Effectiveness of the Diabetic Foot RiskClassification System of the InternationalWorking Group on the Diabetic FootEDGAR J.G. PETERS, MD1

LAWRENCE A. LAVERY, DPM, MPH2

OBJECTIVE To evaluate the effectiveness of a diabetic foot risk classification system by theInternational Working Group on the Diabetic Foot to predict clinical outcomes.

RESEARCH DESIGN AND METHODS A total of 225 diabetic patients were initiallyevaluated as part of a prospective case-control study at the University of Texas Health ScienceCenter at San Antonio. Complete records were available for 213 patients for follow-up evaluationafter 29 months. Upon enrollment, subjects were stratified into four risk groups based on thepresence of risk factors according to the consensus of the International Working Group on theDiabetic Foot. Group 0 consisted of subjects without neuropathy, group 1 consisted of patientswith neuropathy but without deformity or peripheral vascular disease (PVD), group 2 consistedof subjects with neuropathy and deformity or PVD, and group 3 consisted of patients with ahistory of foot ulceration or a lower-extremity amputation.

RESULTS Upon enrollment, patients in higher-risk groups had longer duration of diabetes,worse glycemic control, vascular and neuropathic variables, and more systemic complications ofdiabetes. During 3 years of follow-up, ulceration occurred in 5.1, 14.3, 18.8, and 55.8% of thepatients in groups 0, 1, 2, and 3, respectively (linear-by-linear association, P , 0.001). Allamputations were found in Groups 2 and 3 (3.1 and 20.9%, P , 0.001).

CONCLUSIONS The foot risk classification of the International Working Group on theDiabetic Foot predicts ulceration and amputation and can function as a tool to prevent lower-extremity complications of diabetes.

Diabetes Care 24:14421447, 2001

The purpose of this study was to eval-uate the effectiveness of a diabeticfoot risk classification established bythe International Working Group on theDiabetic Foot (1). There is some evidencethat diabetic foot complications can beprevented. One of the keys to population-based screening and disease managementis risk classification and stratification. Theobjective of the classification system de-veloped by the International WorkingGroup on the Diabetic Foot is to evaluatediabetic patients by using inexpensive

readily available instruments in order tostratify them into risk groups that wouldbe predictive of morbid outcomes. Thetools were intended to be inexpensive andreadily available to ensure worldwide im-plementation of the classification instru-ment. Resources such as therapeuticshoes, education, and clinical visits canthen be allocated to patients at greatestrisk of adverse events. In an internationalenvironment with an increasing preva-lence of diabetes and dwindling medicalresources, this type of tool can be used to

help shift current treatment models toprevention models.

RESEARCH DESIGN ANDMETHODS A total of 236 diabeticpatients were initially evaluated as part ofa case-control study at the University ofTexas Health Science Center at San Anto-nio in 1995 and 1996. After informedconsent was obtained, the subjects weresequentially enrolled from the foot clin-ics. Complete records were available for213 patients for follow-up evaluation af-ter a mean period of 30 months. As part ofstandard care in the foot specialty clinic,we evaluated patients with diabetes andprovided preventive care consisting ofregular podiatry evaluation based on riskcategory and referral to the pedorthist,vascular surgery department, and diabe-tes education or other services as re-quired. Patients with wounds receivedtreatment based on well-accepted proto-cols including wound debridement, dew-eighting the wound site with casts,removable walking boots or healingshoes, vascular testing, and lower-extremity revascularization and infectioncontrol as warranted (2).

Only data of subjects with follow-upwere evaluated. Ulcers were defined asskin lesions distal to the ankle. Subjectswho received an amputation as a directresult of their initial ulceration were dis-qualified from further analysis. These pa-tients had a much higher a priori risk ofamputation; inclusion may lead to selec-tion bias.

All subjects had diabetes according tothe criteria of the World Health Organi-zation (3). The type of diabetes was basedon the algorithm by Mogensen (4). Sev-eral variables previously reported to berisk factors for the development of dia-betic foot ulcers and lower-extremity am-putations were evaluated. These exposurevariables are also shown in Tables 1 and 2.

BMI (Quetelet index) was calculatedas weight in kilograms divided by thesquare of the height. Glycemic controlwas evaluated with HbA1c. Plantar peakpressures were measured with the Novel

c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c c

From the 1Vrije Universiteit, Amsterdam, the Netherlands, and the2Diabetex Foot Care Center, San Antonio,Texas.

