Diabetic Neuropathy: Pathogenesis and Therapy
Aaron I. Vinik, MD, PhD
Diabetic neuropathies are complex, heterogeneous
disorders that encompass a wide range of abnormal-
ities affecting both peripheral and autonomic ner-
vous systems, causing considerable morbidity and
mortality. Treatment should be based upon the un-
derlying etiology and not symptoms alone, although
symptomatic therapy is needed. Neuropathies may
be focal or diffuse, proximal or distal, and involve
somatic and autonomic nerves. Focal syndromes
are classified as (1) entrapment syndromes or (2)
mononeuropathies. Entrapment syndromes are
treated by means of relieving compression within
confined spaces. Mononeuropathies are due to a
vascular insult and resolve spontaneously. They are
best treated by supportive therapy. Proximal neurop-
athies are usually due to an inflammatory, vasculitic,
or autoimmune condition and are best treated with
specific therapies for the underlying disorder based
on biopsy findings. Therapies for distal polyneurop-
athies include metabolic treatments (e.g., aldose re-
ductase inhibitors, aminoguanidine, ␥-linolenic acid),
autoimmune therapies, and nerve growth factors. No
definitive treatment is available for painful diabetic
neuropathy. Several medications have been used,
among them tricyclic antidepressants, antiepileptic
drugs, phenothiazines, calcitonin, local anesthetics,
nonsteroidal anti-inflammatory drugs, and dextro-
methorphan. Nonpharmacologic therapies include
surgical sympathectomy, spinal cord blockade, elec-
trical spinal cord stimulation, and prostaglandin. Am
J Med. 1999;107(2B):17S–26S. © 1999 by Excerpta
Medica, Inc.
From the Strelitz Diabetes Institutes, Departments of Internal Medi-
cine, Anatomy and Neurobiology, Eastern Virginia Medical School,
Norfolk, Virginia.
Requests for reprints should be addressed to Aaron I. Vinik, MD,
PhD, The Strelitz Diabetes Institutes, 855 West Brambleton Avenue,
Norfolk, Virginia 23510.
© 1999 by Excerpta Medica, Inc.
All rights reserved.
D
iabetic neuropathy comprises a number of dif-
ferent syndromes, each with a range of clinical
and subclinical manifestations. The main groups
of neurologic disturbance in diabetes mellitus include1:
(1) subclinical neuropathy determined by abnormalities
in electrodiagnostic and quantitative sensory testing; (2)
diffuse clinical neuropathy with distal symmetric senso-
rimotor and autonomic syndromes; and (3) focal syn-
dromes. Proximal neuropathy may be considered as a
separate entity based upon the nature of the pathology
and the response to treatment.
PATHOGENIC MECHANISMS
Figure 1 shows a current view of the pathogenesis of di-
abetic neuropathy. Although there is increasing evidence
that the pathogenesis of diabetic neuropathy is multifac-
torial, the prevailing theory implicates persistent hyper-
glycemia as the primary factor.2,3 Persistent hyperglyce-
mia increases polyol pathway activity with accumulation
of sorbitol and fructose in nerves, damaging them by an
as yet unknown mechanism. This is accompanied by de-
creased myo-inositol uptake and inhibition of the Na⫹/
K⫹-adenosine triphosphatase, resulting in Na⫹ reten-
tion, edema, myelin swelling, axoglial disjunction, and
nerve degeneration. Deficiency of dihomo-␥-linolenic
acid (GLA) as well as N-acetyl-L-carnitine have also been
implicated.4 Metabolic factors cannot account for all
forms of neuropathy or for the heterogeneity of the clin-
ical syndromes.
Immune Mechanisms
In some patients, immune mechanisms may be responsi-
ble for the clinical neuropathic syndrome, especially in
those with proximal neuropathy and those with a more
marked motor component to their neuropathy. Circulat-
ing antineuronal antibodies are present in diabetic serum
in some patients. The circulating autoantibodies directed
against motor and sensory nerve structures have been
detected by indirect immunofluorescence, and antibody
and complement deposits in various components of sural
nerves have been shown.5–7
A cross-sectional analysis of 154 patients with diabetic
neuropathy revealed that 12% had positive anti-GM1-
ganglioside antibody (an antibody subtype) findings.
Anti-GM1 antibodies were frequently associated with dis-
tal symmetric polyneuropathy (DSPN) characterized by a
slight emphasis on a motor deficit with electrophysi-
ologic signs of demyelination.8 Antiphospholipid anti-
bodies (anti-PLAs) were present in 88% of this popula-
0002-9343/99/$20.00 17S
PII S0002-9343(99)00009-1
 A Symposium: Diabetic Neuropathy—Pathogenesis and Therapy/Vinik

Figure 1. Current view of the pathogenesis of diabetic neuropathy. Diabetic neuropathy is a heterogeneous disorder. Superim-
posed on a genetic predisposition and environmental factors such as smoking and alcohol consumption, hyperglycemia and the other
metabolic abnormalities lead to increased polyol activity, activation of protein kinase C (PKC), and myo-inositol and ␥-linoleic acid
(GLA) deficiency. In addition, there is a tendency to microvascular insufficiency with “oxidative stress,” increase in the vasocon-
strictor endothelin and PKC, and deficiency of the vasodilators nitric oxide (NO) and prostacyclin (PGI). The third arm is perpet-
uation of the nerve insult by autoimmune destruction of nerves. The initial abnormality is functional, culminating in structural
changes with nerve degeneration and impairment of the regenerative response due to deficiencies in important neurotrophic factors.
Ab ⫽ antibody; EDRF ⫽ endothelial-dependent relaxing factor; GF ⫽ growth factor; IGF ⫽ insulin-like growth factor; NGF ⫽ nerve
growth factor; TRK ⫽ tyrosine kinase. (Adapted from Diabetes Rev.4)

tion, compared with 32% in diabetic patients without
apparent neurologic complications and 2% in the general
population.9 There is evidence that PLAs may be injuri-
ous to neural tissue and that damage may be selective for
specific parts of the nervous system.9 Because PLAs are
associated with a tendency to vascular thrombosis, their
presence may provide a link between the immune and
vascular theories of causation of neuropathy.
