1 - diabetes care

N A L A R T I C L E

A Practical Two-StepQuantitative Clinical andElectrophysiologicalAssessment for theDiagnosis and Staging ofDianetic NeuropathyEVA L. FELDMAN, MD, PHD

M. J. STEVENS, MBBCH

P. K. THOMAS, CBE, MD

M. B. BROWN, PHD

N. CANAL, MD

D. A. GREENE, MD

OBJECTIVE — Early diagnosis of distal symmetric sensorimotor polyneuropathy, acommon complication of diabetes, may decrease patient morbidity by allowing forpotential therapeutic interventions. We have designed an outpatient program to facil-itate diagnosis of diabetic neuropathy.

RESEARCH DESIGN AND METHODS— Patients are initially administereda brief questionnaire and screening examination, designated the Michigan NeuropathyScreening Instrument (MNSI). Diabetic neuropathy is confirmed in patients with apositive assessment by a quantitative neurological examination coupled with nerveconduction studies, designated the Michigan Diabetic Neuropathy Score (MDNS). Inthis study, 56 outpatients with confirmed type I or II diabetes were administered thestandardized quantitative components required to diagnose and stage diabetic neurop-athy according to the San Antonio Consensus Statement (1) and the Mayo Clinicprotocol (2). These same patients were then assessed with the MNSI and the MDNS.

RESULTS— Of 29 patients with a clinical MNSI score >2, 28 had neuropathy.Twenty-eight patients with an MDNS of ^ 7 had neuropathy, while 21 non-neuropathic patients had a score ^ 6 . Of 35 patients with diabetic neuropathy, 34 had^ 2 abnormal nerve conductions. Twenty-one normal patients and one patient withneuropathy had ^ 1 abnormal nerve conduction.

CONCLUSIONS — The results indicate that the MNSI is a good screening tool fordiabetic neuropathy and that the MDNS coupled with nerve conductions provides asimple means to confirm this diagnosis.

From the Departments of Neurology (E.L.F.), Internal Medicine (M.J.S., D.A.G.), and Biostatistics(M.B.B.) and the Michigan Diabetes Research and Training Center (E.L.F., M.J.S., M.B.B., D.A.G.),University of Michigan, Ann Arbor, Michigan; the Department of Neurology (P.K.T.), Royal FreeHospital, London, U.K.; and the Department of Neurology (N.C.), San Raffaele Hospital, Milan, Italy.

Address correspondence and reprint requests to Eva L. Feldman, MD, PhD, Department of Neu-rology, University of Michigan, 200 Zina Pitcher Place, Room 4414 Kresge III, Ann Arbor, MI 48109-0588.

Received for publication 12 August 1993 and accepted in revised form 17 June 1994.DN, distal symmetric sensorimotor polyneuropathy; MNSI, Michigan Neuropathy Screening In-

strument; MDNS, Michigan Diabetic Neuropathy Score; NDS, Neuropathy Disability Score; NSP,Neuropathy Symptom Profile; sBP, systolic blood pressure.

D istal symmetric sensorimotor poly-neuropathy (DN) is a commoncomplication of diabetes (3). In a

prospective study of >4,400 outpatientsin a diabetes clinic, ~10% of patients hadneuropathy at the time of diagnosis oftheir diabetes. The incidence of neuropa-thy increased with the duration of diabe-tes, and >50% of patients had DN after25 years of diabetes (4). Variations in theprevalence of DN have historically re-flected the different criteria used to fonnits diagnosis. In a cohort of 278 healthytype I diabetic patients entered into theDiabetes Control and ComplicationsTrial, using criteria that included abnor-malities on a physical exam plus citherabnormal nerve conduction studies intwo nerves or abnormal autonomic func-tion testing, 39% had some clinical man-ifestation of DN (5). In a cross-sectionalmulticenter study of 6,487 diabetic pa-tients in the U.K., the prevalence of DNincreased with age and reached 44.2% inindividuals between 70 and 79 years ofage (6). Similarly, in a population-basedstudy, 66% of patients with diabetes had adecline in vibratory sensation in the feeton repeated testing over time (7). Thishigh prevalence of DN is associated withsignificant morbidity, including recurrentfoot ulcerations, infections, and subse-quent amputations. More hospital bedsare occupied by diabetic patients withfoot problems than with all other remain-ing diabetic complications (8), and, in theU.S., 50-70% of all nontraumatic lower-extremity amputations are secondary todiabetes (9).