Address correspondence and reprint requests to Lawrence A. Lavery, 9331 Fallworth, San Antonio, TX,78250. E-mail: lavery@texas.net.

Received for publication 23 January 2001 and accepted in revised form 17 April 2001.Abbreviations: ABI, ankle-brachial index; OR, odds ratio; PVD, peripheral vascular disease; SWM,

Semmes-Weinstein monofilament; VPT, vibration perception threshold.A table elsewhere in this issue shows conventional and Syste`me International (SI) units and conversion

factors for many substances.

P a t h o p h y s i o l o g y / C o m p l i c a t i o n sO R I G I N A L A R T I C L E

1442 DIABETES CARE, VOLUME 24, NUMBER 8, AUGUST 2001

EMED-SF platform system (Novel Elec-tronics, St. Paul, MN). An average of threesteps was used for analysis. Elevated plan-tar pressures were defined as a peak plan-tar pressure of .70 N/cm2 (5). Several

vascular parameters were evaluated. Thedata consisted of TcPO2 on the dorsal firstintermetatarsal space (TcPO2, ,40mmHg, probe temperature of 45C) (6),an ankle-brachial index (ABI) of ,0.8

(7), absence of dorsal pedal and posteriortibial artery pulsations, and incompress-ibility of pedal arteries (ABI .1.2). Pe-ripheral vascular disease (PVD) wasdefined as an ABI of ,0.8 or any non-

Table 1Distribution of descriptive characteristics and demographics of continuous variables among the four groups of consensus classifi-cation

Group

Total P0

(no neuropathy)1

(neuropathy)

3(neuropathy and/or

vascular diseaseand/or deformity)

4(previous ulcer)

n 79 21 51 62 213Age (years) 49.8 6 10.3 54.1 6 7.5 55.3 6 10.8 53.4 6 10.2 52.6 6 10.4 0.017Diabetes duration (years) 7.7 6 6.7 10.9 6 8.3 13.2 6 12.7 13.5 6 8.0 11.0 6 9.3 0.001Follow-up (months) 29.1 6 6.6 30.6 6 4.5 29.7 6 5.9 29.3 6 7.2 29.4 6 6.4 NSBMI (kg/m2) 31.4 6 5.6 32.1 6 7.2 33.1 6 7.2 30.4 6 5.7 31.6 6 6.2 NSHbA1c (%) 8.7 6 2.0 8.8 6 2.0 8.6 6 1.8 9.7 6 2.3 9.0 6 2.1 0.020Plantar peak pressure (N/cm2) 62.3 6 20.9 65.1 6 27.0 63.9 6 20.7 83.1 6 25.7 68.2 6 24.2 ,0.001ABI 1.2 6 0.3 1.2 6 0.2 1.1 6 0.4 1.33 6 0.42 1.23 6 0.35 0.036VPT (V) 10.2 6 4.1 24.3 6 13.6 20.4 6 11.9 37.0 6 13.2 21.7 6 15.0 ,0.001TcPO2 40.5 6 13.3 40.0 6 10.5 38.2 6 14.6 40.3 6 19.5 39.8 6 15.2 NS

Data are means 6 SD.

Table 2Distribution of descriptive characteristics and demographics of dichotomous variables among the four groups of the consensusclassification

Group 0(no neu-ropathy)

OR(groups

13)

Group 1(neurop-

athy)

OR(groups

13)

Group 2(neuropathy,

vascular diseaseand/or deformity

OR(groups

23)

Group 3(previous

ulcer) Total P

Female 70.9 7.0 (3.314.7) 85.7 17.2 (4.566.7) 47.1 2.6 (1.25.6) 25.8 53.5 ,0.001Type 2 diabetes 94.2 NS 100 1.1 (1.01.1) 94.1 NS 93.4 93.8 NSPrevious or present drinker 45.6 2.3 (1.14.5) 47.6 NS 60.8 NS 65.6 55.2 0.011Present or previous tobacco

