DIABETIC NEUROPATHY: CLINICAL MANAGEMENT - PART 7 pdf

facilitate the recruitment of homogeneous study subject populations. Second, correction
or partial correction of abnormalities in regeneration is a concrete target with a clear
clinical interpretation. Furthermore, an improvement in the regeneration rate will pre-
cede an improvement in nerve fiber density and is likely to be a more sensitive meas-
ure. Finally, such a study designed to detect a 50% normalization of regeneration with
80% power in nonneuropathic subjects with diabetes would require about 65 subjects
per treatment arm. A recent trial of Timcodar dimesylate in healthy control subjects
used such measures of collateral and regenerative sprouting as outcome measures.
Although the compound did not accelerate regeneration by either measure, the trial did
demonstrate that such an approach was feasible and the measures were reproducible and
robust (42).
In conclusion, skin biopsy with determination of epidermal nerve fiber density is a
powerful tool that provides investigators insight into a population of nerve fibers that is
prominently affected in diabetes and yet has been relatively under investigated. The super-
ficial nature of epidermal nerve fibers allows repeated sampling of these nerves in a
304 Polydefkis
Fig. 7. For each subject, a regression line from postcapsaicin time-points is generated and the
slope of this line is used as the rate of regeneration. The mean line for each group is shown as a
thick solid line. The rate of regeneration following denervation is 0.177 ± 0.075 fibers per
mm/day for control subjects (red), 0.10 ± 0.07 fibers per mm/day (p = 0.03) for subjects with
diabetes but no neuropathy (green), and 0.04 ± 0.03 fibers per mm/day (p = 0.03) for subjects
with diabetes and neuropathy (blue).
relatively non invasive fashion, and in sites that cannot be assessed through conventional
electrodiagnostical techniques. These features have allowed investigators to diagnose
neuropathy earlier and to define an association between neuropathy and impaired glucose
tolerance. Finally, the ability to injure these fibers in a standardized fashion has led to
novel measures of human axonal regeneration that may provide a more sensitive scale
by which to assess promising regenerative compounds.
REFERENCES
1. Bolton CF, Winkelmann RK, Dyck PJ. A quantitative study of Meissner’s corpuscles in
man. Neurology 1966;16(1):1–9.

2. Dyck PJ, Winkelmann RK, Bolton CF. Quantitation of Meissner’s corpuscles in hereditary
neurologic disorders. Charcot-Marie-Tooth disease, Roussy-Levy syndrome, Dejerine-
Sottas disease, hereditary sensory neuropathy, spinocerebellar degenerations, and heredi-
tary spastic paraplegia. Neurology 1966;16(1):7–10.
3. Dalsgaard CJ, Rydh M, Haegerstrand A. Cutaneous innervation in man visualized with
protein gene product 9.5 (PGP 9.5) antibodies. Histochemistry 1989;92(5):385–390.
4. Kennedy WR, Wendelschafer-Crabb G, Brelje TC. Innervation and vasculature of human
sweat glands: an immunohistochemistry-laser scanning confocal fluorescence microscopy
study. J Neurosci 1994;14(11 Pt 2):6825–6833.
5. Nolano M, Provitera V, Crisci C, et al. Quantification of myelinated endings and
mechanoreceptors in human digital skin. Ann Neurol 2003;54(2):197–205.
6. Lombardi R, Erne B, Lauria G, et al. IgM deposits on skin nerves in anti-myelin-associated
glycoprotein neuropathy. Ann Neurol 2005;57(2):180–187.
7. McArthur JC, Stocks EA, Hauer P, Cornblath DR, Griffin JW. Epidermal nerve fiber
density:normative reference range and diagnostic efficiency. Arch Neurol 1998;55(12):
1513–1520.
8. Kennedy WR, Wendelschafer-Crabb G, Polydefkis M, McArthur JC. Pathology and quan-
titation of cutaneous innervation, in Peripheral Neuropathy. (Dyck PJ, Thomas PK, eds.)
Elsevier Saunders, Philadelphia, 2005, pp. 869–897.
9. Holland NR, Stocks A, Hauer P, Cornblath DR, Griffin JW, McArthur JC. Intraepidermal nerve
fiber density in patients with painful sensory neuropathy. Neurology 1997;48(3):708–711.
10. Polydefkis M, Yiannoutsos C, Cohen B, et al. Reduced intraepidermal nerve fiber density
in HIV-associated sensory neuropathy. Neurology 2002; in press.
11. Rowbotham MC, Yosipovitch G, Connolly MK, Finlay D, Forde G, Fields HL. Cutaneous
innervation density in the allodynic form of postherpetic neuralgia. Neurobiol Dis
1996;3(3):205–214.
12. Kennedy WR, Wendelschafer-Crabb G, Johnson T. Quantitation of epidermal nerves in
diabetic neuropathy. Neurology 1996;47(4):1042–1048.
13. Silos-Santiago I, Molliver DC, Ozaki S, et al. Non-TrkA-expressing small DRG neurons
are lost in TrkA deficient mice. J Neurosci 1995;15(9):5929–5942.

14. Kennedy WR, Nolano M, Wendelschafer-Crabb G, Johnson TL, Tamura E. A skin blister
method to study epidermal nerves in peripheral nerve disease. Muscle Nerve
1999;22(3):360–371.
15. Kennedy WR, Wendelschafer-Crabb G. The innervation of human epidermis. J Neurol Sci
1993;115(2):184–190.
16. McCarthy BG, Hsieh ST, Stocks A, et al. Cutaneous innervation in sensory neuropathies:
evaluation by skin biopsy. Neurology 1995;45(10):1848–1855.
17. Lauria G, Borgna M, Morbin M, et al. Tubule and neurofilament immunoreactivity in
human hairy skin: markers for intraepidermal nerve fibers. Muscle Nerve 2004;30(3):
310–316.
Punch Skin Biopsy in Diabetic Neuropathy 305
18. Polydefkis M, Hauer P, Sheth S, Sirdofsky M, Griffin JW, McArthur JC. The time course
of epidermal nerve fibre regeneration: studies in normal controls and in people with dia-
betes, with and without neuropathy. Brain 2004;127(Pt 7):1606–1615.
19. Lauria G, Morbin M, Borgna M, et al. Vanilloid receptor (VR1) expression in human
peripheral nervous system. J Periph Nerv Sys 2003;8(S1):1–78.
20. Herrmann DN, Griffin JW, Hauer P, Cornblath DR, McArthur JC. Epidermal nerve fiber
density and sural nerve morphometry in peripheral neuropathies. Neurology 1999;53(8):
1634–1640.
21. Kennedy WR, Said G. Sensory nerves in skin: answers about painful feet? Neurology
1999;53(8):1614–1615.
22. Levy DM, Terenghi G, Gu XH, Abraham RR, Springall DR, Polak JM. Immunohistochemical
measurements of nerves and neuropeptides in diabetic skin: relationship to tests of neuro-
logical function. Diabetologia 1992;35(9):889–897.
23. Brown MJ, Martin JR, Asbury AK. Painful diabetic neuropathy. A morphometric study.
Arch Neurol 1976;33(3):164–171.
24. Hsieh ST, Chiang HY, Lin WM. Pathology of nerve terminal degeneration in the skin.
J Neuropathol Exp Neurol 2000;59(4):297–307.
25. Pan CL, Tseng TJ, Lin YH, Chiang MC, Lin WM, Hsieh ST. Cutaneous innervation in
Guillain-Barre syndrome: pathology and clinical correlations. Brain 2003;126(Pt 2):