DN is frequently insidious in on-set and may result in sensory loss, pre-disposing to callus formation that canfissure, become infected, and may precip-itate plantar ulceration (10). Early diag-nosis of DN can decrease patient morbid-ity by allowing for potential therapeuticinterventions, including patient educa-tion and regular foot surveillance. How-ever, there is no single diagnostic test forthe detection of DN. Historically, the di-agnosis of DN was based on subjective

DIABETES CARE, VOLUME 17, NUMBER 11, NOVEMBER 1994 1281

Dow

nloaded from http://diabetesjournals.org/care/article-pdf/17/11/1281/515039/17-11-1281.pdf by guest on 03 February 2022

Practical assessment for diabetic neuropathy

interpretation of neuropathic symptomsand specific signs, such as reduced anklejerks and loss of vibratory sensation (11).In 1985, Dyck et al. (12) at the MayoClinic in Rochester, Minnesota, intro-duced the first comprehensive set of diag-nostic criteria for DN. In 1988, the SanAntonio Consensus Statement by theAmerican Diabetes Association (13) sug-gested that, to fully classify diabetic neu-ropathy, a patient have at least one mea-surement among the following tests:clinical symptoms, clinical examination,electrodiagnostic studies, quantitativesensory testing, and autonomic functiontesting. Patients were classified as eitherstage I (absence of symptoms) or II (pres-ence of symptoms), and within each stagewere graded "A" through "C" dependingon the number of abnormal physiologicaltests and a symptoms score. The need toadapt or simplify these criteria for epide-miological screening was recognized. Thesecond proposal, developed by the MayoClinic group (2), used the same electro-physiological and symptoms measure-ments, but placed patients in one of threestages (1, 2, or 3). Like the San Antoniocriteria, each stage was defined by thenumber of abnormal physiological testsand a symptom score. These tests not onlydiagnose diabetic neuropathy but alsodistinguish it from other forms of neurop-athy. They require experienced personnelusing specialized equipment that may notbe widely available. The San AntonioConsensus and the Mayo Clinic proposalswere intended for research purposes butare less practical in many routine clinicalsettings where there is a need for a simpleassessment using widely available tech-niques to rapidly screen large numbers ofpatients.

We have designed a simple two-step screening and staging program forclinically significant DN to use in out-patient clinics. The first step, the Michi-gan Neuropathy Screening Instrument(MNSI), is designed to screen large num-bers of patients in a routine clinical set-ting for the presence of DN. The MNSIcan be administered by an endocrinolo-

gist, internist, general practitioner, nursepractitioner, or other health care profes-sionals involved in the treatment of dia-betic patients. Patients who screen posi-tive on the MNSI may be referred foradministration of the Michigan DiabeticNeuropathy Score (MDNS). The MDNS,which consists of a quantitative neurolog-ical examination and a set of nerve con-duction studies, provides a means of di-agnosing and staging DN that is simplerand less time consuming than presentlyaccepted research protocols. The clinicalcomponent of the MDNS can be adminis-tered by any practicing physician, but thenerve conduction component requires aneurologist or a physician skilled in elec-trophysiology.

In this study, 56 patients from anoutpatient diabetes clinic were adminis-tered the MNSI, the MDNS, and all thecomponents required by the Mayo Clinicstaging process, which includes the as-sessment components outlined in the SanAntonio criteria. Numerical scores on theMNSI and MDNS were compared in eachpatient with the stage of DN according tothe Mayo Clinic classification. Results in-dicate that the MNSI is a good screeningtool for the detection of DN and that theMDNS provides a quick, sensitive meansto confirm the diagnosis and potentiallystage DN.

RESEARCH DESIGN ANDMETHODS — Thirty-one men and 25women were recruited from consecutivepresentations to the University of Michi-gan Diabetes and Research Training Cen-ter general diabetic outpatient clinics.These patients met the American DiabetesAssociation criteria for either type I ortype II diabetes. The average age of thepatients was 47.5 ± 13 years with an agerange of 24-65 years. Twenty-seven pa-tients had type I and 29 patients had typeII diabetes with an average disease dura-tion of 15.3 ± 13 years. HbAx varied from7.6 to 19.8% with an average value of11.6 ± 2.8%. The neuropathy status ofthe patients at the time of recruitment wasunknown. All patients were administered

the MNSI (Figs. 1 and 2), followed by theMDNS (Table 3) and the following diag-nostic tests: Neuropathy Disability Score(NDS), Neuropathy Symptom Profile(NSP), quantitative vibration testing, au-tonomic functioning testing, and nerveconductions.

MNSIThe MNSI (Figs. 1 and 2) consists of 15 "yesor no" questions on foot sensation (pain,numbness, and sensitivity to tempera-ture), including one relevant to generalaesthenia and one relevant to peripheralvascular disease, and was administered toall patients. The questions chosen werethought to reflect common symptoms re-ported in DN together with two questionsto record non-neuropathic and primarilyvascular symptoms. The questionnairewas followed by a brief clinical examina-tion involving 1) foot inspection for de-formities, dry skin, callus, infection, orulceration, 2) semiquantitative assess-ment of vibration sensation at the dorsumof the great toe (normal, reduced, or ab-sent), and 3) grading of ankle reflexes(normal, reduced, or absent). Foot defor-mities included hammer toes, overlap-ping toes, halux valgus, joint subluxation,prominent metatarsal heads, and medialconvexity (Charcot foot). Abnormalitywas determined by the number of posi-tive responses or abnormal clinical find-ings. In both the question and examina-tion portion of the MNSI, the greater thescore, the greater is the neuropathy.