Microvascular Insufficiency
Microvascular insufficiency has been implicated as a
cause of diabetic neuropathy.10 –12 The interest in micro-
vascular derangement arises from studies suggesting that
absolute or relative ischemia exists in the nerves of dia-
betic persons due to altered function of the endoneurial
and/or epineurial blood vessels. Histopathologic studies
show the presence of different degrees of endoneurial
and epineurial microvasculopathy, mainly thickening
of blood vessel wall or occlusion.13,14 Functional
disturbances have also been demonstrated in the micro-
vasculature of the nerves of diabetic persons. Studies
show decreased neural blood flow,15 increased vascular
resistance,16 decreased PO2,16,17 and altered vascular per-
meability characteristics such as a loss of the anionic
charge barrier and decreased charge selectivity. It has also

18S August 30, 1999 THE AMERICAN JOURNAL OF MEDICINE威 Volume 107 (2B)
 A Symposium: Diabetic Neuropathy—Pathogenesis and Therapy/Vinik
been shown that abnormalities of cutaneous blood flow
correlate with neuropathy.18
Growth Factor Deficiency
Growth factor deficiency may also play a role in the
pathogenesis of diabetic neuropathy. Many of the char-
acteristic neuronal changes are similar to those seen after
removal of target-derived growth factors by axotomy or
depletion of endogenous growth factors by experimental
induction of growth factor autoimmunity. Because neu-
ronal growth factors can promote survival, maintenance,
and regeneration of neurons subject to the noxious ef-
fects of diabetes, the success of diabetic patients in main-
taining normal nerve morphology and function may de-
pend on expression and efficacy of these factors.4
It is established that sympathetic and dorsal root gan-
glion (DRG) neurons are developmentally dependent on
nerve growth factor (NGF). Recently, it was shown that
adult DRG and sympathetic neurons, both of which are
affected in diabetic neuropathy, are dependent on NGF
for either their maintenance or their survival.19 NGF be-
longs to a gene family encoding structurally and func-
tionally related proteins called neurotrophins (NTs).
NGF has been implicated in diverse activities including
vasodilatation, gut motility, and nociception.4 Some data
suggest that a decrease in NGF synthesis has a role in the
pathogenesis of diabetic neuropathy, namely, in the func-
tional deficit of small fibers. These fibers have a role in
pain and thermal sensation. The effect of NGF depletion
may be mediated through the downregulation of neuro-
filament gene expression or mRNAs that encode the pre-
cursor molecules of substance P, both shown to be NGF
dependent.4 Another member of the neurotrophin fam-
ily, NT3, may be important for the survival and function
of the large nerve fibers subserving proprioception, vibra-
tion, and, possibly, motor function.4
The insulin-like growth factors (IGFs), IGF-I and IGF-
II, have been implicated in the growth and differentiation
of neurons, and IGF receptors are present in nerve tissues
(e.g., neurons, Schwann cells, ganglia) involved in diabe-
tes-associated nerve disorders. IGFs and their binding
proteins (IGF-BPs) are regulated by insulin and the gly-
cemic state.4,20,21 A consequence of insulin insufficiency
is a reduced circulating IGF-I concentration. Abnormal
IGF-I and IGF-II metabolism play causative roles in dia-
betic neuropathy. However, little is known regarding the
other effects of diabetes on local expression, synthesis,
and transport of these growth factors in nerve tissue.
Laminin
Laminin is a large, heteromeric, cruciform glycoprotein
composed of a large A chain and two smaller B chains, B1
and B2.22,23 Laminin promotes neurite extension by cul-
tured neurons.24 –28 Lack of normal expression of laminin
B2 gene may contribute to the pathogenesis of diabetic
neuropathy. In our laboratory, we have shown that all
August 30, 1999
normal adult rat DRG neurons express the laminin B2
chain gene.29 In other experiments, we described the up-
regulation of laminin B2 gene in regenerating neu-
rons.30,31 This finding indicates its importance in the pro-
cess of neuronal regeneration. However, laminin B2 ex-
pression under basal conditions and during neuronal
regeneration was decreased in diabetic animals.32 If these
changes are present in patients with diabetic neuropathy,
it might lay the foundation for exploring the therapeutic
potential of laminin.
FOCAL NEUROPATHIES
(MONONEURITIS AND ENTRAPMENT
SYNDROMES)
Mononeuropathies (when an individual nerve is af-
fected) are due to vasculitis and subsequent ischemia or
infarction of individual nerves.33 The isolated peripheral
nerve lesions involve particularly ulnar, median, radial,
femoral, and lateral cutaneous nerves of the thigh. Com-
mon mononeuropathies may also involve cranial nerves
3, 4, 6, and 7. Their onset is acute, sometimes associated
with pain, and their course is usually self-limiting, resolv-
ing over a period of 6 – 8 weeks. Mononeuropathies must
be distinguished from entrapment syndromes, which
start slowly, progress, and persist without intervention
(Table 1). Common entrapments involve the median
nerve with impaired sensation in the first three fingers
and positive Tinel sign. Ulnar entrapment decreases sen-
sory perception in the little and ring fingers. Medial and
lateral plantar entrapments decrease sensation in the in-
side and outside of the feet, respectively. The entrapment
neuropathies are common in persons with diabetes and
should be actively sought in every patient with signs and
symptoms of neuropathy because the treatment may be
surgical.34
Carpal tunnel syndrome occurs twice as often in the
diabetic population compared with the normal healthy
population. Its increased prevalence in diabetes may be
related to repeated undetected trauma, metabolic
changes, or accumulation of fluid or edema within the
confined space of the carpal tunnel. If suspected, the di-
agnosis can be confirmed by electrophysiologic study.
The physician should be aware that symptoms may
spread to the whole hand or arm, and the signs may ex-
tend beyond those subserved by the entrapped nerve. Be-
cause the nature of the trouble may go unrecognized, an
opportunity for therapeutic intervention is often missed.
The mainstays of nonsurgical treatment are avoidance
of the use of the wrist, placement of a wrist splint in a
neutral position for day and night use, and anti-inflam-
matory medications. Surgical treatment consists of sec-
tioning the volar carpal ligament. The decision to proceed
with surgery should be based on considerations such as
THE AMERICAN JOURNAL OF MEDICINE威 Volume 107 (2B) 19S
 A Symposium: Diabetic Neuropathy—Pathogenesis and Therapy/Vinik

Figure 2. Steps in the proposed pathways that lead to the pathogenesis of diabetic neuropathy and possible interventions at the level
of each proposed mechanism. AII ⫽ angiotensin II; ACE ⫽ angiotensin-converting enzyme; AGE ⫽ advanced glycation end
products; EFA ⫽ essential fatty acid (deficiency); ET ⫽ endothelin; GLA/EPO ⫽ ␥-linolenic acid/evening primrose oil; IGF ⫽
insulin-like growth factor; NGF ⫽ nerve growth factor; Nitrodil ⫽ nitrodilators; NO ⫽ nitric oxide; NT3 ⫽ neurotrophin-3; PDIE ⫽
phosphodiesterase; PGI2 ⫽ prostacyclin.