386–397.
26. Lauria G, Morbin M, Lombardi R, et al. Axonal swellings predict the degeneration of
epidermal nerve fibers in painful neuropathies. Neurology 2003;61(5):631–636.
27. Shun CT, Chang YC, Wu HP, et al. Skin denervation in type 2 diabetes: correlations with
diabetic duration and functional impairments. Brain 2004;127(Pt 7):1593–1605.
28. Sustained effect of intensive treatment of type 1 diabetes mellitus on development and pro-
gression of diabetic nephropathy:the Epidemiology of Diabetes Interventions and
Complications (EDIC) study. Jama 2003;290(16):2159–2167.
29. Singleton JR, Smith AG, Bromberg MB. Increased prevalence of impaired glucose
tolerance in patients with painful sensory neuropathy. Diabetes Care 2001;24(8):
1448–1453.
30. Singleton JR, Smith AG, Bromberg MB. Painful sensory polyneuropathy associated with
impaired glucose tolerance. Muscle Nerve 2001;24(9):1225–1228.
31. Sumner CJ, Sheth S, Griffin JW, Cornblath DR, Polydefkis M. The spectrum of neuropa-
thy in diabetes and impaired glucose tolerance. Neurology 2003;60(1):108–111.
32. Hughes RA, Umapathi T, Gray IA, et al. A controlled investigation of the cause of chronic
idiopathic axonal polyneuropathy. Brain 2004;127(Pt 8):1723–1730.
33. Smith AG, Singleton JR. Peripheral neuropathy and the metabolic syndrome. Annals of
Neurology 2005;58(S9):S31.
34. Knowler WC, Barrett-Connor E, Fowler SE, et al. Reduction in the incidence of type 2 dia-
betes with lifestyle intervention or metformin. N Engl J Med 2002;346(6):393–403.
35. McArthur JC, Yiannoutsos C, Simpson DM, et al. A phase II trial of nerve growth factor
for sensory neuropathy associated with HIV infection. AIDS Clinical Trials Group Team
291. Neurology 2000;54(5):1080–1088.
36. Hart AM, Wilson AD, Montovani C, et al. Acetyl-l-carnitine: a pathogenesis based treat-
ment for HIV-associated antiretroviral toxic neuropathy. Aids 2004;18(11):1549-1560.
37. Pittenger GL, Simmons K, Anandacoomaraswamy D, Rice A, Barlow P, Vinik A.
Topiramate improves intraepidermal nerve fiber morphology and quantitative measures in
diabetic neuropathy patients. J Periph Nerv Sys 2005;10(S1):73.
38. Rajan B, Polydefkis M, Hauer P, Griffin JW, McArthur JC. Epidermal reinnervation after

intracutaneous axotomy in man. J Comp Neurol 2003;457(1):24–36.
306 Polydefkis
39. Hahn K, Brown A, Hauer P, McArthur J, Polydefkis M. Epidermal reinnervation after
mechanical intracutanous axotomy in skin biopsies in normal controls and in people with
HIV. Neurology 2005;64(6 S1):A245–A246.
40. Simone DA, Nolano M, Johnson T, Wendelschafer-Crabb G, Kennedy WR. Intradermal
injection of capsaicin in humans produces degeneration and subsequent reinnervation
of epidermal nerve fibers: correlation with sensory function. J Neurosci 1998;18(21):
8947–8959.
41. Nolano M, Simone DA, Wendelschafer-Crabb G, Johnson T, Hazen E, Kennedy WR.
Topical capsaicin in humans: parallel loss of epidermal nerve fibers and pain sensation.
Pain 1999;81(1–2):135–145.
42. Polydefkis M, Sirdofsky M, Hauer P, Petty BG, Murinson BB, McArthur JC. Factors influ-
encing nerve regeneration in a trial of Timcodar dimesylate. Neurology 2006; 66(2):259–261.
43. Polydefkis M, Hauer P, Griffin JW, McArthur JC. Skin biopsy as a tool to assess distal
small fiber innervation in diabetic neuropathy. Diabetes Technol Ther 2001;3:23–28.
44. Polydefkis M, Griffin JW, McArthur J. New insights into diabetic polyneuropathy. JAMA
2003;290:1371–1376.
Punch Skin Biopsy in Diabetic Neuropathy 307
18
Aldose Reductase Inhibitors for the Treatment
of Diabetic Neuropathy
Aristidis Veves, MD, DSc
SUMMARY
It has been more than 30 years since the first aldose reductase inhibitor (ARI) was tested
in diabetic and galactosemic rats and found to control the polyol accumulation. Since then, a
considerable number of ARIs have been tested in experimental and human diabetes. Despite
the initial encouraging results from tests that were conducted for the past 20 years, ARIs have
not been established for the treatment of diabetic neuropathy yet. The main reasons for this are
inconsistent results and the unacceptable high rate of side-effects associated with the initially

tested compounds. The lack of well-defined end points and the inability to produce an inhibitor
that achieves satisfactory tissue penetration and enzyme inhibition are other major contributing
factors for this failure. This chapter focuses on the clinical trials that have examined the effect of
all tested ARIs on human diabetic neuropathy.
Key Words: Aldose reductase inhibitors; clinical trials; human diabetic neuropathy; efficacy;
side effects; clinical use.
INTRODUCTION
It has been more than 30 years since the first aldose reductase inhibitor (ARI) was
tested in diabetic and galactosemic rats and found to control the polyol accumulation
(1). Since then, a considerable number of ARIs have been tested in experimental and
human diabetes. However, very few new information have become available since the
last edition of this book, probably an indication that either interest in these compounds
is waning down in the scientific community or that despite all the intensive efforts, the
ideal compound that will offer satisfactory enzyme inhibition with minimal side-effect
has not been discovered as yet. A thorough review of work on experimental diabetes
would be out of the spirit of this chapter; however, more information is provided in
chapters of this edition. The following chapter will focus on the results from clinical
trials in diabetic neuropathy (2).
END POINTS FOR CLINICAL TRIALS IN DIABETIC NEUROPATHY
Painful symptoms and foot ulceration are the two most important clinical problems
related to peripheral somatic diabetic neuropathy. The conduction of clinical trials,
which test the efficacy of new therapies for painful neuropathy is straight forward:
From: Contemporary Diabetes: Diabetic Neuropathy: Clinical Management, Second Edition
Edited by: A. Veves and R. Malik © Humana Press Inc., Totowa, NJ
309
patients with this condition are provided trial medication and the primary end point is
the reduction of the symptoms, which is expected to occur during a reasonable period
after the treatment has been initiated. In contrast, foot ulcers develop long after the
initiation of events which lead to nerve damage and, by this time, the possibility of
restoring the nerve lesions, or halting their progression, is close to nonexisting.

Therefore, if a study was to be conducted having as primary end point the prevention
of foot ulceration it should involve patients who have diabetic neuropathy in the early
stages and follow them until they reach the very late stages of the disease. This would
mean that a large number of patients should be followed for prolonged periods of
time, even decades, before any conclusion would be reached.
It is obvious from the aforementioned that more practical end points should be used
in order to conduct clinical therapeutic trials, which will be financially supported by the
pharmaceutical industry where efficient development of new medications are of para-
mount importance. In addition, these end points should give an accurate and more
detailed picture about the effects of the treatment on the progression of the disease,
mainly to what extent it can restore the already established lesions.
Sural nerve biopsies were initially considered to be the best method for evaluating
new medications. However, the interpretation of the biopsy results was found to be more
difficult than was originally believed and as a result, nerve biopsies fall out of favor.
Most recent studies have used surrogate measurements, mainly nerve electrophysiolog-
ical measurements and quantitative sensory testing. Regarding the electrophysiological
measurements, Dyck and O’Brian, based on epidemiological data, initially suggested
that a mean change of 2.9 m per second in the combined conduction velocities of the
ulnar, median, and peroneal nerves, or a change of 2.2 m per second in the peroneal
nerve alone should be achieved in order that the results can have a meaningful clinical
significance (4). However, it should be emphasized that the selection of end points is not
the only issue with the design of new clinical trials. Thus, the current prevailing opin-
ion is that future studies will have to include large number of patients and be of long
enough duration (probably around 18–24 months), pay particular attention to the vari-
ability of the end point measurements, and have rigorous quality control in order to allow
the drawing of definite conclusions regarding the efficacy of ARIs in treating diabetic
neuropathy.
CLINICAL TRIALS WITH ARIS
Alrestatin
Alrestatin was the first ARI to be tried in human diabetic neuropathy. In the first,

uncontrolled study conducted in 1981, 10 patients with symptomatic neuropathy were
treated with intravenous infusions of alrestatin for 5 days (5). Although, symptomatic
improvement was noticed in seven patients, objective measurements failed to improve.
Therefore, as the trial was not controlled, a placebo effect accounting for the sympto-
matic improvement cannot be excluded. No adverse effects of alrestatin were noticed in
this trial (Table 1).
The next trial included nine patients with diabetes with severe symptomatic neuro-
pathy, which had necessitated at least one hospital admission before the study (6). The
trial was a single-blind, nonrandomized, placebo-crossover, which lasted for 4 months.
Each patient received the maximum tolerated oral dose for 2 months and was on
placebo for the other two. Subjective improvement was noted by most of the patients
310 Veves
ARIs for the Treatment of Diabetic Neuropathy 311
Table 1
ARIs Trials in Human Neuropathy
Duration of
Authors Design active treatment Results
Alrestatin
Culebras (1981) Uncntr 5 days Symptomatic improvement
Handelsman (1981) sb, nonrmd, co 4 months Symptomatic improvement
Fagious (1981) db, rmd 12 weeks Improvement of symptoms,
VPT and ulnar mcv
Sorbinil
Judzewitsch (1983) db, rmd 9 weeks Improvement of peroneal mcv
and median mcv and scv
Jaspan (1983) sb 3–5 weeks Symptomatic improvement
Young (1983) db, rmd, co 4 weeks Improvement of symptoms
and sural sap
Lewin (1984) db, rmd, co 4 weeks No improvement
Fagious (1985) db, rmd 6 months Improvement of posterior