MDNSThe MDNS (Table 1 and Fig. 3) consistsof two parts, each of which are easily con-ducted in routine clinical practice: a clin-ical neurological examination followed byroutine nerve conduction measurements.Vibratory threshold perception, pain, andlight touch were assessed with a 128 Hztuning fork, a pin, and a 10-g filament,respectively. Previous work suggests thatmono filaments, as a measure of touchperception, can indicate the pressurethreshold that confers protection againstplantar ulceration (14). In the current

1282 DIABETES CARE, VOLUME 17, NUMBER 1 1 , NOVEMBER 1994

Dow


Feldman and Associates

Neuropathy Screening Instrument Questionnaire

Please take a few minutes to answer the questions below about the feeling in your legs andfeet. Check yes or no based on how you usually feel.

1. Are your legs and/or feet numb?

2. Do you ever have any burning pain in your legs and/or feet?

3. Are your feet too sensitive to touch?

4. Do you get muscle cramps in your legs and/or feet?

5. Do you ever have any prickling feelings in your legs or feet?

6. Does it hurt when the bed covers touch your skin?

7. When you get into the tub or shower, are you able to tell thehot water from the cold water?

8. Have you ever had an open sore on your foot?

9. Has your doctor ever told you that you have diabeticneuropathy?

10. Do you feel weak all over most of the time?

11. Are your symptoms worse at night?

12. Do your legs hurt when you walk?

13. Are you able to sense your feet when you walk?

14. Is the skin on your feet so dry that it cracks open?

15. Have you ever had an amputation?

Figure 1—MNSl questionnaire.

1. yes

1. yes

1. yes

1. yes

1. yes

1. yes

1. yes

1. yes

1. yes

1. yes

1. yes

1. yes

1. yes

1. yes

1. yes

••DDD•D

D••D••••

TOTAL:

2. no

2. no

2. no

2. no

2. no

2. no

2. no

2. no

2. no

2. no

2. no

2. no

2. no

2. no

2. no

•••••••••••••••

/15Pts

reflexes. When evaluating muscle strengthin the face, neck, arms, and legs, normalstrength is coded as 0; 25% weak as 1;50% weak as 2; 75% weak as 3; and 100%weak as 4. Sensation was examined inboth index fingers and in the dorsum ofboth great toes. Modalities examined in-cluded light touch, pinprick, vibratorysense, and joint position sense. These mo-dalities were graded as 0 for normal; 1,2,and 3 for mild, moderate, and severe re-ductions, respectively; and 4 for absentsensation. Muscle stretch reflexes, includ-ing biceps brachii, triceps brachii, brachi-oradialis, quadriceps, and triceps surae(Achilles) were graded as 0 for normal; 1,2, or 3 for mild, moderate, or severe de-creases, respectively; and 4 for an absentreflex. A total NDS ^ 2 was considered tobe abnormal.

NSPThe NSP was administered as previouslydescribed (16) and scored by opticalreader and by computer at the MayoClinic. Responses ^99th percentile forthe neuropathy and individual subsetcomponents (weakness, sensory, and au-tonomic scores) were considered abnor-mal. When two of the components wereabnormal, the entire NSP was consideredabnormal.

protocol, the 10-g filament is applied tothe dorsum of the great toe, and the pa-tient is asked to respond "yes" if he or shefeels the filament. Eight correct responsesout of 10 applications is considered nor-mal (score of 0); 1-7 correct responsesindicates reduced sensation (score of 1);and no correct answers translate into ab-sent sensation (score of 2).

Tendon reflexes are scored as 0for normal, 1 for abnormal, and 2 for ab-sent responses. Muscle strength is scoredas 0 for normal, 1 for mild to moderate,and 2 for severe weakness, while com-plete loss of strength is scored as 3. Forexample, sensation that is present but re-duced, reflexes that are present only with

reinforcement, and mild-to-moderate butnot severe weakness are each scored as 1.Absent reflexes, absent sensation, and se-vere weakness are scored as 2. Nerve con-ductions (sural, peroneal motor, mediansensory and motor, and ulnar sensory)are graded separately: 0 for normal and 1for abnormal values. Each patient is thengiven a composite score based on thenumber of abnormal nerve conductionsand the number of points scored on theclinical examination.

NDSThe NDS has been described in detail(15) and consists of an evaluation of apatient's muscle strength, sensation, and

Quantitative vibration testingVibration thresholds on the plantar as-pects of both great toes were measuredwith a Vibratron II (Physitemp, Clifton,NJ). A two-alternative forced-choicemethod was used in which the patientwas required to decide which of two rodswas vibrating (17). Only one rod was vi-brating as determined in a random fash-ion by the examiner. When the patientgave a correct response, the voltage wasdecreased by 10% for each trial until thepatient made an error. When the patienthad made five errors, testing was com-pleted. Patients with responses of > 3 vi-bration units in both the right and lefttoes were considered to have abnormalvibration sensation.