Table 1. Comparative Features of Mononeuritis and Entrapment

Feature Mononeuritis Entrapment
Onset Sudden Gradual
Pain Acute Chronic
Multiplex Occurs Rare
Course Resolves Persists without intervention
Treatment Physical therapy Rest/splints/surgery

severity of symptoms, appearance of motor weakness,
and failure of nonsurgical treatment.
DISTAL SYMMETRIC POLYNEUROPATHY:
THERAPEUTIC INTERVENTIONS
Glycemic Control
Retrospective and prospective studies suggest a relation
between hyperglycemia and the development and sever-
ity of diabetic neuropathy.35,36 The importance of tight
glycemic control is discussed elsewhere in this supple-
ment by Parry.36a
Many attempts at reversal of neuropathy focus on the
aftermath of hyperglycemia: increased polyol pathway
activity, growth factor insufficiency, and the immune-
mediated consequences of glycemia-induced changes to
protein structure, rendering them antigenic and activat-
ing immune destruction of neurons or their satellites
(Figure 2).37
Aldose Reductase Inhibitors
Aldose reductase inhibitors (ARIs) reduce the flux of glu-
cose through the polyol pathway, inhibiting tissue accu-
mulation of sorbitol and fructose and preventing reduc-
tion of redox potentials. Alrestatin, the first ARI studied,
was shown to improve sensory impairment scores in one
study,38 but in other studies no objective benefits were
noted.39 – 41 Patient selection (i.e., inclusion of patients
with severe neuropathy) may have complicated interpre-
tation of these trials because more severe neuropathy may
be irreversible.

20S August 30, 1999 THE AMERICAN JOURNAL OF MEDICINE威 Volume 107 (2B)
 A Symposium: Diabetic Neuropathy—Pathogenesis and Therapy/Vinik
Sorbinil, another widely studied ARI, generated more
promising clinical effects. In a double-blind crossover
trial, treatment with sorbinil resulted in small but signif-
icant improvements in both motor and sensory conduc-
tion velocities compared with placebo.42 Additional
studies have demonstrated minor improvements in test
results but failed to demonstrate progressive benefit with
long-term therapy.43,44 Enrollment in clinical trials was
discontinued because of toxic effects such as lymphade-
nopathy, rash, fever, and pancytopenia.
In a placebo-controlled, double-blind study of another
ARI, tolrestat, 45 diabetic patients with asymptomatic au-
tonomic neuropathy, as defined by at least one pathologic
cardiovascular reflex, were treated for 1 year.45 Patients
who received tolrestat had significant improvement in
autonomic function tests as well as in vibration percep-
tion, whereas placebo-treated patients showed deteriora-
tion in most of the parameters measured. A large, multi-
center trial is under way in the United States and Canada,
using carefully standardized clinical measures of symp-
toms and physical findings, as well as quantitative physi-
ologic testing and sural nerve biopsy.
Gangliosides
Gangliosides have been shown to promote improvement
in sensation without changes in nerve conduction or
small but significant improvements in motor and sensory
conduction.46 – 49 Although concerns about slow viruses
and transmission of autoimmune neurologic disorders
have been mitigated by refinement of the products, a
moratorium has been placed on their development.
Linoleic3 Linolenic Acid
Linoleic acid, an essential fatty acid metabolized to
dihomo-␥-linolenic acid (GLA), serves as an important
constituent of neuronal membrane phospholipids and as
a substrate for prostaglandin E formation, which may be
important for preservation of nerve blood flow. In dia-
betic patients, conversion of linoleic acid to GLA and sub-
sequent metabolites is impaired, possibly contributing to
the pathogenesis of diabetic neuropathy. A recent multi-
center, double-blind, placebo-controlled trial using GLA
for 1 year revealed significant improvements in both clin-
ical measures and electrophysiologic test results.50 These
findings are particularly encouraging when one considers
that linolenic acid also has lipid-lowering effects and may
be useful in the management of diabetes in general.
Advanced Glycosylation End Products
Advanced glycosylation end product (AGE) inhibition
may be of some value in the treatment of diabetic neu-
ropathy. Studies in rats with streptozocin-induced diabe-
tes using the AGE inhibitor aminoguanidine show im-
provement in nerve conduction velocity.51,52 Controlled
clinical trials are needed to determine its efficacy in hu-
mans.
August 30, 1999
N-acetyl-L-carnitine
N-acetyl-L-carnitine has yielded some exciting informa-
tion on potential therapeutic benefit.18 However, the
early promise for this compound was not fulfilled when
the essentially negative results of the multicenter Cana-
dian and American study were reported.
myo-Inositol
myo-Inositol supplements appear to improve neuropa-
thy,53 but treatment may have to be prolonged for ⱖ6
months to achieve a significant effect.
Human Intravenous Immunoglobulin
Human intravenous immunoglobulin (IVIg) treatment
has the potential to ameliorate disturbances in both non-
neurologic and neurologic autoimmune diseases. Kren-
del et al6 reported significant neurologic improvement in
21 patients with either peripheral diabetic polyneurop-
athy or a demyelinating form of DSPN who were treated
with various immunotherapies. Patients with peripheral
diabetic polyneuropathy showed favorable responses to
IVIg, whereas those with DSPN improved during a
course of treatment with steroids.
Treatment with immunoglobulin is well tolerated and
safe. The major toxicity of IVIg has been an anaphylactic
reaction, but this is uncommon and occurs mainly in
patients with immunoglobulin (usually immunoglobulin
A [IgA]) deficiency. In patients who do not have hypo-
gammaglobulinemia or disorders associated with IgA de-
ficiency, screening for IgA deficiency before IVIg is prob-
ably not justified.