tibial mcv and ulnar nevre
F wl and dsl
O’Hare (1988) db, rmd 12 months No benefit
Guy (1988) db, rmd 12 months No benefit
Sima (1988) db, rmd 12 months Improvement of symptoms,
sural sap and mfd
Ponalrestat
Ziegler (1991) db, rmd 12 months No benefit
Krentz (1992) db, rmd 12 months No benefit
Tolrestat
Ryder (1986) db, rmd 8 weeks Improvement of median mcv
Boulton (1990) db, rmd 12 months Improvement of paraesthetic
symptoms and peroneal mcv
Macleod (1992) db, rmd 6 months Improvement of VPT, median
and ulnar mcv
Boulton (1992) db, rmd, 12 months Improvement of symptoms,
withdrawal median and peroneal mcv
Giugliano (1993) db, rmd 12 months Improvement of autonomic
measurements and VPT
Giugliano (1995) db, rmd 12 months Improvement of autonomic
measurements and VPT
Didangelos (1999) db, rmd 24 months Improvement of autonomic
measurements
Greene (1999) db, rmd 12 months Increase in small diameter
myelinated fibers
Hotta (2001) db, rmd 12 months Improvement of symptoms,
median fcv and median
F-wave latency
Johnson (2004) db, rmd 12 months Exercise LVEF and cardiac
stroke volume

sb, single blind; db, double-blind; Uncntr, uncontrolled; nonrmd, nonrandomized; rmd, randomized;
co, crossover; mcv, motor nerve conduction velocity; scv, sensory nerve conduction velocity; sap, sensory
action potential; wl, wave latency; dsl, distal sensory latency; VPT, vibration perception threshold; mfd,
myelinated fibre density.
(eight out of nine), but electrophysiological measurements remained virtually unchanged.
The most notable side-effects were nausea, and photosensitivity, which was severe in
two cases.
Around the same time, the most comprehensive trial of alrestatin was conducted.
Thirty patients with long-standing diabetes and mild to moderate neuropathy were stud-
ied in a double-blind, randomized, placebo-controlled trial, which lasted 12 weeks (7).
Symptomatic improvement, reduction of the sensory impairment score, and improve-
ment of vibration perception threshold and ulnar nerve conduction velocity were
noticed, but the rest of the electrophysiological measurements in the median, peroneal,
and sural nerves did not show any significant difference.
The earlier-mentioned studies indicated that treatment with ARIs might be helpful in
treating diabetic neuropathy and also highlighted the need for well-conducted long-term
trials in order to fully explore the potential of this new therapeutic approach. On the
down side, the high incidence of side-effects of alrestatin prohibited its further devel-
opment. This led the way for using some newly discovered compounds such as sorbinil
and tolrestat.
Sorbinil
Sorbinil was the second ARI to be tested in human diabetic neuropathy and a con-
siderable number of studies were conducted during the last decade using this drug.
An early study using sorbinil for the treatment of neuropathy was published in 1983
and included 39 patients with stable diabetes and no clinical symptoms of neuropa-
thy (8). The design of the study was randomized, double-blind, crossover and each
patient received active treatment for 9 weeks. The results showed a small but statis-
tically significant increase of the conduction velocity of the peroneal motor nerve
(0.70 m per second), the median motor nerve (0.66 m per second), and the median
sensory nerve (1.16 m per second) during the treatment with the active drug. Another

important finding was that the increase declined rapidly after cessation of the treat-
ment so that the nerve conduction velocity was similar to pretreatment levels 3 weeks
later. Five patients were withdrawn from the study because of fever and rash, which
were attributed to sorbinil.
In contrast with the previous trial, the ones which followed included mainly patients
with diabetes with symptomatic neuropathy. The first one studied 11 patients with
severely painful neuropathy who failed to respond to conventional treatment with anal-
gesics or tricyclic antidepressants (10). In a single blind design the patients were treated
with sorbinil for 3–5 weeks and the pain relief was measured using a graphic scale.
Marked to moderate pain relief was noted in eight patients usually 3–4 days after being
on treatment, whereas the pain returned to pretreatment levels in seven of the respon-
ders when they stopped taking the drug. The motor and sensory conduction velocities
of the median nerve improved in four patients, whereas the peroneal motor conduction
velocity improved in two patients. It is of interest however, that in four patients who
responded to the treatment the pain was related to proximal motor neuropathy, a condi-
tion, which is thought to be caused by mechanisms not related to polyol accumulation.
No significant side-effects were noted in the 11 patients who finished the study whereas
12th patient who started the study was withdrawn because of rash.
312 Veves
The next study had a double-blind, randomized, placebo-controlled crossover design,
and included 15 patients with painful symptoms, which were present for more than
1 year (10). The patients were observed for 16 weeks but they were on active treatment
for only 4 weeks, either from week 5 to 8 or from 9 to 12. Painful symptoms were
assessed using a standardized symptom score, whereas other measurements included
neurological findings on clinical examination, vibration perception threshold, motor and
sensory nerve conduction velocities, and autonomic system function tests. A significant
number of patients reported improvement of painful symptoms while on the active treat-
ment, but when the pain score was calculated using their diaries no difference was found
between sorbinil and placebo treatment. Significant improvement was also noticed in
the sural sensory potential action, whereas the rest of the electrophysiological measure-

ments remained unchanged. The number of patients who withdrew because of side-
effects (mainly rash and fever) had increased in comparison with the previous study;
four patients in total had an idiosyncratic reaction which resolved rapidly after the dis-
continuation of the drug.
The next trial used the same layout, i.e., double-blind, placebo-controlled crossover,
and included 13 patients with diabetes with chronic symptomatic neuropathy (mean
duration of symptoms 6 years) (11). The duration of treatment with sorbinil was the
same as in the previous trial, 4 weeks out of a total study period of 16 weeks. The pain
intensity was measured using a 100-mm visual analog scale whereas other measure-
ments included vibration perception threshold, motor and sensory conduction veloci-
ties, autonomic function tests, and duration of sleep. In contrast to the previous study,
no difference was found in any parameter, including the severity of neuropathic symp-
toms and the objective measurements of peripheral nerve function. Side-effects were
present only in one patient who took sorbinil in the form of a febrile rash necessitating
his withdrawal from the study.
The aforementioned short-term trials were followed by long-term ones, which exam-
ined the effects of aldose reductase inhibition for periods of 6–12 months. The first
long-term study studied 55 male patients with diabetes with symptomatic neuropathy
for 6 months in a double-blind placebo-controlled parallel group design (12). To avoid
a possible long-term effect of the drug, the authors elected to randomize their patients
to active- and placebo-treatment and to avoid the crossover design. Patients assessment
included clinical examination, neurophysiological measurements, thermal and vibration
perception thresholds, and autonomic system function tests.
No significant improvement was found in the sorbinil-treated group when it was
compared with the placebo group, although three sorbinil-treated patients reported a
marked overall improvement compared with none from the placebo group. When these
three patients were compared with the whole sorbinil-treated group their age was lower
than the mean group age and the neuropathy assessed by electrophysiology was less
severe. All three patients worsened to pretreatment levels when sorbinil was discontinued.
No significant changes were found in the vibration and thermal discrimination thresh-

old. From the electrophysiological measurements improvement was noticed in the
motor posterior tibial nerve conduction velocity (approximately 1.5 m per second),
F-wave latency of the ulnar nerve, and the distal sensory latency of the ulnar nerve. From
the autonomic tests, a significant improvement in the R–R interval variation during deep
ARIs for the Treatment of Diabetic Neuropathy 313
breathing was found in the sorbinil-treated group. The number of patients with serious
side-effects was smaller in this study; only two patients had to be withdrawn from the
study because of rash and lymphadenopathy.
The next long-term study included 31 patients with mild to moderate neuropathy and
lasted for 14 months (including a 2-month run-in period) (13). The study was designed
as double-blind, randomized, placebo-controlled, and two-third of patients were treated
with sorbinil whereas one-third received placebo. Assessments of the patients response
were performed every 3 months and included the measurement of symptoms such
as pain, tingling, and temperature insensitivity using a 100 mm visual analog scale, clin-
ical examination, vibration perception thresholds, electrophysiology, and autonomic
function tests. The results indicated no benefit for the sorbinil-treated patients in any of
the measured parameters. In addition, as similar doses of the drug were used in this trial
and the previous ones, and was accompanied by serum sorbinil levels measurements,
inadequate drug dosage or poor patient compliance could not be held responsible for
the observed discrepancies. Hypersensitivity reactions with fever, rash, and myalgia
occurred in two patients who recovered completely after the drug was discontinued.
No improvement was also found in another double-blind, randomized trial, which
lasted for 12 months and included patients with severe neuropathy with or without
symptoms (14). Thirty nine patients took part in this study and the severity of neuropa-
thy is indicated by the fact that a history of foot ulceration was present in 21 patients.
Efficacy assessments included clinical evaluation, vibration and thermal perception
thresholds, nerve conduction velocities in 12 nerves, and somatosensory-evoked poten-
tials. The results showed no difference in any of the above measurements between
sorbinil and placebo-treated patients, both for the lower and upper extremities, despite
the fact that the arms were less severely affected.