use55.7 NS 57.1 NS 52.9 NS 60.7 56.6 NS

Nephropathy 25.3 3.1 (1.56.4) 33.3 NS 27.5 2.8 (1.36.2) 51.6 34.3 0.004Retinopathy 33.8 8.9 (3.721.5) 50.0 4.6 (1.414.7) 44.4 5.7 (2.214.4) 82.0 51.4 ,0.001HbA1c .9% 42.7 2.5 (1.25.0) 57.9 NS 40.9 2.7 (1.26.0) 65.0 50.5 0.033Amputation history 0 2.8 (2.23.6) 0 1.4 (1.21.7) 0 2.2 (1.82.7) 30.6 8.9 ,0.001Elevated plantar peak

pressures33.3 4.4 (2.09.6) 47.1 NS 36.2 3.9 (1.79.1) 68.8 44.4 0.001

ABI ,0.8 0 2.7 (2.03.6) 0 NS 21.2 NS 10.3 8.5 0.015Any missing pulse 0 3.0 (2.33.9) 0 1.6 (1.31.9) 52.9 NS 36.1 23.1 ,0.001Incompressible ankle

pressure (.1.2)40.5 NS 42.1 NS 34.0 NS 47.3 43.0 NS

TcPO2 ,40 mmHg 47.2 NS 60.0 NS 62.5 NS 48.1 52.6 NSVPT .25 V 0 8.9 (5.015.9) 42.9 6.7 (2.220.0) 27.5 13.2 (5.333.0) 83.3 34.6 ,0.001Abnormal SWM test 0 62.5 (9.2500) 89.5 NS 67.4 15.0 (1.8121.1) 96.9 49.0 ,0.001Deformity 36.7 3.3 (1.76.8) 0 2.0 (1.52.7) 68.6 NS 66.1 49.3 ,0.001Hallux limitus 5.1 7.1 (2.422.4) 9.5 NS 11.8 2.8 (1.07.8) 27.4 13.6 ,0.001Pes equinus 17.7 NS 14.3 NS 29.4 NS 29.0 23.5 NSLimited ROM subtalar joint 29.1 2.4 (1.24.9) 14.3 6.0 (1.622.5) 14.3 NS 50 36.2 0.006

Data are % or OR (95% CI) unless otherwise indicated. For VPT, abnormal SWM test, deformity, and hallux limitus, P , 0.001; for limited ROM subtalar joint,P 5 0.006. ROM, range of motion.

Peters and Lavery

DIABETES CARE, VOLUME 24, NUMBER 8, AUGUST 2001 1443

palpable pedal pulsation. Peripheralneurological deficits were assessed withvibration perception threshold (VPT) andSemmes-Weinste in monofilament(SWM). The VPT was measured with aBiothesiometer (Biomedical Instrument,Newbury, OH) at the tip of the hallux (8).A voltage of .25 V was defined as a loss ofprotective sensation (9). The 10-g SWMwas applied to 10 areas of the foot. Im-paired sensation was defined as one ormore unnoticed pinpricks with the SWMor a VPT of .25 V (10).

Joint mobility of the first metatarsalphalangeal joint, the subtalar joint, andthe ankle was assessed by averaging threemeasurements. Limited joint mobility ofthe forefoot was determined by the pres-ence of hallux limitus, which is dorsiflex-ion of the hallux of ,50 degrees (1012).Limited range of motion of the subtalarjoint was determined by ,20 degrees to-tal range of motion, and pes equinus wasdefined as ,0 degrees of dorsiflexion atthe ankle (10). Deformity of the forefootwas defined as hallux valgus, rigid toecontractures (such as hammer or clawtoes), and prominent metatarsal heads(10).

All subjects completed a CAGE (cut-down, annoyed, guilty, and eye-opener)questionnaire. Alcohol abuse was definedas a positive CAGE ($3) (13). Previousalcohol abuse was also noted if it wasstated in the medical record. Nephropa-thy was defined as creatinine .4.0 mg/dl,current renal dialysis, and history of renaltransplantation (14,15). Retinopathy wasdefined as presence of at least backgroundchanges. The retinal pictures had beentaken as part of the routine diabetes careof the subjects (Cannon CR6-45NM Non-

Mydratic Retinal Camera; Cannon, LakeSuccess, NY).

During an average follow-up of 29months, several outcomes were evalu-ated. These included occurrence of an ul-cer or amputation, the level of such anamputation, reamputations of such anamputation, and necessity for a periph-eral arterial bypass.

