European Journal of Neurology 2004, 11 (Suppl.

1): 12–21

Pain in diabetic neuropathy case study: whole patient management

P. Marchettini, L. Teloni, F. Formaglio and M. Lacerenza
Pain Medicine Center, Scientific Institute and Hospital San Raffaele, Milano, Italy

Keywords: Painful diabetic peripheral neuropathy (DPN) is described as a superficial burning

anticonvulsants, anti-
depressants, chronic pain,
comorbidities, diabetic
peripheral neuropathy,
differential diagnosis,
neuropathic pain, quality
of life
pain associated with other positive and/or negative sensory systems affecting the feet
and lower extremities. It is one of the most commonly encountered neuropathic pain
syndromes in clinical practice. Presentation may be complicated by multiple symp-
toms, including allodynia, hyperalgesia, other less well characterized dysesthesias, and
serious disruption of social functioning and mood. Peripheral nerve function may
deteriorate, which can account for patient reports of diminution of pain after several
years of follow-up. Although current understanding holds that the pathogenesis of
DPN is multifactorial in nature, long-term studies have shown that rigorous glycemic
control is the most relevant factor in clinical intervention and can delay the onset and
slow the progression of neuropathy. In addition to glycemic control, other treatment
approaches must be examined in order to restore quality of life for patients experi-
encing painful DPN. Differential diagnosis is required to isolate DPN from other
unexplained chronic pain. Neurologic testing in painful DPN is an area of active
research and is used to assess the neurologic pathways giving rise to the pain, the
degree of neural damage and the degree of subclinical damage. Current treatment
options for DPN include mainly antidepressants and anticonvulsants, with other
agents such as tramadol, dextromethorphan and memantine being employed or
studied. This review article includes a case study of a patient with painful DPN to
demonstrate the current management strategies for this neuropathic pain syndrome.

Case presentation Painful diabetic peripheral neuropathy:

AC, a 29-year-old male with insulin-dependent diabetes
of 12 years’ duration, complained of sudden onset of
pain in both legs and feet. At the clinic, he described the
pain as an excruciating burning, squeezing sensation in
his feet and legs (visual analog scale 9), accompanied by
cramps in his calf muscles. During the day, he is unable
to wear shoes comfortably, and experiences worsening
pain when walking or bearing weight. At night, he is
unable to tolerate the touch of the blankets on his skin.
He has become exhausted from the pain and lack of
sleep, complains of frequent episodes of diarrhea, and
has lost more than 10% of his body weight over the
preceding 3 months.
Physical and neurologic examination revealed the
presence of warm, red feet, and mild general atrophy of
the legs. Muscle strength and tendon reflexes at the knee
and ankle were spared. AC was evaluated further by
clinical and electrophysiologic testing to determine the
character and extent of his neurologic deficits, and the
best course of treatment.
Correspondence: Paolo Marchettini, Director, Pain Medicine Center,
Scientific Institute and Hospital San Raffaele, Via Stamira D’Ancona
20, Milano, Italy 20127 (tel.: +39 02 26433393;
fax: +39 02 26433394; e-mail: marchettini.paolo@hsr.it).
clinical setting
Painful diabetic peripheral neuropathy (DPN) is one of
the most commonly encountered chronic neuropathic
pain syndromes in clinical practice. The first published
description of pain in the setting of diabetes mellitus is
more than 200 years old (Benbow et al., 1999). It ap-
pears that any peripheral or cranial nerve in the body
may be affected by DPN. Most commonly it is des-
cribed as a superficial burning pain associated with
other positive and/or negative sensory symptoms
affecting the feet and lower extremities. As with other
neuropathic pain syndromes, presentation is often
complicated by multiple symptoms, including allody-
nia, hyperalgesia and other less well characterized
dysesthesias, and by serious disruption of social func-
tioning and mood (Jensen and Larson, 2001).
Epidemiology
Diffuse symmetrical distal sensorimotor neuropathy is
the most common manifestation of DPN and is thought
to affect up to 34% of all patients with diabetes
(Benbow et al., 1999). Age at diagnosis, duration of
diabetes and poor glycemic control correlate with the
presence of DPN (Ziegler et al., 1992). Of some