As it can be seen from the aforementioned studies, the beneficial results, which were
initially reported failed to be confirmed in subsequent, better designed, long-term trials.
In an effort to clear the confusion, the next trial used sural nerve biopsies, which allow
more precise evaluation of the therapeutic efficacy (15). This trial included 16 patients
with established peripheral neuropathy and involved subjects undergoing fascicular
sural nerve biopsies of the same limb at the beginning and the end of the study (16). The
design of the trial was double-blind, randomized, placebo-controlled, and lasted
12 months. Additional investigations included clinical neurological assessments, ther-
mal perception thresholds, and electrophysiological measurements. Although both
actively- and placebo-treated groups showed some clinical improvement at the end of
the study this was more pronounced in the sorbinil-treated group. The nonbiopsised sural
nerve of the sorbinil group showed an improvement of 1 µV in the action-potential
amplitude and of 2 m per second in the sensory conduction velocity (2 m per second),
results that were not found in the placebo group.
The analysis of the sural nerve biopsies showed that the sorbitol levels in the sorbinil
group were reduced, indicating a successful aldose reductase inhibition in the nerve
tissue. The myelinated fibers density, the best single histopathological criterion to quan-
tify neuropathy, was similarly reduced at baseline by 50% in both the sorbinil and
placebo groups when they were compared with age-matched nondiabetic subjects. After
12 months of treatment, a significant increase of 33% was found in the sorbinil group,
314 Veves
whereas no difference was noted in the placebo group. The regeneration and remyeli-
nation activity in the sorbinil group was also increased, whereas no change was noticed
in the placebo group. Important changes were also noticed in the degree of paranodal
demyelination, segmental demyelination, and myelin wrinkling. The main importance of
this study lies in the fact that it was the first to demonstrate morphological improvements
in nerve biopsies after long-term aldose reductase inhibition in humans and suggested
that long-term treatment in properly selected patients might be the most beneficial.
A second clinical trial, which used repeated sural nerve biopsies, assessed the
changes in nerve concentrations of alcohol sugars after a 12-month period with sorbinil

treatment (17). Six patients took part in this study and histochemical measurements
showed a significant decrease in nerve sorbitol and fructose levels in the follow-up visit
compared with baseline, whereas the levels of glucose and myo-inositol remained
unchanged. The earlier findings were interpreted by the investigators as indicating that
sorbinil is an effective inhibitor of aldose reductase, but raised doubts about the role of
myo-inositol in the pathogenesis of diabetic neuropathy.
A common factor, present in virtually all the earlier studies which used sorbinil was
the relatively high rate of side-effects. The main adverse reactions were rash, fever, and
lymphadenopathy, which subsided when the drug was discontinued. Nevertheless,
these adverse reactions would make the use of sorbinil for prolonged period of time
in relatively asymptomatic patients unacceptable, and therefore, the compound was
withdrawn.
Ponalrestat
The main characteristic of ponalrestat compared with the previous two drugs was its
safety profile: very few adverse reaction were reported during the preliminary safety tri-
als, making it ideal for long-term usage. These early expectations were soon dashed as
it became apparent that the nerve-tissue concentration levels were probably insufficient
to inhibit aldose reductase. Therefore, it is hardly surprising that the few properly con-
ducted trials with this compound reported negative results, despite some modest
improvements, which were reported in short, preliminary trials (18,19).
An example of a published paper with ponalrestat was that by Ziegler et al. (20), who
reported a randomized, double-blind, placebo-controlled trial of 60 patients with
chronic symptomatic peripheral diabetic neuropathy for 12 months. No difference in
any peripheral nerve function measurements, including electrophysiology, were docu-
mented at the end of the study. As was expected, the drug was well tolerated and no sig-
nificant side-effects were present during the study. Similar results were subsequently
reported by Krentz et al. (21) in a study with almost identical design.
Tolrestat
Tolrestat was the first ARI to be licensed for the treatment of diabetic neuropathy in
certain countries all over the world including Italy, Mexico, and Ireland. Given orally,

tolrestat is rapidly absorbed at a rate of 60–70%. Its plasma half-life is 10 hours and in
clinical practice a dose of 200 mg per day is sufficient to provide satisfactory inhibition
of the aldose reductase for 24 hours. Excretion is mainly through kidneys (70%),
whereas a further 25% of the dose is excreted in the feces.
ARIs for the Treatment of Diabetic Neuropathy 315
In a multicenter, double-blind, randomized, placebo-controlled trial, which lasted for
12 months the efficacy of tolrestat on symptomatic neuropathy was studied in 556 patients
with either type 1 or type 2 diabetes (22). Inclusion criteria were stable or increasing
severity of neuropathic symptoms, and abnormal motor or sensory nerve electrophysio-
logical measurements in at least three of six tested nerves. Patients were randomized to
doses from 50 to 200 mg daily and efficacy assessments included the response of the
painful and paraesthetic symptoms and electrophysiological measurements.
The painful symptoms improved in both the tolrestat- and placebo-treated patients but
the paraesthetic symptoms improved significantly in patients treated with 200 mg tolre-
stat daily over placebo. From the objective measurements, a significant improvement
(up to 2 m per second) was noticed for the tibial and peroneal nerve conduction veloci-
ties when they were compared both with the baseline measurements and with the
placebo-treated group. Improvement in both symptoms and electrophysiological meas-
urements was found in 28% of tolrestat-treated patients, significantly higher when com-
pared with the 5% of the placebo-treated patients who had a similar response. The
adverse reaction profile of the tolrestat was also satisfactory. The only symptom which
occurred more frequently in the tolrestat group was dizziness. Elevation of transaminases
was found in 13 (2.9%) patients with diabetes treated with tolrestat on any dose, but the
transaminases returned to normal levels within 8–16 weeks after the drug was discontin-
ued. There was no evidence of severe liver dysfunction in any of the patients. A small but
significant drop of the blood pressure, up to 7 mmHg in the systolic and 3.4 mm in the
diastolic was also noticed without any consequences. No hypersensitivity reactions sim-
ilar to the ones which were present with other ARIs were noticed.
The same design with the previous study was adopted by a multicentre European study,
which enrolled 190 patients with symptomatic diabetic neuropathy (23). The study lasted

for 6 months and patients were randomized to take either placebo or tolrestat 200 mg per
day. The efficacy analysis included measurements of painful and paraesthetic symptoms,
vibration perception threshold in three sites, and nerve conduction velocities of four
motor and two sensory nerves. No difference in the painful symptoms was found
between the placebo and tolrestat group at the end of the study, although both groups
improved in comparison with baseline measurements. In contrast, a significant improve-
ment of paraesthetic symptoms was noticed in the placebo group compared both with tol-
restat group and with baseline measurements. Regarding the vibration perception
threshold measurements, a significant change in favor of tolrestat-treated patients was
found in one of the three sites it was measured (carpal site, which was located at the dor-
sum of the second metacarpal bone).
Significant increases in the motor conduction velocities in tolrestat-treated patients
were recorded at the median nerve compared both with baseline (2 m per second) and
with the placebo group, and in the ulnar nerve compared with baseline. When the changes
of all motor conduction velocities were combined together, a significant improvement was
found at the end of the study, compared with baseline measurements and with the placebo
group. All the above changes were present only at the end of the study, after
24 weeks of treatment. At the same time, 48% of the tolrestat-treated patients showed
an improvement in three of the four motor nerve conduction velocities, whereas in the
placebo-treated patients similar response was noticed in 28%. No changes in the two
316 Veves
sensory nerve function measurements were present at the end of the study, although the
heart rate in the tolrestat group was slower compared with baseline measurements of the
same group and with the placebo group. Six tolrestat-treated and two placebo-treated
patients were discontinued from the study because of elevated liver enzymes.
A considerable number of patients who took part in the aforementioned studies con-
tinued to take the drug for several years after the studies were completed and were the
cohort of the subsequent trial, which was designed as a randomized, double-blind,
placebo-controlled withdrawal study (24). Thus, 372 patients who had already received
tolrestat for a mean period of 4.2 years were randomly selected either to continue