For continuous exposure variables, aone-way analysis of variation was used todetect any statistical differences amongthe groups. This test is used to test thehypothesis that several means are equaland function as an extension to the Stu-dents t test. For dichotomous exposurevariables and for detection of trendsamong the four and five groups of theclassification and of the alternative classi-fication, respectively, a linear-by-linearassociation (Mantel-Haenszel x2 test fortrend) was calculated to evaluate trendsfrom group 0 through group 3. This testcalculates the linear association betweenrow and column variables in a cross-tabulation. Odds ratios (ORs) of the datain groups 0, 1, and 2 were calculated in-dependently from one another to the datain group 3 in a cross-tabulation. To eval-uate the incidence of morbidity in thegroup of patients with a history of ulcer-ation, the data of groups 0, 1, and 2 werecombined and compared with the data ofthe patients in group 3 with a x2 test.

After this, the same procedure wasfollowed for two subgroups within group3. Group 3 was split into a group of pa-tients without a history of amputation(group 3a) and one with a history of am-putation (group 3b), thereby resulting ina total of five groups. P , 0.05 was con-sidered statistically significant. All statis-

tical analyses were performed with SPSS8.0 software.

RESULTS Of 236 initially enrolledpatients, follow-up was available for 213patients. As shown in Tables 13 , it ap-pears that initial measurements for mostof the factors assessed in the study aremore often of a more severe nature in thehigher-risk groups. The data of groups 0,1, and 2 were combined and comparedwith the data of the patients in group 3with a x2 test. Patients with a history ofulceration were 2.8 times (95% CI 1.55.3) more likely to have nephropathy, 6.9times (3.115.4) more likely to have ret-inopathy, and 3.9 times (1.97.9) morelikely to have elevated plantar peak pres-sures than patients in groups 0, 1, and 2.

Table 3 displays the incidence of ul-cerations and amputations in the 3-yearfollow-up period. There were signifi-cantly more ulcerations and amputations(P , 0.001) as well as an increased pro-portion of proximal amputations in thehigher-risk groups (P , 0.001). Patientsin the highest-risk group were 34.1 times(95% CI 11.0105.8) more likely to de-velop an ulceration in the follow-up pe-riod.

x2 Tests showed that patients with ahistory of an ulceration were 17.8 times(95% CI 8.337.9) more likely to developan ulceration within the 3-year follow-upperiod than patients without an ulcer his-tory (groups 0, 1, and 2 combined). Like-wise, they were 52.2 times (6.7404.1)more likely to receive a lower-extremityamputation and 13.2 times (1.5115.1)more likely to undergo a peripheral arte-rial bypass.

Table 4 shows the result of the variant

Table 3Distribution of adverse outcomes after 3 years among the four groups of the consensus classification

Group 0(no neu-ropathy)

OR(groups

13)

Group 1(neurop-

athy)

OR(groups

13)

Group 2(neuropathy,

vascular diseaseand/or deformity

OR(groups

23)

Group 3(previous

ulcer) Total P

Follow-up ulcer 4 (5.1) 34.1 (11.0105.8) 3 (14.3) 10.9 (2.941.2) 7 (13.3) 11.4 (4.429.6) 40 (64.5) 54 (25.4) ,0.001Follow-up amputation 0 2.7 (2.23.4) 0 1.4 (1.21.6) 1 (2.0) 17.4 (2.2136.4) 16 (25.8) 17 (8.0) ,0.001Toe/ray 0 0 1 (2.0) 7 (11.3) 8 (3.8) ,0.001Transmetatarsal 0 0 0 5 (8.1) 5 (2.3)Transtibial or higher 0 0 0 4 (6.5) 4 (1.9)Reamputations 0 2.4 (2.03.0) 0 1.4 (1.21.6) 0 1.9 (1.62.3) 7 (43.8) 7 (41.1) 0.001Follow-up bypass 0 2.4 (2.02.9) 0 1.4 (1.21.6) 1 (2.0) NS 5 (8.1) 6 (2.8) 0.006

Data are n (%) and OR (95% CI) unless otherwise indicated. For follow-up ulcer, follow-up amputation, and for all amputations, P , 0.001; for reamputations, P 50.001; and for follow-up bypass, P 5 0.006.