DIABETES CARE, VOLUME 17, NUMBER 1 1 , NOVEMBER 1994 1283

Dow



Appearance of feet Right Normal

Yes( 0 )

No( 1 )

Appearance of feet

Ulceration Right

Left

RightLeft

RightLeft

Present(0 )

Present(0 )

Ankle reflexes

Vibration perceptionat great toe

Figure 2—MNSI/orm.

Autonomic functioning testingThe resting heart rate was calculated aftera period of 20-min supine rest. The heartrate variability response to six deepbreaths/min (5 s in and 5 s out) was re-corded for 1 min on a continuous electro-cardiogram trace. The maximum andminimum R-R intervals during eachbreathing cycle were measured and con-verted to beats/min, and a mean value wascalculated for the six measured cycles.The heart rate response to the Valsalvamaneuver (expiration against a pressureof 40 mmHg for a period of 15 s) wasperformed three times, and a mean valuewas calculated for the ratio of the longestR-R interval after the maneuver to theshortest R-R interval during the maneu-ver. The patient then rested supine for 20min. After a mean supine systolic blood

If no, check all that apply:

DeformedDry skin, CallusInfection, Fissure

Left Normal

Yes No( 0 ) (1 )

If no, check all that apply:

DeformedDry skin, CallusInfection, Fissure

Absent( 0 )

Present(1)

Present/reinforcement(0.5)

Reduced(0-5)

TOTAL

Absent(1)

Absent(1)

/8Pts.

pressure (sBP) was measured, the patientstood erect, and the sBP was recorded im-mediately and at 1-min intervals thereaf-ter for a further 5-min period. The loweststanding sBP was recorded, and the sBPfall calculated. If two of four tests wereoutside published normal values (18,19),the patient was considered to have abnor-mal autonomic function.

Nerve conductionNerve conduction studies were per-formed on the nondominant side with thelower limb maintained at 32°C and theupper limb at 33°C. Sural, median, andulnar sensory-evoked potential ampli-tudes and distal and peak latencies wereevaluated. The amplitudes of the com-pound muscle action potentials for theperoneal and median motor nerves and

their respective distal latencies and con-duction velocities were performed. Anerve was considered abnormal if any at-tribute (amplitude, distal latency, or con-duction velocity) was not within the nor-mal limits, defined as values between thefirst and 99th percentiles (20). When twoor more nerves were abnormal, nerveconduction was considered abnormal ac-cording to the Mayo Clinic staging criteria(2).

Neuropathy classificationPatients were staged according to theMayo Clinic classification (2). Patientswith stage 0 (no neuropathy) had fewerthan two abnormalities among nerve con-ductions, NDS, vibration thresholds orautonomic function testing, or NSP. Stage1 (asymptomatic neuropathy) was de-fined as two or more abnormalitiesamong nerve conductions, NDS, or vibra-tion threshold or autonomic functiontesting, but no abnormality of NSP. Instage 2 (symptomatic neuropathy), therewere two or more abnormalities amongnerve conductions, NDS, vibrationthreshold or autonomic function, or NSP.In stage 3 (disabling neuropathy), therewere also two or more abnormalitiesamong nerve conductions, NDS, vibra-tion threshold or autonomic function, orNSP. However, when compared with pa-tients in stage 2, patients in stage 3 weredisabled by their neuropathy.

Statistical analysisSummary data are expressed as means ±SD of the mean. Correlations were com-puted between summary scores. P values< 0.05 are considered statistically signif-icant.

RESULTS

Prevalence and staging ofneuropathySixty-nine percent of patients had abnor-mal nerve conductions, 61% had abnor-mal NDSs, 53% had abnormal vibrationthresholds, 34% had abnormal auto-nomic function testing, and 30% had ab-


Dow



Table 1—Diabetic neuropathy score

Sensory impairmentRight

Vibration at big toe10-g filamentPin prick on dorsum of great toe

LeftVibration at big toe10-g filamentPin prick on dorsum of great toe

Muscle strength testingRight

Finger spreadGreat toe extensionAnkle dorsiflexion

LeftFinger spreadGreat toe extensionAnkle dorsiflexion

ReflexesRight

Biceps brachiiTriceps brachiiQuadriceps femorisAchilles

LeftBiceps brachiiTriceps brachiiQuadriceps femorisAchilles

Total: 46 points

Normal00

Painful0

Normal00

Painful0

Normal000

Normal000

Present0000

Present0000

Decreased11

Not painful2

Decreased11

Not painful2

Mild to moderate111

Mild to moderate111

Present with reinforcement1111

Present with reinforcement1111

Absent22

Absent22

Severe222

Severe222

Absent2222

Absent2222

Absent333

Absent333

normal NSPs. Using the Mayo Clinic clas-sification, 38% of the patients werenormal (no neuropathy), 39% had stage 1neuropathy (asymptomatic neuropathy),18% had stage 2 neuropathy (symptom-atic neuropathy), and 5% had stage 3neuropathy (disabling neuropathy). Thepercentage of patients with neuropathy inthe current study agrees with previouslyreported data obtained from a larger sam-ple size (21).