receiving tolrestat at a dose of 200 or 400 mg or to switch to placebo for 1 year. Another
interesting feature of the design of this trial was the fact that patients were given the
option to change treatment on one occasion after the first three months of the study
without breaking the code and therefore, maintaining the double-blind design of the
trial. The symptom score and the motor conduction velocities of four nerves were used
as end points.
A significant deterioration of the symptom score was noticed at the 24th and 36th
weeks in the placebo group compared with the tolrestat group. However, at the end of the
study, although a small difference still existed between the two groups, it failed to reach
statistical significance. The conduction velocities of three out of the four motor nerves
also deteriorated considerably in the patients who switched to placebo, whereas no
change was noticed in the patients who continued on tolrestat. Thus, in the median nerve
there was a drop of 0.9 m per second, in the ulnar 1.3 m per second, and in the peroneal
0.8 m per second, whereas the mean reduction of both nerves was 0.9 m per second. In
addition, in patients who switched from tolrestat to placebo during the study there was a
mean drop of 1.3 m per second for all four nerves, whereas in the patients who switched
from placebo to tolrestat an improvement of 1 m per second was recorded. Therefore, a
small but significant benefit of long-term treatment with tolrestat which can disappear
when the treatment is discontinued was the main finding of the aforementioned study.
In a parallel study, sural nerve biopsies were obtained at the end of the earlier-
mentioned trial from 13 patients who continued to receive tolrestat and 14 patients who
received placebo (25). Morphometric analysis showed no difference between the above
two groups, but when compared with nerve biopsies from untreated neuropathic
patients both groups showed increased nerve fiber regeneration. In addition, treatment
with tolrestat was found to ameliorate the increase in the sorbitol and fructose levels in
the nerve tissue indicating that tolrestat can achieve satisfactory concentration levels in
the peripheral nerves.
The following two trials with tolrestat were performed at the University of Naples
and were both randomized, placebo-controlled, double-blind, parallel trials of 52 weeks
duration. The first one examined the effect of 200 mg daily tolrestat on patients with

asymptomatic autonomic diabetic neuropathy, defined as at least one abnormal cardio-
vascular reflex (26). At end of the study, improvement in the tolrestat-treated group was
found in all autonomic tests, which included deep breathing (E/I ratio), lying to stand-
ing (30/15) ratio, Valsalva (L/S ratio), and postural hypertension. In contrast to this
improvement, a worsening in all the above parameters except the orthostatic hypoten-
sion was observed in the placebo-treated group. Similar results, namely an improvement
ARIs for the Treatment of Diabetic Neuropathy 317
in the tolrestat group and a worsening in the placebo group, were found in vibration per-
ception threshold measurements, the only reported assessment of the peripheral somatic
nerve function. Similar results were reported in the second study, which included
patients with subclinical neuropathy, defined as abnormality in only one autonomic test,
the squatting test (27). Improvement was found in all the autonomic tests and the vibra-
tion perception thresholds in the tolrestat group, whereas deterioration was observed in
the placebo group in all but the orthostatic hypotension tests.
In one of the last studies that were conducted with tolrestat, patients with diabetes
with clinical autonomic neuropathy were randomized to either tolrestat (200 mg daily)
or placebo for a period of 2 years. As with the earlier studies, treatment with tolrestat
resulted in improvement of most standard cardiovascular reflex test, in comparison with
both the baseline measurements and also to the changes that were observed in the
placebo group (28). Despite the initially promising results, tolrestat was subsequently
withdrawn from clinical use as it was associated to serious side-effects, mainly related
to liver failure.
Zenarestat
Zenarestat was an ARI that was shown to achieve very good penetration in the nerve
tissue. Initial studies indicated that in patients who achieved more than 80% sorbitol
suppression in sural nerve biopsies after a 52-week treatment with agent, there was a
significant increase in the density of small diameter sural nerve myelinated fibers (29).
However, an unfavorable risk/benefit ratio, mainly related to kidney problems, prohib-
ited the conduction of large pivotal studies, and zenarestat was withdrawn.
Fidarestat

In animal models, Fidarestat was shown to be one of the most potent AR inhibitors.
As a result, a large multicenter placebo-controlled randomized study that used 1 mg of
Fidarestat once a day and lasted for 52 weeks was conducted (30). A total of 279 patients
with mild diabetic neuropathy were included. Fidarestat-treated patients showed an
improvement in two out of eight nerve electrophysiological measures (median nerve
FCV and F-wave latency) that were recorded and in subjective symptoms.
Zopolrestat
Zopolrestat is an interesting ARI as there are no published clinical trials indicating
favorable effects on diabetic neuropathy. However, a recently published randomized,
placebo-controlled, double-blind, parallel trial of 52 weeks duration examined the effi-
cacy of 500 or 1000 mg daily in patients with low diastolic peak filling rate or impaired
augmentation of left ventricular ejection fraction (LVEF) and absence of coronary
artery disease, left ventricular hypertrophy, and valvular heart disease (30). Treatment
with either dose of zopolrestat resulted in a small improvement of the exercise LVEF
and stroke volume when compared with the placebo-treated patients. Although, the clin-
ical significance of these results is small, they do indicate that diabetic cardiomyopathy
may not be exclusively related to coronary artery disease, but it might also be associ-
ated to activation of the polyol pathway in the cardiac myocytes. It is also of interest
that no improvement was noticed in the peripheral somatic or autonomic neuropathy in
the patients who participated in the aforementioned study.
318 Veves
CONCLUSION
Despite the initial encouraging results from trials that were conducted during the last
20 years, ARIs have not been established for the treatment of diabetic neuropathy yet.
The main reasons for this are inconsistent results in subsequent trials and the unaccept-
able high rate of side-effects associated with the initially tested compounds. The lack of
well-defined end points and the conduction of numerous small trials, instead of
focussing on the most promising agents and conducting large pivotal trials is one of the
reasons that are related to this outcome. In addition, the inability to produce an inhibitor
that achieves satisfactory tissue penetration and enzyme inhibition, whereas at the same

time is devoid of serious side-effects have also played a significant role. Currently, it
seems that the interest in ARIs is significantly reduced and is doubtful if new ARIs will
be tested clinically in the near future.
REFERENCES
1. Dvornik D, Simard-Duquesne N, Krami M, et al. Polyol accumulation in galactosemic and
diabetic rats: control by an aldose reductase inhibitor. Science 1973;182:1146–1148.
2. Tomlinson DR, Willars GB, Carrington AL. Aldose reductase inhibitors and diabetic com-
plications. Pharmacol Ther 1992;54:151–194.
3. Pfeifer MA, Schumer MP, Gelber DA. Aldose reductase inhibitors: the end of an era or the
need for different trial designs? Diabetes 1997;46(Suppl 2):S82–S89.
4. Dyck PJ, O’Brian PC. Meaningful degrees of prevention or improvement of nerve con-
duction in controlled clinical trials of diabetic neuropathy. Diabetes Care 1989;12:
649–652.
5. Culebras A, Alio J, Herrera JL, Lopez-Fraile IP. Effect of an aldose reductase inhibitor on
diabetic peripheral neuropathy. Arch Neurol 1981;38:133–134.
6. Handelsman DJ, Turtle JR. Clinical trial of an aldose reductase inhibitor in diabetic neu-
ropathy. Diabetes 1981;30:459–464.
7. Fagius J, Jameson S. Effects of aldose reductase inhibitor treatment in diabetic polyneu-
ropathy—a clinical and neurophysiological study. J Neurol Neurosurg Psychiatr 1981;44:
991–1001.
8. Judzewitsch RG, Jaspan JB, Polonsky KS, et al. Aldose reductase inhibition improves
nerve conduction velocity in diabetic patients. N Engl J Med 1983;308:119–125.
9. Jaspan J, Maselli R, Herold K, Bartkus C. Treatment of severely painful diabetic neuropa-
thy with an aldose reductase inhibitor: relief of pain and improved somatic and autonomic
nerve function. Lancet 1983;ii:758–762.
10. Young RJ, Ewing DJ, Clarke BF. A controlled trial of sorbinil, an aldose reductase
inhibitor, in chronic painful diabetic neuropathy. Diabetes 1983;32:938–942.
11. Lewin IG, O’Brien AD, Morgan MH, Corrall RJM. Clinical and neurophysiological stud-
ies with the aldose reductase inhibitor, sorbinil, in symptomatic diabetic neuropathy.
Diabetologia 1984;26:445–448.