Effectiveness of foot risk classification

1444 DIABETES CARE, VOLUME 24, NUMBER 8, AUGUST 2001

of the consensus classification, in whichthe group of patients with a history of ul-ceration was split into a group of patientswith and without a history of amputation.It appeared that the higher-risk groupshad a significantly higher number ofulcerations, amputations, and a higherproportion of proximal amputations thanpatients in the lower-risk groups (,0.001).Patients in the highest-risk group (group3b) were 100 times (95% CI 20.4491.0)more likely to ulcerate than patients in thelowest-risk group (group 0).

CONCLUSIONS This study evalu-ates the effectiveness of a classificationsystem to predict diabetic foot compli-cations. Our data suggests that the clas-sification system proposed by theInternational Working Group is effectivein predicting groups that are more likelyto develop diabetes-related foot compli-cations. There was a clear trend for in-creased morbid events in progressivestages of the classification scheme. Ourresults show that patients in a higher-riskgroup are 34 times more likely to ulceratethan patients in the lowest-risk group.Likewise, high-risk patients are a littleover 17 times more likely to receive anamputation in a 3-year follow-up than pa-tients in a lower-risk group. Patients witha history of amputation are even morelikely to develop a foot complication.These patients are 100 times more likelyto ulcerate and 32 times more likely toreceive an additional amputation than pa-tients in the lower-risk groups (Table 4).Overall, patients in higher-risk groupshad a longer history of diabetes, worseglycemic control, more neuropathy, andincreased plantar pressures. Further-more, they were more likely to be maleand to have a history of alcohol abuse andend-organ complications of diabetes,such as nephropathy and retinopathy.The stratification process easily explainsthese differences among the risk groupsat enrollment of the patients; there is anincreased prevalence of angiopathy, neu-ropathy, deformity, and earlier amputa-tions in the higher risk groups. It seemsonly logical that these conditions aremore often seen in certain groups of pa-tients, such as men with a longer historyof diabetes and worse glucose level con-trol. It has been demonstrated that footdeformities and other ulcer risk factorsare indeed exacerbated by poor glucose

control and longer duration of diabetes(16).

The goal of the study was not to assessthe responsible cause or cluster of causesthat lead to an ulceration or amputationbut rather to determine if a classificationsystem would predict an ulceration oramputation. By themselves, the risk fac-tors have been shown to be possiblecausal factors in ulceration and amputa-tion in diverse case-control studies. Neu-ropathy has been shown to be a pivotalrisk factor by itself for both ulceration andamputation (6,10,12,17,18). Vasculardisease has been strongly associated withamputations (19,20); its role as a risk fac-tor for ulceration seems likely, but severalstudies provided mixed results. An ABI of,0.8, for instance, was not identified as asignificant risk factor in previous studies(6,21). It has been suggested that vasculardisease might be a more important riskfactor for delayed wound healing andsubsequent amputation than for ulcer-ation (10). Deformities of the foot mightcontribute to ulceration because they cancreate areas of increased pressure on thefoot. Likewise, amputations have beenshown to increase the risk of additionalulceration (22).

Several authors have proposed classi-fication systems in the past. Most of themincorporated neuropathy, bony foot de-formity, history of ulceration or amputa-tion, or various combinations of these(10,2329). Three of them also includedPVD (10,28,29). Although most earlyclassification systems were solely basedon expert clinicians experiences, three ofthese systems were validated with patientdata (10,26,29). Mayfield et al. (29) vali-dated their classification for amputations,Lavery et al. (10) validated their classifi-cation for ulcerations, and Rith-Najarianet al. (26) validated their classification forboth ulcers and amputations. The con-sensus classification differs from earlierclassification systems because cliniciansand researchers from various parts of theworld and from various fields of workwere involved in its conceptualization.Some of the authors of other classificationsystems were also members of the Inter-national Working Group on the DiabeticFoot (1).

One of the classifications, which re-sembles the international consensus, isthe University of Texas Treatment-BasedDiabetic Foot Classification System(10,30,31). It consists of two differentT

able

4D

istr

ibut

ion

ofad

vers

eou

tcom

esaf

ter

3ye

ars

amon

gth

egr

oups

ofth

eco

nsen

sus

clas

sifi

cati

on

Gro

up0

(no

neu-

ropa

thy)

OR

(gro

ups

13b

)

Gro

up1

(neu

rop-

athy

)

OR

(gro

ups

13b

)

Gro

up2

(neu

ropa

thy,

PVD

and/

orde

form

ity)

OR

(gro

ups

23b

)

Gro

up3a

(pre

viou

sul

cer,

noam

puta

tion

)

OR

(gro

ups

3a3

b)

Gro

up3b

(pre

viou

sam

puta

tion

)T

otal

P

n79

2151

4319

213

Follo

w-u

pul

cer

4(5

.1)

100.