Neurotrophic Therapy
NGF, discussed by Apfel53a elsewhere in this supplement,
is one of several neurotrophic factors that play an impor-
tant role in the development, maintenance, and survival
of neuronal tissues. Recombinant NGF administration
restores these neuropeptide levels toward normal and
prevents manifestations of sensory neuropathy in ani-
mals.54
AUTONOMIC NEUROPATHY
The development of autonomic neuropathy is of partic-
ular significance. Cardiac function is markedly impaired,
exercise tolerance is limited, and the ability to withstand
heat is impaired, making it essential that the patient be
protected from temperature extremes.
Diabetic patients with proteinuria have a severalfold
increase in mortality risk as compared with nondiabetic
patients. Abnormalities in well-established cardiovascu-
lar risk factors (e.g., dyslipidemia, hypertension, smok-
ing, and obesity) cannot fully account for this finding.55
Patients with type 1 or type 2 diabetes and albuminuria
have a less pronounced nocturnal decrease in blood pres-
sure56,57 compared with nondiabetic persons. The prev-
THE AMERICAN JOURNAL OF MEDICINE威 Volume 107 (2B) 21S
 A Symposium: Diabetic Neuropathy—Pathogenesis and Therapy/Vinik
alence of sympathetic and parasympathetic autonomic
insufficiency is increased, indicating a possible role of
persistently high cardiac output and total peripheral re-
sistance in the pathogenesis of cardiovascular events that
occur in this high-risk diabetic population.58 Reduced
vagal activity may be an important factor in the patho-
physiology of sudden cardiovascular death, and silent
myocardial ischemia/infarction occurs with loss of sym-
pathetic innervation. These findings explain the poor
prognosis in diabetic patients with autonomic failure.
Patients at increased risk for cardiovascular complica-
tions secondary to autonomic neuropathy may be iden-
tified by examining heart rate variability response to deep
breathing, Valsalva maneuver, and going from a lying
down to a standing position. Unfortunately, there is not
yet an accepted treatment for this condition.
Postural Hypotension
Postural hypotension in the patient with diabetic auto-
nomic neuropathy can present a difficult management
problem. Elevating the blood pressure in the standing
position must be balanced against preventing hyperten-
sion in the supine position.
Attempts should be made to increase venous return
from the periphery using total-body stockings. The pa-
tient should be instructed to put them on while lying
down and to not remove them until returning to the su-
pine position. They can be uncomfortable, especially in
hot weather, making compliance an issue. In severe cases,
an Air Force antigravity suit may be needed.
Some patients with postural hypotension may benefit
from treatment with 9-␣-fluorohydrocortisone, which
increases plasma volume. This mineral corticoid is usu-
ally started at low doses of 50 –200 ␮g at night, but may be
increased up to 0.4 mg qd, although at these doses there is
a significantly greater risk of developing hypokalemia and
excessive fluid retention. Supplementary salt intake may
also be beneficial. These treatments must be used with
caution, because they may cause edema, and there is a
significant risk of congestive heart failure and hyperten-
sion. Various adrenergic agonists may be used instead;
for example midodrine, an ␣1-adrenergic agonist may be
used in doses of 2.5–10 mg tid. Dihydroergotamine in
combination with caffeine, indomethacin, and the ␣2-
adrenergic antagonist yohimbine may also be tried in re-
fractory patients.
Gastropathy
The first step in management of diabetic gastroparesis
consists of multiple small feedings.33 The amount of di-
etary fat should be decreased because it tends to delay
gastric emptying. Metoclopramide, which sensitizes tis-
sues to the actions of acetylcholine, stimulating the mo-
tility of the upper gastrointestinal tract, may be added, if
necessary. It is often given in 10-mg doses about 30 min-
utes before meals. The drug will not be well absorbed,
22S August 30, 1999 THE AMERICAN JOURNAL OF MEDICINE威 Volume 107 (2B)
however, if large gastric residuals secondary to severe gas-
troparesis are present. If so, gastric suctioning and intra-
venous nutrition, as well as intravenous drug administra-
tion, may be needed initially. Other medications, such as
bethanechol and the pyridostigmine, have been advo-
cated but generally are of little therapeutic benefit and
may cause side effects including dry mouth, blurred vi-
sion, and colic.
Cisapride is effective in some patients and may be ini-
tiated at a dose of 10 mg at least 15 minutes before meals
and at bedtime. If medications fail and severe gastropare-
sis persists, jejunostomy placement into normally func-
tioning bowel may be needed. Feedings can be given at
night, freeing patients during the day, and insulin can
accordingly be adjusted to accommodate their feeding
schedule.
Enteropathy
Enteropathy involving the small bowel and colon can
produce both chronic constipation and explosive dia-
betic diarrhea.33 Stasis of bowel contents with bacterial
overgrowth may contribute to the diarrhea. Treatment
with broad-spectrum antibiotics such as ampicillin or
tetracycline (250 mg every 8 hours for either drug) is the
mainstay of therapy. Metronidazole (500 mg every 6
hours) may also be effective. Treatment may be empiric
or, when possible, a hydrogen breath test may be used to
determine whether the cause of the diarrhea is bacterial
overgrowth and whether the overgrowth has been con-
trolled.
Retention of bile may occur and can be highly irritating
to the gut. Chelation of bile salts with cholestyramine
mixed with fluid may offer relief of symptoms. Diphen-
oxylate and atropine (Lomotil; Searle, Chicago, IL) may
help to control the diarrhea. However, toxic megacolon
can occur; extreme care should be used.
Patients with poor digestion may benefit from a glu-
ten-free diet. Certain fibers can lead to bezoar formation
because of bowel stasis in gastroparetic or constipated
patients. The soluble fiber psyllium can be useful for
diarrhea.
Cystopathy
Patients with neurogenic bladder may not be able to sense
when their bladders are full. They should be instructed to
palpate their bladders. If unable to initiate micturition
when the bladder is full, Crede’s maneuver—manual
squeezing of the bladder from all sides in a downward
motion so that urine is expressed—should be used to
start the flow of urine. Parasympathomimetic agents such
as bethanechol are sometimes helpful but often do not
help to fully empty the bladder.33 The patient is given 5
mg or 10 mg initially and then hourly until a satisfactory
response has been achieved or a maximum dose of 50 mg
has been administered. The total dose is then used as a
maintenance dose three or four times a day. Extended
 A Symposium: Diabetic Neuropathy—Pathogenesis and Therapy/Vinik

Table 2. Responses of Specific Pain Syndromes in Diabetic Neuropathy

Syndrome
Paresthesias, dysesthesias,
lancinating
Superficial burning, allodynia
Pain of mononeuropathies
and focal neuropathies
TCAs ⫽ tricyclic antidepressants.