Ó 2004 EFNS 12
 Pain in diabetic neuropathy 13

importance is the finding that newly diagnosed patients Similar benefits were observed in type 2 diabetes in the
have significant levels of neuropathy, particularly of an UK Prospective Diabetes Study (Jensen and Larson,
occult or subclinical nature (Ziegler et al., 1992). It has 2001). It is also worth noting that poor glycemic control
been more difficult to estimate the number of DPN may exacerbate painful symptoms on an acute basis,
patients who experience chronic pain, as such an esti- since hyperglycemia per se, along with rapid fluctua-
mate depends on the stage of diabetes being considered tions in blood glucose levels, have both been shown to
and the definition of chronic pain. Furthermore, there is decrease pain tolerance (Jensen and Larson, 2001).
a tendency for diabetic patients who experience chronic
pain to report a diminution in pain after several years of
Differential diagnosis of DPN
follow-up, a change that may parallel deterioration in
peripheral nerve function (Benbow et al., 1999). This Regardless of its exact etiology, DPN is most clearly
variability is responsible for the widely disparate esti- associated with long-standing diabetes (Arezzo, 1999).
mates (11%)25%) of painful DPN incidence reported It is characterized by a gradual loss or reduction of
in several sources (Table 1) (Benbow et al., 1999). sensation in the feet, and in some cases the hands, and
by pain and muscle weakness in the lower extremities.
Symptoms are often slight at first and may go unnoticed
Etiology
for many years. The first noticeable sign of DPN is
Although evidence supporting various pathogenic usually numbness, pain or tingling in the hands, feet or
mechanisms in DPN has been accumulating at a steady legs. This may be followed after several years by
pace over the last 10–15 years (Table 2), no single weakness in the muscles of the legs and feet. The loss of
theory has come to predominate, and current under- sensation in the feet is particularly serious as it may lead
standing still holds that the pathogenesis of DPN is to inattention to foot injuries and ulcers. At times
multifactorial in nature (Arezzo, 1999; Jensen and symptoms of small fiber neuropathy prevail, with cramps
Larson, 2001). Multiple etiologies notwithstanding, in the legs, spontaneous burning and hyperalgesia in the
several substantial long-term studies have shown that feet. Nerve damage caused by diabetes may also result in
glycemic control is the most relevant etiologic factor dysfunction of the visceral and reproductive organs
from the standpoint of clinical intervention. In support resulting in indigestion, diarrhea or constipation, dizzi-
of this, the Diabetes Control and Complications Trial ness, bladder infections and impotence.
showed that rigorous glycemic control in patients with The symptoms of DPN at presentation can vary
type 1 diabetes mellitus can delay the onset and slow the widely among patients, and this has considerable
progression of neuropathy (Jensen and Larson, 2001). bearing on both diagnosis and treatment. For instance,
DPN may manifest as a sudden onset of symptoms in
Table 1 Incidence of painful DPN. (Adapted with permission from
the form of a focal neuropathy (FN) of a specific nerve
Benbow et al., 1999) or nerves (National Diabetes Information, 2002). FN is
unpredictable, often painful, and occurs most often in
Study population Incidence Source
older people. FN of the cranial nerves may cause
Insulin-treated diabetics < 60 years of age 11% Boulton et al. diplopia or Bell’s palsy. The nerve trunk pain due to
Type 2 diabetics with disease > 10 years 21% Partanen et al. sudden ischemia of cranial nerves causes headache
In-hospital diabetic clinic population Chan et al. often localized behind the eye. FN of the nerves
Diabetics with pain symptoms 25%
serving the dorsal and lumbar nerve trunks can produce
Diabetics with lower limb symptoms 8%
Controls (without diabetes) 16% sensory symptoms and pain in the lower back,
thoracic areas, pelvis, stomach or flank that could be
mistaken for cardiovascular or visceral organ disease.
Affected nerves serving the lower extremities can cause
Table 2 Pathogenic mechanisms in DPN. (Adapted with permission
pain on the outside of the shin or inside of the foot that
from Arezzo, 1999)
seems more like the pain of a sudden injury. Since FN
Proposed mechanism Year Source may also improve by itself over weeks or months and is
Increased polyol pathway activity 1987 Greene et al.
not always associated with long-term nerve damage, it
Microvascular ischemia 1989 Dyck must be distinguished from the chronic neuropathic
Protein glycation/glycosylation 1990 Brownlee pain symptoms of DPN by means of a careful clinical
Altered fatty acid metabolism 1990 Jamal history and neurologic testing.
Support tissue overgrowth or 1995 Giannini and Dyck Since elevated blood glucose levels can both aggra-
degeneration
Diminished neurotrophic support 1996 Tomlinson et al.
vate the perception of pain acutely and contribute to
the progression of DPN, the first step in evaluating