0(2

0.4

491.

0)3

(14.

3)32

.0(5

.61

81.6

)7

(13.

7)33

.5(7

.71

45.6

)24

(55.

8)4.

2(1

.11

6.7)

16(8

4.2)

54(2

5.4)

,0.

001

Follo

w-u

pam

puta

tion

07.

6(4

.51

2.8)

01.

6(1

.12

.2)

1(2

.0)

29.2

(3.3

260

.1)

9(2

0.9)

NS

7(3

6.8)

17(1

8.1)

,0.

001

Toe

/ray

0

0

1(2

.0)

6

(14.

0)

1(5

.3)

8(3

.8)

Mid

foot

/TM

A0

0

0

2

(4.7

)

3(1

5.8)

5(2

.3)

,0.

001

BKA

orA

KA

0

0

0

1(2

.3)

3

(15.

8)4

(1.9

)R

eam

puta

tion

s0

0

0

4.2

(2.7

6.4

)4

(9.3

)N

S3

(15.

8)7

(41.

1),

0.00

1Fo

llow

-up

bypa

ss0

6.0

(3.8

9.3

)0

2.3

(1.6

3.3

)1

(2.0

)N

S2

(4.7

)N

S3

(15.

8)6

(2.8

)0.

001

Dat

aar

en

(%)

orO

R(9

5%C

I)un

less

othe

rwis

ein

dica

ted.

For

follo

w-u

pul

cer,

follo

w-u

pam

puta

tion

,toe

/ray

,mid

foot

/tra

nsm

etat

arsa

lam

puta

tion

(TM

A),

belo

w-t

he-k

nee

ampu

tati

on(B

KA

),or

abov

e-th

e-kn

eeam

puta

tion

(AK

A),

and

ream

puta

tion

s,P

,0

.00

1;f

orfo

llow

-up

bypa

ss,P

50

.00

1.

Peters and Lavery

DIABETES CARE, VOLUME 24, NUMBER 8, AUGUST 2001 1445

classifications. The first one stratifies pa-tients into risk groups for ulceration, andthe second one stratifies patients with anexisting ulceration into risk groups foramputation (31). Although it is very log-ical and the first part for ulcerations hasbeen validated, it is not the simplest ofclassifications. Incorporating this classifi-cation scheme in diabetes clinics that arenot specialized in problems of the diabeticfoot might be difficult and impractical.This is especially true because there arewide intercountry variations in healthcare resources and infrastructure. Even inthe Western world, which is known forreadily available medical care, the qualityof diabetic foot examinations is oftenlacking (32,33). The complexity of mostof the classification systems might makethem hard to implement globally. Thegoal of the consensus was to achieveglobal consistency in adequate diagnos-tic, preventative, and therapeutic strate-gies. These strategies use simplediagnostic tools to help prevent diabeticfoot problems from occurring. Therefore,a shift might take place from therapeuticmedical and surgical treatment to preven-tative measures.

The International Working Group onthe Diabetic Foot did not specifically de-scribe the default methods to measureneuropathy or angiopathy. However, theinstruments we used to define thesemethods are described in the Consensus(1). VPTs, SWMs, ABIs, and pedal pulsesare widely accepted clinical measures (810,34,35). All of these techniques pro-vide easily obtainable data. Only the VPT,measurable with a biothesiometer or neu-roesthesiometer, could be expensive forsome centers. It was used in this studybecause of its convenience and because allof the researchers involved were familiarwith its use. As an alternative, a 128-Hztuning fork could be used. The data col-lected with a tuning fork has been shownto correlate well with the data of the VPT(36).

The recommended check-up fre-quency for the group without neuropathywas once a year, for the group with neu-ropathy but without deformity or vascu-lar disease wasonce every half year, for thegroup with neuropathy, deformity,and/or vascular disease was once every 3months, and for the group of patientswith a history of ulceration was once ev-ery 13 months (1). In general, patients inthis study were not strictly seen according

to the frequency recommended by theconsensus. At the University of Texas, pa-tients are often seen more frequently thanthe guidelines in the consensus. There-fore, the chance that diabetic foot ulcersoccurred and healed outside of a patientsappointment schedule is small.