Treatment Side Effects
TCAs ⫹ fluphenazine TCAs: Drowsiness, fatigue,
lethargy, dry mouth,
blurred vision, urinary
retention orthostasis;
fluphenazine: shakiness
Topical capsaicin Care with application to
sensitive areas
Gabapentin or Dizziness, nausea, skin rash,
carbamazepine marrow toxicity,
leukopenia, idiosyncratic
reactions

sphincter relaxation can be achieved with an ␣1 blocker
such as doxazosin. Self-catheterization can be helpful,
and the risk of infection is generally low. In male patients,
bladder neck surgery may help to relieve spasm of the
internal sphincter if ␣1 blockade fails.
Sexual Dysfunction
For the patient, sexual dysfunction can be one of the most
disturbing complications of diabetes. A thorough assess-
ment of sexual function is useful to determine whether a
psychogenic or vascular component is involved in its
pathogenesis. Several physiologic and pharmacologic
treatment modalities are available.33
Sildenafil, at a dose of 25–100 mg, can be taken orally
about 1 hour before anticipated intercourse. Men with
underlying vascular or cardiac disease should use the
drug with caution as sildenafil has serious potential for
interacting with nitrates. Sildenafil is effective in ⱕ50% of
diabetic patients; for those in whom it is not effective,
other agents may soon be available. Intrapenile injection
of regitine and papaverine have been used with some suc-
cess,59 but they carry a risk of infection, priapism, and
fibrosis. Inflatable devices that obstruct venous outflow
while promoting inflow have also been used successfully.
If these therapies do not help, rigid and semirigid pros-
theses are available.
CONTROLLING THE PAIN OF DIABETIC
NEUROPATHY
Control of pain is one of the most difficult management
issues in diabetic neuropathy. Many different pain syn-
dromes have been described, reflecting pathology at dif-
ferent levels of the neuroaxis, including nerve terminals,
the length of the axon, abnormal activity at central syn-
apses in the spinal cord, abnormal activation of periph-
eral sympathetic neurons, and abnormalities in central
pain regions of the brain. The various medications that
are commonly used are active at these different anatomic
sites. There is no single correct approach to the manage-
ment of any given patient with painful neuropathy. Often
it requires patience on the part of the patient and physi-
cian who must try a variety of different medications on a
trial-and-error basis until a satisfactory regimen is estab-
lished.
The syndrome of acute painful neuropathy is a distinct
variant of diabetic polyneuropathy, which is character-
ized by the onset of severe, often unrelenting pain in the
legs. This pain is usually worse at night. In contrast to the
severe painful symptoms, the neurologic examination is
often unimpressive, with only moderate loss of small-
fiber sensory modalities. This syndrome is characteristi-
cally associated with marked weight loss and poor control
of hyperglycemia, although it can be precipitated by a
sudden improvement in glycemic control. Symptoms
generally improve with persistent control of hyperglyce-
mia, although it can take 6 –18 months before significant
improvement is achieved. Pharmacologic treatment of
acute painful neuropathy is not always effective, although
many of the same drugs used have been effective to some
degree in other types of painful neuropathy (Table 2).
These agents will be reviewed in detail below. Nonphar-
macologic therapies that have been tried with limited suc-
cess include sympathectomy, spinal cord blockade, and
electrical spinal cord stimulation.
Analgesics are rarely effective in the treatment of pain-
ful diabetic neuropathy, although they may be of some
use on a short-term basis for self-limited syndromes, such
as painful diabetic third-nerve palsy, or for chronic pain-
ful neuropathy associated with musculoskeletal and joint
abnormalities. Analgesic use should be restricted to non-
steroidal anti-inflammatory drugs (NSAIDS), because
narcotics are addicting and induce constipation, which
may exacerbate features of the autonomic neuropathy.
Clinical trials have demonstrated efficacy with ibuprofen
(600 mg qd) and sulindac (200 mg bid) in relieving neu-
ropathic pain.60 NSAIDs must be used cautiously in dia-
betic patients, however, because of the risk of nephrotox-

August 30, 1999 THE AMERICAN JOURNAL OF MEDICINE威 Volume 107 (2B) 23S
 A Symposium: Diabetic Neuropathy—Pathogenesis and Therapy/Vinik
icity in patients who may already have compromised kid-
neys.
Tricyclic antidepressants used in combination with or
without the phenothiazine fluphenazine have been
shown to be effective in treating painful diabetic neurop-
athy.61,62 In a meta-analysis of 21 different clinical trials,
tricyclics provided significant pain relief to 30% of pa-
tients with neuropathic pain.63 Tricyclics act on the cen-
tral nervous system, preventing the reuptake of norepi-
nephrine and serotonin at synapses involved in pain in-
hibition. The benefits of tricyclics appear to be unrelated
to relief of depression. The most commonly used tricyclic
antidepressants for the treatment of neuropathic pain are
amitriptyline and nortriptyline. There is no evidence to
suggest that one is more efficacious than the other; how-
ever, nortriptyline may have fewer anticholinergic side
effects such as dysautonomia, sleepiness, and dry mouth.
Initial doses are usually low (10 or 25 mg at night) and
then gradually increased until some relief is attained or
until side effects become prohibitive. Precautions should
be taken to avoid precipitating cardiac arrhythmias, uri-
nary retention, and glaucoma, especially in older pa-
tients.
Anticonvulsants have also been used to control neuro-
pathic pain. Diphenylhydantoin (DPH) has been used in
the past; however, double-blind crossover studies have
not demonstrated a therapeutic benefit in diabetic neu-
ropathy.64,65 In addition, DPH has been shown to sup-
press insulin secretion, resulting in the precipitation of
hyperosmolar diabetic coma. Carbamazepine has also
been widely used and has been shown to be efficacious in
clinical trials.66,67 When initiating carbamazepine, it is
advisable to begin with a low dose, 100 mg bid, and then
increase gradually until side effects are encountered or
until there is significant relief of symptoms. Complete
blood counts should be checked at the outset and then on
a frequent basis over the first 3 months because leukope-
nia is a common complication. Gabapentin is a relatively
new anticonvulsant that has been used with increasing
frequency for the relief of neuropathic pain. Gabapentin
has demonstrated significant efficacy in a recent multi-
center, double-blind, randomized, placebo-controlled
study of 165 patients with type 1 and type 2 diabetes.68,69
Gabapentin significantly decreased the mean pain score
and improved the profile of mood states and quality of
life. Adverse effects were, however, significant. Dizziness
was reported in 24%, somnolence in 23%, headache in
11%, and diarrhea in 11% of patients. Gabapentin should
be started at a dose of 300 mg per day, and gradually
increased until either efficacy or until a maximum dose of
2,400 mg per day is reached.