MNSIThe results of the patient scores on theclinical examination and question portionof the MNSI were compared with the pa-tient's neuropathic status according to theMayo Clinic criteria (Table 2). When the15 questions were analyzed, 18 patients

without DN and 20 patients with DN an-swered positively on ^ 6 questions,thereby indicating that many patientswithout DN would answer positively tothese questions. Two patients with no DNand 14 patients with DN answered posi-tively on ^ 7 questions (Table 4). The re-sults of the MNSI clinical portion are alsopresented in Table 4. Nineteen normalpatients and six patients with DN had ascore of ^ 2 , while no normal patients and26 patients with DN had a score of >2.This indicates that a score >2 accuratelyidentifies patients with DN. The sensitiv-ity of the clinical portion of the MNSI is80%, and the specificity is 95%. The pos-itive predictive value is 97%, and the neg-ative predictive value is 74%.

Correlations were computed be-tween the summary score for the varioustests (Table 3). The clinical portion of theMNSI correlated with the MDNS, theNDS, vibration thresholds and autonomicfunction testing, and nerve conductions.In contrast, the questionnaire portion ofthe MNSI correlated only with the NSP.The clinical examination, therefore, ap-pears to be a good predictor of DN as di-agnosed by abnormalities in electrophysi-ological tests. However, there is a lowcorrelation between positive answers onthe symptom questionnaire and the pres-ence of defined DN.

MDNSThe results of the patient's scores on theclinical portions and the nerve conduc-tion components of the MDNS were com-pared with the patient's neuropathic sta-tus according to the Mayo Criteria (Table4). Twenty-one patients with an MDNS of^ 6 had no neuropathy, while seven hadneuropathy according to the Mayo classi-fication. All 28 patients with an MDNS^ 7 had neuropathy according to theMayo classification. When evaluating theelectrodiagnostic component of theMDNS, 21 patients with no neuropathyand 1 patient with neuropathy had ^ 1abnormal nerve conduction, while 34 pa-tients with neuropathy had ^ 2 abnormalnerve conductions (Table 4).

Based on the electrophysiologicaland clinical results of the MDNS, DN wasdivided into four classes defined by theclinical MDNS and the number of nerveconductions abnormalities (Fig. 4). Pa-tients with < 1 abnormal nerve conduc-tion and ^ 6 points on the clinical portionof the MDNS were graded as class 0 or noneuropathy. Class I or mild neuropathywas defined as two abnormalities onnerve conductions and a clinical score of^ 1 2 ; class II neuropathy, defined asmoderate neuropathy, included patientswith either three or four abnormal nerveconductions and a clinical score of <29;and class III neuropathy, considered se-vere neuropathy, was defined as abnor-malities in all five nerve conductions with

DIABETES CARE, VOLUME 17, NUMBER 11, NOVEMBER 1994 1285

Dow



VISITAbnormal

Nerves

0-1

2

3-4

5

ClinicalPoints

o .i

7

5A

>6<7

i

R0

7 n io , _1112

>12<13

] - ,

14I S

161718

13-29 ^71M

242S26272829

>29<30

31

3334

36

3839

41M

43 . . _44

Date

Score

Date

Score

Date

Score

Date

Score

Date

ScoreCLASS

0no neuropathy

1mild

neuropathy

2moderate

neuropathy

3severe

neuropathy

Figure 3—MDNS score sheet.

time of or anytime after diagnosis of dia-betes and in type I patients after 5 years ofdiabetes (22). It is imperative to performat least an annual neurological historyand clinical examination to detect evi-dence of DN, which can result in an in-sensitive foot with the inherent risk of cal-lus and fissure formation with plantarulceration (10). Early detection of DN isessential for the initiation of potential pre-ventative measures, patient education,and evaluation of therapeutic options. Pa-tient education concerning foot care maymake a substantial impact on reducingthe inherent morbidity of DN (23).

To facilitate the early diagnosis ofDN, we created the MNSI for use by gen-eral and nurse practitioners, internists,endocrinologists, and diabetologists. Indeveloping the MNSI, we hoped to assessboth clinical symptoms and the key clin-ical signs of DN. A 15-item, yes/no ques-tionnaire that included 13 questions rele-vant to DN, one to general aesthenia andone to peripheral vascular disease, wasadministered to all patients. An essen-tially equal number of patients with andwithout DN answered positively on upto six questions, demonstrating thatsymptoms may not always indicate un-derlying neuropathy. Accurate assess-ment of symptoms in DN is known to bedifficult, as evidenced by the poor repro-ducibility of the current standard proto-cols (24). In the Rochester Diabetic Neu-

clinical scores ^46. Within each class,there was a range of clinical scores with apositive correlation between higher clini-cal scores and number of abnormal nerveconductions.

Examination of the MDNS copaircorrelation matrix revealed that theMDNS correlated highly with the NDS,vibration thresholds, autonomic functiontesting, and nerve conductions (Table 3),but not with the NSP.