12. Fagius J, Brattberg A, Jameson S, Berne C. Limited benefit of treatment of diabetic neu-
ropathy with an aldose reductase inhibitor: a 24-week controlled trial. Diabetologia
1985;28:323–329.
13. O’Hare JP, Morgan MH, Alden P, Chissel S, O’Brien AD, Corrall RJM. Aldose reductase
inhibition in diabetic neuropathy: Clinical and neurophysiological studies of one year’s
treatment with sorbinil. Diabet Med 1988;5:537–542.
14. Guy RJC, Gilbey SG, Sheehy M, Asselman P, Watkins P. Diabetic neuropathy in the
upper limb and the effect of twelve months sorbinil treatment. Diabetologia 1988;31:
214–220.
ARIs for the Treatment of Diabetic Neuropathy 319
15. Consensus Statement. Report and Recommendations of the San Antonio Conference on
Diabetic Neuropathy. Diabetes 1988;37:1000–1004.
16. Sima AAF, Brill V, Nathaniel T, et al. Regeneration and repair of myelinated fibres in sural
nerve biopsy specimens from patients with diabetic neuropathy treated with sorbinil.
N Engl J Med 1988;319:548–555.
17. Dyck PJ, Zimmerman BR, Vilen TH, et al. Nerve glucose, fructose, sorbitol, myo-inositol,
and fiber degeneration and regeneration in diabetic neuropathy. N Engl J Med 1988;319:
542–548.
18. Gill JS, Williams G, Ghatei MA, Hetreed AH, Mather HM, Bloom SR. Effect of the aldose
reductase inhibitor, ponalrestat, on diabetic neuropathy. Diabete Metab 1990;16:296–302.
19. Price DE, Alani SM, Wales JK. Effect of aldose reductase inhibition on resistance to
ischemic conduction block in diabetic subjects. Diabetes Care 1991;14:411–413.
20. Ziegler D, Mayer P, Rathmann W, Gries FA. One-year treatment with the aldose reductase
inhibitor, ponalrestat, in diabetic neuropathy. Diabetes Res Clin Pract 1991;14:63–73.
21. Krentz AJ, Honigsberger L, Ellis SH, Hardman M, Nattrass M. A 12-month randomized
controlled study of the aldose reductase inhibitor ponalrestat in patients with chronic
symptomatic diabetic neuropathy. Diabet Med 1992;9:463–468.
22. Boulton AJM, Levin S, Comstock J. A multicentre trial of the aldose reductase inhibitor,
tolrestat, in patients with symptomatic diabetic neuropathy. Diabetologia 1990;33:
431–437.

23. Macleod AF, Boulton AJM, Owens DR, et al. A multicentre trial of the aldose reductase
inhibitor tolrestat in patients with symptomatic diabetic peripheral neuropathy. Diabete
Metab 1992;18:14–20.
24. Santiago JV, Sonksen PH, Boulton AJM, et al. Withdrawal of the aldose reductase inhibitor
tolrestat in patients with diabetic neuropathy: Effect on nerve function. J Diab Comp
1993;7:170–178.
25. Sima AAF, Greene DA, Brown MB, et al. Effect of hyperglycemia and the aldose inhibitor
tolrestat on sural nerve biochemistry and morphometry in advanced diabetic peripheral
polyneuropathy. J Diab Comp 1993;7:157–169.
26. Giugliano D, Marfella R, Quatraro A, et al. Tolrestat for mild diabetic neuropathy. A 52-
week, randomized, placebo controlled trial. Ann Int Med 1993;118:7–11.
27. Giugliano D, Acampora R, Marfella R, et al. Tolrestat in the primary prevention of diabetic
neuropathy. Diabetes Care 1995;18:536–541.
28. Didangelos TP, Karamitsos DT, Athyros VG, Kourtoglou GI. Effect of aldose reductase
inhibition on cardiovascular reflex tests in patients with definite diabetic autonomic neu-
ropathy over a period of 2 years. J Diab Comp 1998;12:201–207.
29. Greene DA, Arezzo J, Brown M. Effects of aldose reductase inhibition on nerve conduc-
tion and morphometry in diabetic neuropathy. Neurology 1999;53:580–591.
30. Hotta N, Toyota T, Matsuoka K, et al. Clinical efficacy of fidarestat, a novel aldose reduc-
tase inhibitor, for diabetic peripheral neuropathy: a 52-week multicenter placebo-controlled
double-blind parallel group study. Diabetes Care 2001;24:1776–1782.
31. Johnson BF, Nesto RW, Pfeifer MA, et al. Cardiac abnormalities in diabetic patients with
neuropathy: effects of aldose reductase inhibitor administration. Diabetes Care 2004;27:
448–454.
320 Veves
19
Other Therapeutic Agents for the Treatment
of Diabetic Neuropathy
Gary L. Pittenger PhD, Henri Pharson PhD, Jagdeesh Ullal MD,
and Aaron I. Vinik

MD, PhD
SUMMARY
The pathogenesis of diabetic neuropathy is complex and it is important to understand
the underlying pathology leading to the complication in order to best tailor treatment for
each individual patient. It is unlikely that reversing any single mechanism will prove
sufficient for reversing nerve damage. Several drugs, such as antioxidant, PKC
inhibitors and nerve growth factors can have effects on multiple systems that are com-
promised in diabetic neuropathy, yet even those may not be enough in and of themselves
to completely restore neurological function. Combination therapy may prove to be the
best long term approach, and studies of those combinations should prove revealing as
to the relative roles of metabolic dysfunction, microvascular insufficiency and autoim-
munity in the diabetic neuropathy patient population.
Key Words: Antioxidant; PKC inhibitors and nerve growth factors; VEGF.
INTRODUCTION
Neuropathy is one of the most common complication of diabetes with a heterogeneous
clinical presentation and a wide range of abnormalities. As a result, there is still not a
single therapeutic agent for diabetic neuropathy that consistently gives more than mild
relief. There is a plethora of studies in the literature indicating that metabolic, microvas-
cular, and autoimmune dysfunction all play a role in the progression of diabetic neu-
ropathy. Given the potential significance of each of these different systems in diabetes,
it is not surprising that treatments targeted to specific pathways in these systems have
commonly proven disappointing in their efficacy. Our “simple” concept of the various
functional changes leading to peripheral nerve disease in diabetes is presented in (Fig. 1)
(1). Thus, although clinical studies with thousands of patients have shown that hyper-
glycemia is at the center of diabetic neuropathy (2,3), finding specific biochemical
pathways to manipulate for therapy has proven difficult.
From: Contemporary Diabetes: Diabetic Neuropathy: Clinical Management, Second Edition
Edited by: A. Veves and R. Malik © Humana Press Inc., Totowa, NJ
321
END POINTS FOR CLINICAL TRIALS IN DIABETIC NEUROPATHY

Among the challenges for studies of diabetic neuropathy is the selection of the
measures that can be used to determine the efficacy of the agents. Common end points
include:
1. Subjective measures of symptoms (e.g., nerve symptom scores).
2. Questionnaires that allow for quantification of any changes in the symptoms of diabetic
neuropathy, or the quality of life of the patient (e.g., neuropathy symptom scores, total neu-
ropathy scores).
3. Objective neurological examinations focussing on distal sensorimotor function, including
combinations giving a single score (e.g., NTSS-6, a combination of numbness, prickling
sensation, aching pain, burning pain, lancinating pain, and allodynia scores).
4. Electrophysiology (nerve conduction velocity, amplitudes [especially sural nerve ampli-
tude], F-wave latencies).
5. Quantitative tests of sensory and motor modalities that allow more precise measures of
nerve function.
6. Autonomic measures (e.g., cardiac function or neurovascular measures).
7. Skin biopsy, which allows direct observation of the morphology of sensory nerve fibers.
8. Instruments that entail various combinations of the above.
The following summarizes the view of prospective end points in diabetic neuropathy
trials:
322 Pittenger et al.
Fig. 1. Diagram of the pathologies underlying the development of diabetic neuropathy, begin-
ning with hyperglycemia, possibly exacerbated by alcohol or smoking Adapted from ref. 1.
1. Symptoms (forms and questionnaires)—subjective.
2. Quality of life (questionnaire)—subjective.
3. Quantitative neurological examination focussing on distal sensorimotor function (sensation,
strength, and reflexes)—semiobjective.
4. Neurophysiology (nerve conduction velocity, amplitudes, F-wave latencies; multiple nerves,
both sensory and motor)—objective.
5. Quantitative sensory testing (vibration, thermal, pain, and NTSS-6)—semiobjective 6.
Autonomic testing (microvascular, cardiac measures)—objective.