Even though several descriptive stud-ies have reported the effectiveness of dia-betic foot prevention programs, we stilldo not regularly inspect the foot or pre-scribe therapeutic shoes and insoles toprotect high-risk patients from repetitiveinjuries. The International WorkingGroups Classification provides a simpleand effective globally implementableframework for assessment and treatment.It can easily be used as a checklist in med-ical records to help health professionalsallocate resources, such as therapeuticshoes and insoles, to determine the fre-quency of follow-up visits, and to docu-ment that foot screening is beingperformed on a regular basis.

References1. Apelqvist J, Bakker K, Van Houtum WH,

Nabuurs-Fransen MH, Schaper NC: In-ternational Consensus on the DiabeticFoot. In The International Working Groupon the Diabetic Foot. Amsterdam, TheNetherlands, John Wiley & Sons, 1999,p. 67

2. Armstrong DG, Harkless LB: Outcomes ofpreventative care in a diabetic foot spe-cialty clinic. J Foot Ankle Surg 37:460466, 1998

3. World Health Organization: WHO ExpertCommittee on Diabetes Mellitus. Second Re-port. Geneva, World Health Org., 1980(641-680)

4. Mogensen CE: Renoproductive role ofACE inhibitors in diabetic nephropathy.Br Heart J 72 (Suppl. 3):S38S45, 1994

5. Armstrong DG, Peters EJ, Athanasiou KA,Lavery LA: Is there a critical level of plan-tar foot pressure to identify patients at riskfor neuropathic foot ulceration? J Foot An-kle Surg 37:303307, 1998

6. McNeely MJ, Boyko EJ, Ahroni JE, StenselVL, Reiber GE, Smith DG, Pecoraro RE:The independent contributions of dia-betic neuropathy and vasculopathy infoot ulceration: how great are the risks?Diabetes Care 18:216219, 1995

7. Beach KW, Strandness DE: Arteriosclero-sis obliterans and associated risk factorsin insulin-dependent and non-insulin-de-pendent diabetes. Diabetes 29:882888,1980

8. Young MJ, Breddy JL, Veves A, BoultonAJM: The prediction of diabetic neuro-

pathic foot ulceration using vibration per-ception thresholds: a prospective study.Diabetes Care 17:557560, 1994

9. Armstrong DG, Lavery LA, Vela SA, Que-bedeaux TL, Fleischli JG: Choosing a prac-tical screening instrument to identifypatients at risk for diabetic foot ulcer-ation. Arch Int Med 158:289292, 1998

10. Lavery LA, Armstrong DG, Vela SA,Quebedeaux TL, Fleischli JG: Practicalcriteria for screening patients at high riskfor diabetic foot ulceration. Arch InternMed 158:158162, 1998

11. Birke JA, Cornall MA, Jackson M: Rela-tionship between hallux limitus and ul-ceration of the great toe. Sports Phys TherJ Orthop 10:172176, 1988

12. Birke JA, Franks D, Foto JG: First ray jointlimitation, pressure, and ulceration of thefirst metatarsal head in diabetes mellitus.Foot Ankle 16:277284, 1995

13. Magruder-Habib K, Stevens HA, AllingWC: Relative performance of the MAST,VAST, and CAGE versus DM-III-R criteriafor alcohol dependence. J Clin Epidemiol46:435441, 1993

14. Savage S, Estacio RO, Jeffers B, SchrierRW: Urinary albumin excretion as apredictor of diabetic retinopathy, neu-ropathy, and cardiovascular disease inNIDDM. Diabetes Care 19:12431248,1996

15. Humphrey LL, Ballard DJ, FrohnertPP, Chu CP, OFallon M, Palumbo PJ:Chronic renal failure in non-insulin-de-pendent diabetes mellitus: a population-based study in Rochester, Minnesota. AnnIntern Med 111:788796, 1989

16. Strowig S, Raskin P: Glycemic control anddiabetic complications. Diabetes Care 15:11261140, 1992

17. Kumar S, Fernando DJS, Veves A, KnowlesEA, Young MJ, Boulton AJM: Semmes-Weinstein monofilaments: a simple, ef-fective and inexpensive screening devicefor identifying diabetic patients at risk offoot ulceration. Diabetes Res Clin Practi 13:6368, 1991