CONCLUSIONS — DN may be pres-ent in all type II diabetic patients at the

Table 2—MNSI

Clinical examscore

012

>3Total

<2

Normal

6320

11

DN

12069

Question score positive responses

3-4

Normal

3——03

DN

1——

23

5-6

Normal

1214

DN

0088

>7

Normal

20002

DN

011

1214

Total

Normal

11441

20

DN

231

2834

Comparison of positive patients' responses on the questionnaire to the clinical examination score of theMNSI. Two patients had incomplete questionnaires. Normal, non-neuropathic patients; DN, neuropathicpatients as defined by the Mayo Clinic (2), stages 1-3.


Dow



Table 3—Summary score correlations

MNSIQMNSICMDNSNDSNSPVTAFNC

MNSIQ

0.430.300.400.60*0.350.190.30

MNSIC

0.43—

0.77*0.73*0.340.65*0.62*0.55*

MDNS

0.360.77*

—0.67*0.360.61*0.69*0.59*

NDS

0.400.73*0.67*

—0.320.60*0.490.66*

NSP

0.60*0.340.360.32

—0.160.160.22

VT

0.350.65*0.61*0.60*0.16

—0.54*0.54*

AF

0.190.62*0.69*0.490.160.54*

—0.39

NC

0.300.55*0.50*0.66*0.220.54*0.39

—

*A correlation coefficient of SO.5 is significant at the P < 0.05 level. MNSIQ, questionnaire portion ofMNSI;MNSIC, clinical component ofMNSI; VT, vibration testing; AF, autonomic functioning; NC, nerve conduc-tion.

ropathy Study, assessment of symptomsby the Neuropathy Symptom Score or theNSP was not highly reproducible (24).Significant differences in judging symp-tom severity by the Neuropathy Symp-tom Score occurred even among highlytrained neurologists who used the samesymptom grading criteria (24).

In contrast, the clinical portion ofthe MNSI showed both high specificity(95%) and sensitivity (80%) in predictingthose patients with DN. The examinationwas brief and involved inspection of thefeet for abnormalities, particularly defor-mities, callus, fissure, or infection, as wellas assessment of vibration perception andankle tendon reflexes. These last twocomponents of the MNSI have a high de-gree of reproducibility and sensitivity inthe diagnosis of DN (21,24). In the cur-

rent study, there was a strong correlationbetween a positive score (i.e., indicativeof DN) on the clinical component of theMNSI and a positive reading on the MayoClinic's NDS, nerve conductions, quanti-tative vibration threshold, and the newlydeveloped MDNS. Consequently, 28 of29 patients who had an MNSI clinical ex-amination score of >2 had DN. Combin-ing the results of the MNSI questionnairewith the scores on the clinical componentdid not further improve the diagnosis ofDN, highlighting the great importance ofthe examination component. We suggestthat patients who have an MNSI score of>2 on the clinical examination undergo a

Table 4—Diabetic neuropathy score

more thorough neuropathy evaluation asoutlined in the MDNS.

The two-part MDNS was devel-oped to confirm the diagnosis of DN. Itssimple initial clinical component requiresno specialized equipment apart from thewidely available monofilament. Applica-tion of the clinical component aloneshould be a useful tool in everyday prac-tice to confirm the presence of neuropa-thy and allow for the instigation of pre-ventative measures. More detailed stagingof neuropathy, in particular to monitor itsprogression or response to treatment, re-quires that the patient undergo routinenerve conduction studies. Certainly, thecomprehensive criteria set forth by boththe Mayo Clinic and the San AntonioConsensus are required for a full assess-ment of diabetic patients for DN clinicaltrials. However, because these diagnosticcriteria are both time consuming and ex-pensive, we believe that the MDNS mayplay an important role in the diagnosisand staging of DN.

The clinical parameters chosenfor the MDNS were based on previous re-ports of the sensitivity and specificity ofboth the 10-g filament (25) and the MayoClinic's NDS (21,24) to diagnose neurop-athy. The subsets of the NDS present inthe MDNS (specifically distal sensation,

za4

Class0

21*

4

3

Class1

. .

3

1

Class2

DNS Score fClinican

12zl& 12=22

..1

1 2

Class3

£30

Figure 4—Neuropathy classes as defined by the

MDNS by comparison of patient scores on the

clinical examination and nerve conduction com-

ponents of the MDNS. "'Non-neuropathic patients

as defined by the Mayo Clinic (2).

Clinical componentsSensory component

Weakness component

Reflex component

Total

Electrodiagnostic componentABNNC

core

<6>7

010

>1<6>7

<1>2

Non-neuropathicpatients

210

210

210

210

210

Neuropathicpatients

926

9137

287

28

134

Neuropathic patients were defined by the criteria of the Mayo Clinic (2), stages 1-3.


Dow



strength, and reflexes) are the most repro-ducible components of the NDS (24) andare highly predictive of DN (21). It is,therefore, not surprising that the clinicalportion of the MDNS correlated well withthe NDS, as well as with vibration thresh-old, autonomic function testing, andnerve conductions.