7. Morphology—(skin biopsy—nerve density and length, branching pattern)—objective.
8. Combinations (e.g., NIS-LL + 7 = objective neurological examination of the lower limbs +
4 measures of nerve electrophysiology + QAFT + QST).
Unfortunately, subjective measures may not be sensitive enough to discern small
changes in nerve fiber functions in a small test group during a short period of time.
Unless a study includes a large number of subjects and a sufficient period of time sig-
nificant changes may not be detectable, particularly in the short time periods that most
efficacy trials are designed for. Thus, objective tests may be more useful for studies of
shorter duration.
TREATMENTS
Therapies for diabetic neuropathy are generally directed at the 3 primary mechanisms
shown in Fig. 1: metabolic dysfunction, autoimmunity, or microvascular insufficiency.
Several of these agents, for example, antioxidants or neurotrophins (NTs), can affect
more than one of the pathogenetic mechanisms, as shown in Table 1.
THERAPIES TARGETED TO METABOLIC PATHWAYS
Note: The aldose reductase inhibitors are discussed elsewhere (see Chapter 18).
Antioxidants
Hyperglycemia has been shown in a number of studies to cause oxidative stress in
tissues that are susceptible to the complications of diabetes, including peripheral nerves.
In turn, the oxidative stress leads to the generation of free radicals that can attack the
lipids, proteins, and nucleic acids of the affected tissues directly, compromising physi-
ological function. The end result is the loss of axons and disruption of the microvascu-
lature in the peripheral nervous system (Fig. 2). It has been shown that there is an
increased presence of markers of oxidative stress, such as superoxide and peroxynitrite
ions, and that antioxidant moieties were reduced in patients with diabetic peripheral
Other Therapeutic Agents for the Treatment of Diabetic Neuropathy 323
Table 1
Mechanisms for Pathogenesis of Diabetic Neuropathy and Therapies
That Have Been Tested Addressing Them
Microvascular Nerve

Metabolic dysfunction insufficiency Autoimmunity regeneration
ARIs PKC-β Inhibitor Steroidal anti-inflammatories Neurotrophins
Antioxidants Vasodilators IVIg Topiramate
Inhibitors of glycation Antioxidants Plasmaphoresis –
neuropathy (4). Therefore, it is reasonable to use therapies that are known to reduce
oxidative stress in tissues, and antioxidants. Although a host of antioxidants have been
tested in animal models (for review see ref. 5), those that have been tested in human
studies will be addressed.
α-Lipoic Acid
α-lipoic acid is the best studied antioxidant therapy used in diabetic neuropathy.
Lipoic acid (1,2-dithiolane-3-pentanoic acid), a derivative of octanoic acid, is present in
food and is also synthesized by the liver. It is a natural cofactor in the pyruvate dehy-
drogenase complex where it binds acyl groups and transfers them from one part of the
complex to another. α-lipoic acid, also known as thioctic acid, has generated consider-
able interest as a thiol replenishing and redox modulating agent. In streptozotocin
(STZ)-diabetic rats α-lipoic acid has been shown to prevent slowing of peripheral nerve
conduction velocity and to maintain peripheral nerve blood flow (6–8). It has also been
shown to be effective in ameliorating both the somatic and autonomic neuropathies in
diabetes (9–11). It is not clear that the positive effects are limited to the antioxidant
properties of α-lipoic acid, and studies are ongoing to determine its range of effects.
α-lipoic acid is licensed for use in diabetes in Germany and it is currently undergoing
324 Pittenger et al.
Fig. 2. The interactions of pathways leading to neurovascular and endothelial dysfunction and
the actions of drugs that alter those mechanisms and are being tested in models of neuropathy.
extensive trials in the US as both an antidiabetic agent and for the treatment of diabetic
neuropathy.
γ-Linolenic Acid
Linoleic acid is an essential fatty acid that is metabolized to di-homo-γ-linolenic acid
(GLA), which in turn serves as an important constituent of neuronal membrane phos-
pholipids. In addition it can serve as a substrate for prostaglandin E synthesis, which

may be important for preservation of nerve blood flow. In diabetes, conversion of
linoleic acid to GLA and subsequent metabolites is impaired, possibly contributing to
the pathogenesis of diabetic neuropathy (12,13). A multicenter, double-blind placebo-
controlled trial using GLA for 1 year demonstrated significant improvements in both
clinical measures and electrophysiological testing (14).
Tocopherol (Vitamin E)
The tocopherols, especially the α-tocopherol isoform, have been promoted as effective
antioxidant therapy for a number of neurological diseases, including Alzheimer’s disease,
epilepsy, cerebellar ataxia, and diabetic neuropathy. In studies of patients with diabetes
vitamin E has been shown to decrease 8-isoprostane F2α (15), decrease low density
lipoprotein-C oxidation at high doses (16), and increase skin blood flow and reduce
free radicals in skin with topical application (17). Human studies of vitamin E using com-
bined oral therapy with vitamin C, another well-established antioxidant have shown
improved vascular function, but only in type 1 patients (18). Thus, vitamin E appears to
exert antioxidant protective effects on neurons in diabetes although efficacy has not yet
been demonstrated.
INHIBITORS OF GLYCATION
It is apparent that advanced glycosylation end products (AGEs) contribute to nerve
damage either by direct action on neurons and myelin or by enhancing oxidative stress
under hyperglycemic conditions. Thus, there is a great deal of interest in agents that
either prevent the formation of AGEs or agents that reverse the nonenzymatic glycation
of proteins.
Aminoguanidine
Animal studies using aminoguanidine, an inhibitor of the formation of AGEs,
showed improvement in nerve conduction velocity in rats with STZ-induced diabetic
neuropathy. However, controlled clinical trials to determine its efficacy in humans have
been discontinued because of toxicity (19,20). However, there are compounds related
to aminoguanidine that reduced AGE formation and hold promise for this approach,
although these have not been systematically studied in humans (21–24).
THERAPIES FOR MICROVASCULAR INSUFFICIENCY

Although it is clear that there are significant alterations of blood vessels in diabetes,
data thus far has been conflicting whether neuropathy promotes the changes in the
microvasculature or whether it is the changes in the microvessels of the nerves that lead
to neuropathy. Whatever the case, evidence is growing that re-establishing more normal
patterns of blood flow to the nerves results in improved neurological function.
Other Therapeutic Agents for the Treatment of Diabetic Neuropathy 325
Protein Kinase C Inhibitors
Protein kinase C (PKC) and diacylglycerol (DAG) are intracellular signaling mole-
cules that regulate vasculature by endothelial permeability and vasodilation. The PKC
isozymes are a family of 12 related serine/threonine kinases (25) whose normal func-
tion is the activation of essential proteins and lipids in cells essential for cell survival.
PKC-β is expressed in the vasculature (26,27) and belived to be involved in cell prolif-
eration, differentiation, and apoptosis. PKC is activated by oxidative and osmolar stress,
both of which are a consequence of the dysmetabolism of diabetes. Increased polyol
pathway activity and pro-oxidants bind to the catalytic domain of PKC and it is disin-
hibited. PKC-β overactivation is induced by hyperglycemia or fatty acids through receptor-
mediated activation by phospholipase C. It is hypothesized that AGEs and oxidants
produced by nonenzymatic glycation and the polyol pathway, respectively, increase the
production of DAG (28). Increased DAG and calcium promotes the overactivation of
PKC-β (29). Activation of PKC-β activates MAP kinase and, subsequently, phosphory-
lation of transcription factors that are involved in angiogenesis, increased stress-related
genes, c-Jun kinases and heat shock proteins, all of which can damage cells and vascu-
lar endothelial growth factor (VEGF) (30), which is known to play a critical role in
nerve development (31). Diabetic animal models have shown high levels of PKC-β in a
number of tissues (28), including nerves and endothelium (32). Activation of PKC-β
causes vasoconstriction and tissue ischemia, whereas high levels may impair neuro-
chemical regulation. PKC-β hyperactivity leads to increased vascular permeability,
nitric oxide dysregulation (33), increased leukocyte adhesion (34), and altered blood
flow (35). Furthermore, PKC-β hyperactivity in the neural microvessels causes vaso-
constriction, which might lead to decreased blood flow, resulting in nerve dysfunction

and hypoxia (35). Nerve hypoxia, oxidative nitrosative stress, and an increase in NFκB
causes endothelial damage, leading to depletion of nerve growth factors, VEGF, and
TGF-α autoimmunity and may further accelerate the loss of nerve conduction (31).
Both animal models and human clinical trials investigating complications of diabetes
have shown that blockade of PKC-β slows the progression of complications (33,36–47).
Multiple studies using a specific PKC-β inhibitor, ruboxistaurin mesylate (LY333531),
have shown improvements in diabetic neuropathy. One study in obese rats observed that
ruboxistaurin increased resting nitric oxide concentration, and reduced nitric oxide by
15%, indicating that this action is a PKC-β dependent phenomenon (33). Ruboxistaurin
has been shown to improve nitric oxide-dependent vascular and autonomic nerve dys-
function in diabetic mice (46). In addition to improving nitric oxide levels, ruboxis-
taurin improves nerve function and blood flow. Ruboxistaurin corrected the diabetic
reduction in sciatic endoneurial blood flow, sciatic motor, and saphenous sensory nerve
conduction velocity in diabetic rats (40,43). In another study, the investigators measured
sciatic nerve, superior cervical ganglion blood flow, and nerve conduction velocity in
STZ treated rats. After 8 weeks, the authors observed that diabetes reduced sciatic nerve
and superior cervical ganglion blood flow by 50% and produced deficits in saphenous
nerve sensory conduction velocity (48). After 2 weeks of treatment with ruboxistaurin,
the sciatic nerve, and ganglion blood flow were improved. Additionally, nerve dysfunction
is commonly attributed to alterations of the nerve transporters. Other studies demon-
strated that a specific inhibitor of the PKC-β, (ruboxistaurin), prevents PMA-dependent
326 Pittenger et al.
activation of Na
+
,K
+
-ATPase in rats (44,45). In addition to improvements with blood
flow nerve function and ion transport, ruboxistaurin corrected thermal hyperalgesia
(48,49).
These observations have been supported in preliminary clinical studies. In healthy