Ó 2004 EFNS European Journal of Neurology 11 (Suppl. 1), 12–21

14 P. Marchettini et al.

chronic pain in a diabetic patient for which there is no Table 3 Effect of painful DPN on quality of life measures. (Reprinted
known cause related to diabetes is to assess the degree with permission from Benbow et al., 1998)

of current glycemic control (Jensen and Larson, 2001). Diabetic patients Non-diabetic controls
Nottingham
In addition to determining the period of onset and Health Profile
With pain Without pain Without pain
character of the pain, an effort should be made to Domain
determine the likely pattern of adherence to dietary and
Emotional 27.3 0 0
medication regimens in the period of time preceding the reactions (12–43.8)a,b (0–16.4) )
onset of pain. Relatively lax glycemic control over an Energy 63.2 24.0 0
extended period of time would suggest the greater (38.6–76.0)c,b (0–63.0)e )
likelihood of diabetic neuropathy. In this case, prompt Pain 53.5 0 0
(43.6–79.1)d,b (0–14.8)f )
pharmacologic intervention of pain would be the pri-
Physical 22.0 10.7 0
ority, together with instituting better control of blood mobility (21.4–32.6)c,b (0–22.1)f (0–9.3)
glucose. Even if there is a long history of good com- Sleep 50.4 0 0
pliance with diabetes control measures, DPN may still (12.6–77.6)c,b (0–19.4)f (0–12.6)
be the underlying cause, but other causes should be Social isolation 0 12.6 0
(0–22.0) (0–16.1) )
carefully ruled out before beginning treatment for
neuropathic pain. Reported as median score (95% confidence interval); higher scores
correspond to greater disability. Scores range from 0 to 100.
Comparisons versus diabetics without pain: aP < 0.001; cP < 0.01;
Comorbid conditions and functional disturbances d
P < 0.0001. Comparisons versus non-diabetics without pain:
b
associated with DPN P < 0.0001; eP < 0.01; fP < 0.05.

While maintenance of glycemic control to prevent dia-

betic sequelae must remain the principal clinical focus
in treatment, another important objective must be to
control chronic pain and its potential to disrupt the
mood and functioning of patients who have a signifi-
cant component of chronic painful DPN. Many
reviewers cite general concerns or incidental reports in
support of the potential negative impact of DPN on
quality of life, but there appear to be few studies spe-
cifically addressing this issue in a comparative way
(Benbow et al., 1998; 1999; Backonja, 1999; Hay-
thornthwaite and Benrud-Larson, 2000; Jensen and
Larson, 2001). Benbow et al. (1998) characterized the
quality of life in 41 type 1 and type 2 diabetic patients
with pain compared with that of diabetics without a
history of chronic pain and with a randomly selected,
age-matched control group, also without chronic pain.
This study clearly demonstrated the negative impact of
chronic pain in DPN (Table 3) and confirmed a tight
association between chronic pain and poor sleep, most
probably due to the worsening perception of pain
symptoms at night (Benbow et al., 1998). In a larger
sample, Galer et al. (2000) confirmed these comorbidi-
ties in a survey of 105 patients who were experiencing
painful diabetic neuropathy and who had inquired
about participating in a clinical trial for the treatment
of their pain; there was no control group in the survey
for comparison. These patients were about 50 years old
at the time of diagnosis of diabetes and encountered
their first neuropathologic symptoms at an average age
of 54 years and their first painful symptoms at a mean
age of about 57 years (Galer et al., 2000). More than
half of these patients reported substantial interference
from pain with one or more aspects of their quality of
life: mobility, employment, sleep, enjoyment of life and
recreational and social activities (Galer et al., 2000).
Painful DPN: treatment approaches
Several recent controlled clinical studies with various
antidepressant and anticonvulsant drugs suggest that
the prognosis for painful DPN need not be nearly so
bleak as it once was. However, as treatment options for
painful DPN become increasingly effective, there will
doubtless be a temptation to assume that any significant
unexplainable chronic pain in the context of long-
standing diabetes is DPN, and that treatment for
neuropathic pain may start immediately. This variable
presentation underscores the need for differential
diagnosis by obtaining a careful history and staging the
diabetic patient with neurologic testing.
The need for detailed neurologic examination is
supported by a number of logical clinical objectives: to
characterize the neurologic pathways that are likely to
have given rise to the chronic pain, to assess the degree
of neural damage in the affected systems, and to assess
the degree of subclinical damage that may be present in
other neurologic systems. Neurologic testing in painful
DPN is an area of active research, and results are
beginning to coalesce into protocols for standardized
testing. While any necessary assessments are being
carried out, it may be worthwhile asking the patient to
make entries into a standardized pain diary for a week
or so to provide a baseline with which to compare
progress.