The use of nerve conduction ab-normalities to diagnose DN is well estab-lished (21,22,24) and formed the secondcomponent of the MDNS. The high re-producibility of both the compound mus-cle and sensory nerve action potentials,along with motor nerve conduction ve-locities, in the Rochester Diabetic Neu-ropathy Study has led to the suggestionthat these tests, which reflect actualchanges in peripheral nerve pathology,should be emphasized in the diagnosis ofDN and should be used in controlled clin-ical and epidemiological trials (21). In theRochester Diabetic Neuropathy Study,nerve conductions showed both good re-producibility and a high correlation of di-rect measurement of neurological deficitscompatible with DN (21,24). Certainly,recent comparisons of sural nerve biop-sies and nerve conductions in diabetic pa-tients with clinically significant DN revealthat the two key electrophysiological hall-marks of DN (decreased sural sensory-evoked amplitude and slowed peronealmotor conduction velocity) correlate wellwith characteristic structural lesions ofmyelinated nerve fibers (26,27). The am-plitude of the sural nerve action potentialcorrelates with the density of myelinatedfibers, which decreases over time in pa-tients with DN. Similarly, the observeddecline in peroneal motor nerve conduc-tion velocity in patients with known DNis likely to reflect a combination of demy-elination and loss of the largest myeli-nated fibers.

The high reproducibility of nerveconductions studies and their close corre-lation with nerve fiber loss and structuralinsults in DN make these tests sensitiveindicators of the presence of DN. In theMDNS, therefore, we graded the severityof DN in diabetic patients depending on

the number of abnormal nerve conduc-tions and coupled this parameter with thepatient's score on the clinical examina-tion. In our simple staging scale, patientswith <1 abnormality on nerve conduc-tions, coupled with a clinical score of ^ 6 ,are considered normal or class 0 with noevidence of DN. In this study, all 21 pa-tients staged normal by the Mayo Clinicprotocol fell into our class 0. Patients withtwo abnormalities of nerve conductionsare placed into class 1 (mild DN), patientswith three or four abnormalities into class2 (moderate DN), and patients with fiveabnormalities into class 3 (severe DN).Within each class, DN is further quanti-fied depending on the patient's clinicalscore. For example, if a patient has threeabnormal nerve conductions and a clini-cal MDNS of 10, he is given a compositescore of 3.10. We believe this currentclassification system allows a short andaccurate assessment of the degree of se-verity of DN and allows us to efficientlymonitor the progression of DN in a pa-tient over time.

In a similar fashion to the Roches-ter Diabetic Neuropathy Study (21,24),we initially included median sensory andmotor nerves as part of our assessment forDN. We are currently studying the utilityof substituting the ulnar motor and sen-sory nerves for their median counterpartsin the diagnosis of DN. In our cohort, aswas seen in the Rochester Diabetic Neu-ropathy Study, there exists asymptomaticmedian mononeuropathies at the wrist(carpal tunnel syndrome) (21,24). In thefuture, the ulnar motor and sensorynerves may prove to be equally as diag-nostic for DN as are the median motorand sensory nerves and, because of thedecreased frequency of ulnar entrapment,allow for simpler analysis of nerve con-ductions results.

In summary, early detection ofDN in diabetic patients is essential to de-crease morbidity. We have developed abrief screening tool, the MNSI, to be usedby health care professionals in the outpa-tient setting to screen for possible DN.Patients with a positive MNSI may then

be administered the MDNS, a quantita-tive classification of DN. The MDNS wasdeveloped not to replace the current cri-teria outlined by the San Antonio Con-sensus and the Mayo Clinic, but rather tobe used in the outpatient setting whereboth time and resources are limited. Be-cause we believe that abnormalities innerve conductions reflect the pathologi-cal abnormalities present in nerves of pa-tients with DN, we constructed our clas-sification of DN based on the number ofabnormalities in nerve conductions and aclinical score obtained in the MDNS. Pa-tients with DN are placed in one of fourstages: 0, no neuropathy; 1, mild neurop-athy; 2, moderate neuropathy; and 3, se-vere neuropathy. We are currently exam-ining the use of the MNSI and the MDNSin the diagnosis and classification of DNin larger groups of diabetic patients fromboth the U.S. and Europe. Our long-range goal is to achieve early diagnosisand classification of DN, which will allowfor the development and institution of ra-tional therapies early in the disease whenwe believe therapeutic potential is greatest.

Acknowledgments— These studies weresupported by National Institutes of HealthGrant NS01380 (E.L.F.), USPHS ResearchGrants RO1-DK-38304 (D.A.G.), and theMichigan Diabetes Research and TrainingCenter P60-DK-20572 (D.A.G., M.B.B.,E.L.F.). The authors acknowledge the experttechnical assistance of Susan Nalepa and sec-retarial assistance of Judy Kiger.