Marked hyperesthesia with burning, lancinating, dys-
esthetic pain is typical of C-fiber pain and may respond to
capsaicin. Capsaicin is extracted from chili pepper and
can be prepared at home by mixing 3 teaspoons of cay-
24S August 30, 1999 THE AMERICAN JOURNAL OF MEDICINE威
enne pepper with a jarful of cold cream. The mixture is
applied to the painful area. Capsaicin works by depleting
the neuropeptide substance P, which is involved in pain
transmission. Several studies have demonstrated efficacy
with capsaicin in treating neuropathic pain,70,71 although
it is still somewhat controversial.72 Capsaicin is applied at
a dose of 0.075% four times a day to the skin over the
painful areas. Neuropathic pain may frequently worsen
for several days after therapy is initiated before improv-
ing.
Lidocaine provides relief of intractable pain for 3–21
days. This therapy is most useful in self-limited forms of
neuropathy. If successful, therapy may be continued with
oral mexiletine, a structural analogue of lidocaine. Mexi-
letine has demonstrated efficacy in several clinical
trials.73–75 It is initiated with doses of 150 mg daily and
then increased weekly until the pain is improved or until
a maximum of 600 –900 mg a day is achieved.
Although there is no single correct way to treat every
patient with painful neuropathy, there are some guide-
lines that might be useful for the physician to follow.
Many initiate treatment with either a tricyclic or an anti-
convulsant. The choice of which particular drug to use
within these general classes of agents may depend on the
individual side-effect profiles. These drugs should be in-
creased until the maximum dose or until side effects be-
come prohibitive before trying other therapies. Some-
times a combination of an anticonvulsant plus a tricyclic
is effective. If this regimen is ineffective and the pain is
burning and restricted to particular anatomic sites or re-
flects allodynia (a marked hypersensitivity of the skin), it
is reasonable to add capsaicin to the regimen. If the pain is
still refractory to therapy, then other drugs discussed
above may be tried as well as nonpharmacologic treat-
ments.
CONCLUSION
It was once said that all we could do for patients with
diabetic neuropathy was commiserate with them. We
now can offer more hope. Some forms of pain respond to
drugs, but better options are needed. Proximal neuropa-
thies are likely to be inflammatory or vasculitic and gen-
erally respond to immunotherapy. A large, definitive
clinical trial of the various forms of immunotherapy ver-
sus placebo therapy is needed. Distal symmetric polyneu-
ropathy remains a therapeutic challenge, but ongoing tri-
als, particularly those targeted at specific nerve fiber def-
icits, hold great promise for arrest and even reversal of the
disorder.
REFERENCES
1. American Diabetes Association, American Academy of
Neurology. Consensus statement: report and recommen-
dations of the San Antonio conference on diabetic neurop-
athy. Diabetes Care. 1988;11:592–597.
2. Brownlee M. Advanced products of nonenzymatic glyco-
Volume 107 (2B)
 A Symposium: Diabetic Neuropathy—Pathogenesis and Therapy/Vinik
sylation and the pathogenesis of diabetic complications. In:
Rifkin H, Porte D, eds. Diabetes Mellitus: Theory and Prac-
tice. New York: Elsevier Science, 1990:279 –291.
3. Diabetes Control and Complications Trial Research Group.
Effect of intensive diabetes treatment on nerve conduction
in the Diabetes Control and Complications Trial. Ann Neu-
rol. 1995;38:869 – 880.
4. Vinik AI, Newlon PG, Lauterio TJ, et al. Nerve survival and
regeneration in diabetes. Diabetes Rev. 1995;3:139 –157.
5. Said G, Goulon-Goeau C, Lacroix C, Moulonguet A. Nerve
biopsy findings in different patterns of proximal diabetic
neuropathy. Ann Neurol. 1994;35:559 –569.
6. Krendel DA, Costigan DA, Hopkins LC. Successful treat-
ment of neuropathies in patients with diabetes mellitus.
Arch Neurol. 1995;52:1053–1061.
7. Vinik AI, Pittenger GL, Milicevic Z, Knezevic-Cuca J. Auto-
immune mechanisms in the pathogenesis of diabetic neu-
ropathy. In: Eisenbarth G, ed. Molecular Mechanisms of
Endocrine and Organ Specific Autoimmunity. Austin, TX:
R.G. Landes Company, 1998:217–251.
8. Vinik AI, Milicevic Z. Preventive measures and treatment
options for diabetic neuropathy. Contemp Intern Med.
1994;16:41– 42,47–55.
9. Vinik AI, Leichter SB, Pittenger GL, et al. Phospholipid and
glutamic acid decarboxylase antibodies in diabetic neurop-
athy. Diabetes Care. 1995;18:1225–1232.
10. Malik RA, Tesfaye S, Thompson SD, et al. Transperineurial
capillary abnormalities in the sural nerve of patients with
diabetic neuropathy. Microvasc Res. 1994;48:236 –245.
11. Dyck P, Hansen S, Karnes J, et al. Capillary number and
percentage closed in human diabetic sural nerve. Proc Natl
Acad Sci USA. 1985;82:2513–2517.
12. Low P, Lagerlund T, McManis P. Nerve blood flow and
oxygen delivery in normal, diabetic, and ischemic neurop-
athy. Int Rev Neurobiol. 1989;31:355– 438.
13. Yasuda H, Dyck P. Abnormalities of endoneurial microves-
sels and sural nerve pathology in diabetic neuropathy. Neu-
rology. 1987;37:20 –28.
14. McNeil HP, Chesterman CN, Krilis SA. Immunology and
clinical importance of antiphospholipid antibodies. Adv Im-
munol. 1991;49:193–280.
15. Tuck RR, Schmelzer JD, Low PA. Endoneurial blood flow
and oxygen tension in the sciatic nerves of rats with exper-
imental diabetic neuropathy. Brain. 1984;107:935–950.
16. Newrick PG, Wilson AJ, Jakubowski J, et al. Sural nerve
oxygen tension in diabetes. BMJ. 1986;293:1053–1054.