humans, ruboxistaurin blocked the reduction in endothelium-dependent vasodilation
induced by acute hyperglycemia (47), suggesting that the hyperglycemic effects on
vasodilation are mediated through PKC-β. More recently, a 1-year double-blind, paral-
lel clinical trial with 205 patients with type 1 or 2 diabetes and DPN was performed to
assess the impact of ruboxistaurin on vibration perception in patients with DPN com-
pared with placebo. In patients with DPN, ruboxistaurin treatment improved symptoms
and vibratory sensation with a significant correlation between the two compared with
placebo group (50). Another recent report indicates that ruboxistaurin is particularly
effective in neuropathy patients with intact sural nerve amplitudes (51). A phase II study
using NTSS-6 to assess the intensity and frequency of sensory neuropathy symptoms
further suggested that ruboxistaurin slows the progression of hyperglycemia-induced
microvascular damage (52). Together, these studies support the belief underlying a role
for PKC-β in the etiology of diabetes-induced neuropathy.
Vascular Endothelial Growth Factor
The most potent stimulus for angiogenesis is VEGF. If the pathogenesis of diabetic
neuropathy goes through loss of vasa nervorum, it is likely that appropriate application
of VEGF would reverse the dysfunction. Normally, VEGF activity is induced by tissue
hypoxia (53,54). In diabetes it is just such hypoxia that results in increased VEGF activ-
ity in the retina, with subsequent pathological angiogenesis (55). Conflicting reports
indicated that VEGF in diabetes goes up (56) and goes down (57). One possibility to
explain this is whether the animals were treated with insulin, which can reduce VEGF
expression (56). Only recently it has been demonstrated that there is a reduction in
VEGF activity in STZ-diabetic mice that results in failure of neovascularization in
hypoxic tissue in the lower limb (57). Furthermore, in the same study it was shown that
intramuscular injection of an adenoviral vector encoding for VEGF could induce nor-
mal neovascularization in the hindlimb. There have been no human studies of VEGF,
and caution is the best approach given the demonstrated pathological effects of VEGF
in the retina in diabetes.
Vasodilators
Microvascular insufficiency, endoneurial blood flow, and hemodynamic factors lead

to nerve damage in patients with DPN (58–62). Although, the sequence of events is not
well understood, investigators propose that microvascular vasoconstriction, edema, and
ischemia play a role in DPN development. Endoneurial edema increases endoneurial
pressure (63), thereby causing capillary closure and subsequent nerve ischemia and
damage (64,65). A diminished regulation of the endoneurial blood flow and ischemia
may result from decreased nerve density and innervation of vessels (66), measured with
laser Doppler (67). Nerve ischemia stimulates VEGF production exacerbating DPN
through overactivation of PKC-β (56,68–72). As a result, ischemia and low blood flow
reduces both endothelial dependent and nitric-oxide dependent vasorelaxation nangle
Other Therapeutic Agents for the Treatment of Diabetic Neuropathy 327
(46,73,74). Vascular defects also result in changes in endoneurial vessels. Epineurial
changes include arteriolar attenuation, venous distension, arteriovenous shunting that
leads to new vessel formation (75). Neural regulation of blood flow is complicated by
arteriovenous anastomoses and shunting, which deviate the blood flow from the skin
creating an ischemic microenvironment (76). There is thickening and deposition of sub-
stances in the vessel wall associated with endothelial cell growth, pericyte loss (in eyes),
and occlusion (77) Changes in blood flow correlate with changes in oxygen saturation
(60,78) and reduced sural nerve endoneurial oxygen tension (79). These changes are
followed by increased expression or action of vasoconstrictors, such as endothelin and
angiotensin and decreased activity of vasodilators, such as prostacyclin, substance P,
CGRP, endothelial derived hyperpolarizing factor, and bradykinin (80,81). Based on
this theory, investigators have given oxygen and vasodilatory agents to patients, how-
ever, these therapies have not improved DPN (82,83). Additionally, methods of assessing
skin blood flow have demonstrated that diabetes disturbs microvasculature, tissue PO
2
,
and vascular permeability. In particular, in patients with DPN there is disruption in
vasomotion, the rhythmic contraction exhibited by arterioles, and small arteries (84,85).
In Type 2 diabetes, skin blood flow is abnormal and the loss of neurogenic vasodilative
mechanism in hairy skin may precede lower limb microangiopathic processes and C-

fiber dysfunction (85,86). Changes in endoneurial blood flow often are reflected by
changes in nerve conduction (8,87–89). In addition, impaired blood flow can predict
ulceration (2,90–93). Therefore, both vascular or endoneural alterations may cause
damage over time in the peripheral nerves of patients with diabetes (Fig. 3).
328 Pittenger et al.
Fig. 3. Mechanisms of peripheral vasodilation and agents being investigated for improving
peripheral neurovascular function in diabetes.
THERAPIES TARGETING AUTOIMMUNITY
Traditional therapies for autoimmune neuropathies have proven beneficial for certain
types of diabetic neuropathy (94). Plasmaphoresis and steroidal anti-inflammatories
should be considered if the diagnosis is proximal diabetic neuropathy (diabetic amyo-
trophy) or demyelinating neuropathy. Failure of these treatments or evidence of auto-
immunity in typical diabetic polyneuropathy might warrant anti-immune approaches.
Human Intravenous Immunoglobulin
Immune intervention with human intravenous immunoglobulin (IVIg) has become
appropriate in some patients with forms of peripheral diabetic neuropathy that are
associated with signs of antineuronal autoimmunity (94,95). Chronic inflammatory
demyelinating polyneuropathy associated with diabetes is particularly responsive to
IVIg infusion. Treatment with immunoglobulin is well tolerated and is considered safe,
especially with respect to viral transmission (96). The major toxicity of IVIg has been
an anaphylactic reaction, but the frequency of these reactions is now low and confined
mainly to patients with immunoglobulin (usually IgA) deficiency. Patients may experi-
ence severe headache because of aseptic meningitis, which resolves spontaneously.
In some instances, it may be necessary to combine treatment with prednisone and/or
azathioprine. Relapses may occur requiring repeated courses of therapy.
THERAPIES FOR NERVE REGENERATION
Neurotrophins
Neurotrophic factors are proteins that promote survival of neurons regulating gene
expression through second messenger systems. These proteins may induce morpholog-
ical changes, nerve differentiation, nerve cell proliferation, and induce neurotransmitter

expression and release. Subsequently, reduction in levels of neurotrophic factors (97)
can lead to neuronal loss, possibly through activation of apoptosis (98). Many proteins
have properties and characteristics of neurotrophic factors, including cytokine-like
growth factors, TGF-β, NT3, nerve growth factor (NGF), insulin-like growth factor
(IGF)-1, and VEGF. Although only a few neurotrophic factors have been extensively
investigated, there are number of proteins that have been identified as neurotrophic fac-
tors (31,99). Many of these proteins appear to have altered expression in nerves of
patients with diabetes (77). For example, interleukin-6, a cytokine-like growth factor
may play a role in cell proliferation (100). Although their function is not well under-
stood, IGFs-I and -II have been shown to regulate growth and differentiation of neurons
(101). IGF, NGF, and other neurotrophins have been shown to be members of a family
of proteins supporting the growth and regeneration of neurons. Often these growth fac-
tors are associated with changes in nerve structure through apoptosis or proliferation.
The laminin γ gene is upregulated in normal animals undergoing postsection sciatic
nerve regeneration (102,103). This process is impaired in diabetes. Other extracellular
matrix proteins are also altered in neuropathic nerves (104). Therefore, understanding
the role of neurotrophic factors has been the focus of much investigation.
Nerve Growth Factor
Neurons affected in diabetic neuropathy are developmentally dependent on NGF.
Therefore, a decline in NGF synthesis in patients with diabetes plays a role in the
Other Therapeutic Agents for the Treatment of Diabetic Neuropathy 329