18. Mueller MJ: Identifying patients with dia-betes who are at risk for lower extremitycomplications: use of Semmes-Weinsteinmonofilaments. Phys Ther 76:6871, 1996

19. Pecoraro RE, Reiber GE, Burgess EM:Pathways to diabetic limb amputation:basis for prevention. Diabetes Care 13:513521, 1990

20. Reiber GE, Pecoraro RE, Koepsell TD:Risk factors for amputation in patientswith diabetes mellitus: a case controlstudy. Ann Intern Med 117:97105, 1992

21. Sriussadaporn S, Mekanandha P, Vanna-saeng S, Nitiyanant W, Kolmoltri C,Ploybutr S, Yamwong P, Peerapatdit T,Vichayanrat A: Factors associated with di-abetic foot ulceration in Thailand: a case-control study. Diabet Med 14:5056, 1997

Effectiveness of foot risk classification

1446 DIABETES CARE, VOLUME 24, NUMBER 8, AUGUST 2001

22. Lavery LA, Lavery DC, Quebedeaux TL:Increased foot pressures after great toeamputation in diabetes. Diabetes Care 18:14601462, 1995

23. Birke JA, Sims DS: The insensitive foot. InPhysical Therapy of the Foot and Ankle.Hunt GC, Ed. New York, Churchill Liv-ingstone, 1988, p. 133168

24. Sims DS, Cavanaugh PR, Ulbrecht JS: Riskfactors in the diabetic foot: recognitionand management. J Am Phys Ther Assoc68:18871902, 1988

25. Duffy JC, Patout CA: Management ofthe insensitive foot in diabetes: lessonslearned from Hansens disease. Milit Med155:575579, 1990

26. Rith-Najarian SJ, Stolusky T, Gohdes DM:Identifying diabetic patients at high riskfor lower extremity amputation in a pri-mary health care setting: a prospectiveevaluation of simple screening criteria.Diabetes Care 15:13861389, 1992

27. Brill LR, Cavanaugh PR, Doucette MM,

Ulbrecht JS: Prevention, Protection and Re-currence Reduction of Diabetic NeuropathicFoot Ulcers. East Setauket, New York,Curative Technologies, 1994

28. Plummer ES, Albert SG: Foot care assess-ment in patients with diabetes: a screen-ing algorithm for patient education andreferral. Diabetes Educ 21:4751, 1995

29. Mayfield JA, Reiber GE, Nelson RG,Greene T: A foot risk classification systemto predict diabetic amputation in Pima In-dians. Diabetes Care 19:704709, 1996

30. Armstrong DG, Lavery LA: Diabetic footulcers: prevention, diagnosis and classifi-cation. Am Fam Physician 57:13251340,1998

31. Armstrong DG, Lavery LA, Harkless LB:Treatment-based classification system forassessment and care of diabetic feet. J AmPodiatr Med Assn 86:311316, 1996

32. Edelson GW, Armstrong DG, Lavery LA,Caicco G: The acutely infected diabeticfoot is not adequately evaluated in an in-

patient setting. Arch Intern Med 156:23732376, 1996

33. Rosset J, Walker EA, Shamoon H, Engel S,Basch C, Zybert P: Assessment of docu-mented foot examinations for patientswith diabetes in inner-city primary careclinics. Arch Fam Med 4:4650, 1990

34. Apelqvist J, Larsson J, Agardh CD: Theimportance of peripheral pulses, periph-eral oedema and local pain for the out-come of diabetic foot ulcers. Diabet Med7:590594, 1990

35. Boyko EJ, Ahroni JH, Stensel V, ForsbergRC, Davignon DR, Smith DG: A prospec-tive study of risk factors for diabetic footulcer: the Seattle Diabetic Foot Study Di-abetes Care 22:10361042, 1999

36. Hilz MJ, Axelrod FB, Hermann K, HaertlU, Duetsch M, Neundorfer B: Norma-tive values of vibratory perception in530 children, juveniles and adults aged379 years. J Neurol Sci 159:219 225,1998

Peters and Lavery

DIABETES CARE, VOLUME 24, NUMBER 8, AUGUST 2001 1447