17. Zochodne DW, Ho LT. Normal blood flow but lower oxygen
tension in diabetes of young rats: microenvironment and
the influence of sympathectomy. Can J Physiol Pharmacol.
1992;70:651– 659.
18. Hotta N, Koh N, Sakakibara F, et al. Effect of propionyl-L-
carnitine on motor nerve conduction, autonomic cardiac
function, and nerve blood flow in rats with streptozotocin-
induced diabetes: comparison with an aldose reductase
inhibitor. J Pharmacol Exp Ther. 1996;276:49 –55.
19. Rich KM, Luszczynski JR, Osborne PA, Johnson EM Jr.
Nerve growth factor protects adult sensory neurons from
cell death and atrophy caused by nerve injury. J Neurocy-
tol. 1987;16:261–268.
20. Merimee TJ, Gardner DF, Zapf J, Froesch ER. Effect of
glycemic control on serum insulin-like growth factors in
diabetes mellitus. Diabetes. 1984;33:790 –793.
21. Bowsher RR, Lee W-H, Apathy JM, et al. Measurement of
insulin-like growth factor-II in physiological fluids and tis-
sues, I: an improved extraction procedure and radioimmu-
August 30, 1999
noassay for human and rat fluids. Endocrinology. 1991;128:
805– 814.
22. Timpl R, Rohde H, Robey PG, et al. Laminin—a glycopro-
tein from basement membranes. J Biol Chem. 1979;254:
9933–9937.
23. Timpl R. Structure and biological activity of basement
membrane proteins. Eur J Biochem. 1989;180:487–502.
24. Baron–Van Evercooren A, Kleiman HK, Ohno S, et al. Nerve
growth factor, laminin and fibronectin promote neurite
growth from human fetal sensory ganglion cultures. J Neu-
rosci Res. 1982;8:179 –193.
25. Lander AD, Fujii DK, Reichardt LF. Laminin is associated
with the “neurite outgrowth-promoting factors” found in
conditioned media. Proc Natl Acad Sci USA. 1985;82:
2183–2187.
26. Longo FM, Hayman EG, Davis GE, et al. Neurite-promoting
factors and extracellular matrix components accumulating
in vivo within nerve regeneration chambers. Brain Res.
1984;309:105–117.
27. Manthorpe M, Engvall E, Ruoslahti E, et al. Laminin pro-
motes neuritic regeneration from cultured peripheral and
central neurons. J Cell Biol. 1983;97:1882–1890.
28. Rivas RJ, Burmeister DW, Goldberg DJ. Rapid effects of
laminin on the growth cone. Neuron. 1992;8:107–115.
29. Le Beau JM, Liuzzi FJ, Depto AJ, Vinik AI. Differential
laminin gene expression in dorsal root ganglion neurons
and non-neuronal cells. Exp Neurol. 1994;127:1– 8.
30. Le Beau JM, Liuzzi FJ. Laminin B2 mRNA is up-regulated in
sensory neurons and Schwann cells during peripheral
nerve regeneration. Soc Neurosci Abstr. 1991;17:1500.
31. Le Beau JM, Liuzzi FJ, Depto AJ, Vinik AI. Up-regulation of
laminin B2 gene expression in dorsal root ganglion neurons
and non-neuronal cells during sciatic nerve regeneration.
[Brief communication.] Exp Neurol. 1995;134:150 –155.
32. Depto AS, Le Beau JM, Liuzzi FJ, Vinik AI. Laminin gene
expression in diabetic rat dorsal root ganglia. Soc Neurosci
Abstr. 1993;19:835.
33. Vinik AI, Holland MT, LeBeau JM, et al. Diabetic neuropa-
thies. Diabetes Care. 1992;15:1926 –1975.
34. Dawson DM. Entrapment neuropathies of the upper ex-
tremities. N Engl J Med. 1993;329:2013–2018.
35. Pirart J. Diabetes mellitus and its degenerative
complications: a prospective study of 4,400 patients ob-
served between 1947 and 1973. Diabetes Care. 1978;1:
168 –188.
36. Diabetes Control and Complications Trial Research Group.
The effect of intensive treatment of diabetes on the devel-
opment and progression of long-term complications in in-
sulin-dependent diabetes mellitus. N Engl J Med. 1993;
329:977–986.
36a.Parry GJ. Management of diabetic neuropathy. Am J Med.
1999;107:27S–33S.
37. Pittenger GL, Liu D, Vinik AI. The apoptotic death of neu-
roblastoma cells caused by serum from patients with insu-
lin-dependent diabetes and neuropathy may be Fas-medi-
ated. J Neuroimmunol. 1997;76:153–160.
38. Fagius J, Jameson S. Effects of aldose reductase inhibitor
treatment in diabetic polyneuropathy: a clinical and neuro-
physiological study. J Neurol Neurosurg Psychiatry. 1981;
44:991–1001.
39. Handelsman D, Turtle J. Clinical trial of an aldose reductase
inhibitor in diabetic neuropathy. Diabetes. 1981;30:459 –
464.
40. Culebras A, Alio J, Herrera JL, Lopez-Fraile MI. Effect of an
aldose reductase inhibitor on diabetic peripheral
THE AMERICAN JOURNAL OF MEDICINE威 Volume 107 (2B) 25S
 A Symposium: Diabetic Neuropathy—Pathogenesis and Therapy/Vinik
neuropathy: preliminary report. Arch Neurol. 1981;38:133–
134.
41. Daniele E, Coco MP. Tolrestat in the therapy of diabetic
peripheral neuropathy: Is this drug really useful? Clin Ther.
1995;146:793–799.
42. Judzewitsch R, Jaspan J, Polonsky J. Aldose reductase
inhibition improves nerve conduction velocity in diabetic
patients. N Engl J Med. 1983;308:119 –125.
43. Young R, Ewing D, Clarke B. A controlled trial of sorbinil, an
aldose reductase inhibitor, in chronic painful diabetic neu-
ropathy. Diabetes. 1983;32:938 –942.
44. Christensen J, Varnek L, Gregersen G. The effect of an
aldose reductase inhibitor (sorbinil) on diabetic neuropathy
and neural function of the retina: a double-blind study. Acta
Neurol Scand. 1985;71:164 –167.
45. Boulton A. Effects of tolrestat, a new aldose reductase
inhibitor, on nerve conduction and paresthetic symptoms in
diabetic neuropathy. (Abstr.) Diabetologia. 1986;29:521.
