2. J Vase Surg 9:796-800, 1989

Cina C, Katsamouris A, MegermanJ, Brew-ster DC, Strayhom EC, Robison JG, AbbottWM: Utility of transcutaneous oxygen ten-sion measurements in peripheral arterial oc-clusive disease. J Vase Surg 1:362-71, 1984

Carter SA: Response of ankle systolic pres-sure to leg exercise in mild or questionablearterial disease. N Engl J Med 287:578-82,1972

DIABETES CARE, VOLUME 16, NUMBER 8, AUGUST 1993 1209