References

1. Consensus Panel: Report and recommen-dations of the San Antonio Conference onDiabetic Neuropathy (Consensus state-ment). Diabetes 37:1000-1004, 1988

2. Dyck PJ: Detection, characterization, andstaging of polyneuropathy: assessed in di-abetics. Muscle & Nerve 11:21-32, 1988

3. Committee Health Care Issues AmericanNeurological Association: Does improvedcontrol of glycemia prevent or amelioratediabetic neuropathy? Ann Neurol 19:288-290,1986


Dow



4. Pirart J: Diabetes mellitus and its degen-erative complications: a prospective studyof 4,400 patients observed between 1947and 1973. Diabetes Care 1:168-188,1978

5. The Diabetes Control and ComplicationsTrial Research Group: Factors in the de-velopment of diabetic neuropathy: base-line analysis of neuropathy in the feasibil-ity phase of the Diabetes Control andComplications Trial (DCCT). Diabetes 37:476-481, 1988

6. Young MJ, Boulton AJM, Macleod AF,Williams DRR, Sonksen PH: A multicen-tre study of the prevalence of diabeticperipheral neuropathy in the UnitedKingdom hospital clinic population. Dia-betologia 36:150-154, 1993

7. Comi G, Negrin P, Troni W: Quantitativesensory testing, cardiovascular auto-nomic tests and nerve conduction studiesin diabetic neuropathy. In Diabetic Neu-ropathy. WardJ, Goto Y, Eds. Chichester,U.K., Wiley, 1990, p. 351-359

8. Greene DA, Sima AAF, Stevens MJ, Feld-man EL, Lattimer SA: Complications:neuropathy, pathogenetic considerations.Diabetes Care 15:1902-1925, 1992

9. Brand PW: The diabetic foot. In DiabetesMellitus. Ellenberg M, Rifkin H, Eds. NewYork, Medical Examination Publishing,1983,p.829-849

10. Edmonds ME: The diabetic foot: patho-physiology and treatment. C!in EndocrinolMetab 15:889-916, 1986

11. Mulder DW, Lambert EH, Bastian JA: Theneuropathies associated with diabetes

mellitus: a clinical and electromyographicstudy of 103 unselected diabetic patients.Neurology 11:275-284, 1961

12. Dyck PJ, Karnes JL, Daube J: Clinical andneuropathological criteria for the diagno-sis and staging of diabetic polyneuropa-thy. Brain 108:861-880, 1985

13. American Diabetes Association, AmericanAcademy of Neurology: Report and rec-ommendations of the San Antonio Con-ference on Diabetic Neuropathy. DiabetesCare 11:592-597, 1988

14. Stevens MJ, Edmonds ME, Foster AVM,Watkins PJ: Selective neuropathy andpreserved vascular responses in the dia-betic Charcot foot. Diabetologia 35:148-154,1992

15. Dyck PJ, Sherman WR, Hallchen LM: Hu-man diabetic endoneural sorbitol, fruc-tose, and myoinositol related to suralnerve morphometry. Ann Neurol 8:590-596,1980

16. Dyck PJ, Karnes J, O'Brien PC, SwansonCJ: Neuropathy symptom profile inhealth, motor neuron disease, diabeticneuropathy, and amyloidosis. Neurology36:1300-1308, 1986

17. Maser RE, Nielsen VK, Bass EB: Measur-ing diabetic neuropathy: assessment andcomparison of clinical examination andquantitative sensory testing. Diabetes Care12:270-275, 1989

18. Wheeler T, Watkins PJ: Cardiac denerva-tion in diabetes. Br MedJ 4:584-586,1973

19. Ewing DJ, Clarke BF: Diagnosis and man-agement of diabetic autonomic neuropa-thy. Br MedJ 385:916-918, 1982

20. KimuraJ: Principles of nerve conductionstudies. In Electrodiagnosis in Diseases of

Nerve and Muscle: Principles and Practice.

Philadelphia, Davis, 1983, p. 83-10421. Dyck PJ, Karnes JL, O'Brien PC: The

Rochester Diabetic Neuropathy Study: re-assessment of tests and criteria for diag-nosis and staged severity. Neurology 42:1164-1170,1992

22. Greene DA, Sima A, Pfeifer MA, AlbersJW: Diabetic neuropathy. Ann Rev Med41:303-317, 1990

23. Levin ME: Pathogenesis and managementof diabetic foot lesions. In The DiabeticFoot. Levin ME, O'Neal LW, Bowker JH,Eds. St. Louis, MO, Mosby Year Book,1993,p. 17-60

24. Dyck PJ, Kratz KM, Lehman KA: TheRochester Diabetic Neuropathy Study:design, criteria for types of neuropathy,selection bias, and reproducibility of neu-ropathic tests. Neurology 41:799-807,1991

25. Birke JA, Sims DS: Plantar sensory thresh-old in the ulcerative foot. Lepr Rev 57:261-267, 1986

26. Sima A, Bril V, Nathaniel Y: Degenerationand repair of myelinated fibers in suralnerve biopsies from patients with diabeticneuropathy treated with sorbinol, an in-vestigational aldose reductase inhibitor.NEnglJMed 19:548-555, 1988

27. Sima A, Cherian V, Albers JW, GreeneDA, The Tolrestat Study Group: Nerve fi-ber loss in diabetic neuropathy correlateswith impaired evoked potential ampli-tudes and nerve conduction velocity (Ab-stract). Diabetologia 35:606, 1992


Dow


1 - diabetes care

Documents