46. Abraham R, Wynn V. A double-blind placebo-controlled
trial of mixed gangliosides in diabetic peripheral and auto-
nomic neuropathy. Adv Exp Med Biol. 1984;174:607– 624.
47. Naarden A, Davidson J, Harris L, et al. Treatment of painful
diabetic polyneuropathy with mixed gangliosides. Adv Exp
Med Biol. 1983;174:581–592.
48. Crepaldi G, Fedele D, Tiengo A. Ganglioside treatment in
diabetic peripheral neuropathy: a multicenter trial. Acta Dia-
betol Lat. 1983;10:265–276.
49. Horowitz S. Ganglioside therapy in diabetic neuropathy.
Muscle Nerve. 1986;9:531–536.
50. Keen H, Payan J, Allawi J, et al. Treatment of diabetic
neuropathy with ␥-linolenic acid. Diabetes Care. 1993;16:
8 –15.
51. Miyauchi Y, Shikama H, Takasu T, et al. Slowing of periph-
eral motor nerve conduction was ameliorated by aminogua-
nidine in streptozocin-induced diabetic rats. Eur J Endocri-
nol. 1996;134:467– 473.
52. Schmidt RE, Dorsey DA, Beaudet LN, et al. Effect of ami-
noguanidine on the frequency of neuroaxonal dystrophy in
the superior mesenteric sympathetic autonomic ganglia of
rats with streptozocin-induced diabetes. Diabetes. 1996;
45:284 –290.
53. Gregersen G, Bertelsen B, Harbo H, et al. Oral supplemen-
tation of myoinositol: effects on peripheral nerve function in
human diabetics and on the concentration in plasma, eryth-
rocytes, urine and muscle tissue in human diabetics and
normals. Acta Neurol Scand. 1983;67:164 –172.
53a.Apfel SC. Neurotrophic factors in the therapy of diabetic
neuropathy. Am J Med. 1999;107:34S– 42S.
54. Apfel SC, Kessler JA. Neurotrophic factors in the therapy of
peripheral neuropathy. Baillieres Clin Neurol. 1995;4:593–
606.
55. Nielsen FS, Voldsgaard AI, Gall M-A, et al. Apolipoprotein(a)
and cardiovascular disease in type 2 (non-insulin-depen-
dent) diabetic patients with and without diabetic nephrop-
athy. Diabetologia. 1993;36:438 – 444.
56. Schmitz A, Pedersen MM, Hansen KW. Blood pressure by
24 h ambulatory recordings in type 2 (non-insulin depen-
dent) diabetics. Relationship to urinary albumin excretion.
Diabetic Metab. 1991;17:301–307.
57. Hansen KW, Pedersen MM, Christiansen JS, Mogensen
26S August 30, 1999 THE AMERICAN JOURNAL OF MEDICINE威 Volume 107 (2B)
CE. Diurnal blood pressure variations in normoalbuminuric
type 1 diabetic patients. J Intern Med. 1993;234:175–180.
58. Mann S, Altman DG, Raftery EB, Bannister R. Circadian
variation of blood pressure in autonomic failure. Circulation.
1983;68:477– 483.
59. Shmueli J, Israilov S, Segenreich E, Baniel J, Livne P.
Progressive treatment of erectile dysfunction with intracor-
poreal injections of different combinations of vasoactive
agents. Int J Impot Res. 1999;11:15–19.
60. Cohen KL, Harris S. Efficacy and safety of nonsteroidal
anti-inflammatory drugs in the therapy of diabetic neurop-
athy. Arch Intern Med. 1987;147:1442–1444.
61. Davis J, Lewis S, Gerich J, et al. Peripheral diabetic neu-
ropathy treated with amitriptyline and fluphenazine. JAMA.
1977;238:2291–2292.
62. Gomez Perez F, Rull J, Dies H, et al. Nortriptyline and
fluphenazine in the symptomatic treatment of diabetic
neuropathy: a double-blind cross– over study. Pain. 1985;
23:395– 400.
63. McQuay HJ, Tramer M, Nye BA, Carroll D, Wiffen PJ,
Moore RA. A systematic review of antidepressants in neu-
ropathic pain. Pain. 1996;68:217–227.
64. Ellenberg M. Treatment of diabetic neuropathy with diphen-
ylhydantoin. NY State J Med. 1968;68:2653–2655.
65. Saudek C, Werns S, Reidenberg M. Phenytoin in the treat-
ment of diabetic symmetrical polyneuropathy. Clin Pharma-
col Ther. 1977;22:196 –199.
66. Rull JA, Quibrera R, Gonzalez-Millan H, et al. Symptomatic
treatment of peripheral diabetic neuropathy with carba-
mazepine (Tegretol): double blind crossover trial. Diabeto-
logia. 1969;5:565–568.
67. McQuay H, Carroll D, Jadad AR, Wiffen P, Moore A. Anti-
convulsant drugs for management of pain: a systematic
review. BMJ. 1995;311:1047–1052.
68. Rosenberg JM, Harrell C, Ristic H, et al. The effect of
gabapentin on neuropathic pain. Clin J Pain. 1997;13:251–
255.
69. Backonja M, Beydoun A, Edwards KR, et al. Gabapentin for
the symptomatic treatment of painful neuropathy in pa-
tients with diabetes mellitus: a randomized controlled trial.
JAMA. 1998;280:1831–1836.
70. Capsaicin Study Group. Treatment of painful diabetic neu-
ropathy with topical capsaicin. Arch Intern Med. 1991;151:
2225–2229.
71. Capsaicin Study Group. Effect of treatment with capsaicin
on daily activities of patients with painful diabetic neurop-
athy. Diabetes Care. 1992;15:159 –165.
72. Low PA, Dotson RM. Symptomatic treatment of painful
neuropathy. JAMA. 1998;280:1863–1864.
73. Dejgard A, Petersen P, Kastrup J. Mexiletine for treatment
of chronic painful diabetic neuropathy. Lancet. 1988:8575–
8576.
74. Stracke H, Meyer UE, Schumacher HE, Federlin K. Mexil-
etine in the treatment of diabetic neuropathy. Diabetes
Care. 1992;15:1550 –1555.
75. Oskarsson P, Ljunggren J-G, Lins P-E, the Mexiletine Study
Group. Efficacy and safety of mexiletine in the treatment of
painful diabetic neuropathy. Diabetes Care. 1997;20:1594 –
1597.