Ó 2004 EFNS European Journal of Neurology 11 (Suppl. 1), 12–21


Treatment options
With the publication of several well-controlled clinical
trials over the past 5 years, the treatment options for
painful DPN are proving to be more effective than
previously thought. Of particular note are the results of
three publications: a systematic review of painful DPN
treatment with antidepressants (McQuay et al., 1996), a
systematic review of painful DPN treatment with anti-
convulsants (McQuay et al., 1995) and a large, stand-
alone trial of the anticonvulsant gabapentin (Backonja
et al., 1998).
Antidepressants
Over the last 30 years, there have been numerous con-
trolled clinical trials of various antidepressants for the
relief of neuropathic pain, but many of these trials were
relatively small and used a wide range of antidepressant
drugs and doses (McQuay et al., 1996). McQuay et al.
(1996) performed a meta-analysis of this therapeutic
category based on published studies, reviews and meta-
analyses appearing between the years 1950 and 1994.
To limit the scope and improve the clinical value of
their analysis, they applied fairly stringent selection
criteria. Studies were to have at least 10 patients per
treatment arm and were strictly graded for elements of
experimental design that would lend themselves to the
meta-analytical technique, including adequacy of
randomization techniques, adequacy of blinding tech-
niques, uniformity of control groups (placebo- and
active-controlled studies were analyzed separately),
uniformity of patient groupings, observation periods,
efficacy endpoints and withdrawal rates based on cause.
Comparative statements were based on the proportion
of patients experiencing a more than 50% reduction
in pain.
For the 13 reports detailing treatment of DPN, the
combined odds ratio for benefit from an antidepressant
relative to placebo was highly significant. Within these
13 reports, nine different agents were tested and six
reports independently showed a significant benefit over
placebo. Results for imipramine alone, desipramine
alone and all tricyclic antidepressants combined varied
only slightly from one another, implying interchange-
ability of the tricyclics with regard to efficacy. Though
lesser numbers of patients were treated with the two
serotonin re-uptake inhibitors, paroxetine and fluoxe-
tine, these agents appeared to be distinctly less effective
in relieving painful DPN.
Analysis of adverse events was based on the com-
bined results from all regimens and all neuropathic pain
populations (DPN, postherpetic neuralgia and several
other syndromes) and showed that minor adverse
events occurred with approximately the same incidence
Ó 2004 EFNS European Journal of Neurology 11 (Suppl. 1), 12–21
Pain in diabetic neuropathy 15
as overall benefit, while major adverse events (i.e. those
leading to withdrawal) were less common.
Antidepressants were judged to have a clear analgesic
effect in each of several different pain syndromes, and
the efficacy in each syndrome separately was similar to
the efficacy overall, despite the presumption of differing
etiologic mechanisms (McQuay et al., 1996). There
were several additional clinical insights stemming from
this broad view of antidepressant use in neuropathic
pain. In clinical practice, titration of tricyclic anti-
depressants to optimal dose with minimal side-effects
can be accomplished in 5 days or less; analgesic effect
occurs without significant change in mood measure-
ments; and benefit seems to occur to the same extent
regardless of the character of the pain at presentation
(e.g. burning versus shooting).
Anticonvulsants
McQuay et al. (1995) also reviewed the anticonvulsant
category for efficacy in neuropathic pain syndromes,
again covering the period 1950–94 (McQuay et al., 1995).
The authors also used the same selection criteria for
including studies in the analysis, as well as similar efficacy
and safety endpoints (equivalent to at least a 50% pain
reduction; all patients included in safety). Overall, 20
published studies were included, employing four anti-
convulsant drugs (carbamazepine, phenytoin, clonaze-
pam and sodium valproate). Only three studies focused
on painful DPN; improvement relative to placebo was
significant in only two and ranged from 30% to 50% of
the actively treated patients. For the three studies com-
bined, efficacy and minor adverse events occurred with
similar probabilities; adverse events leading to with-
drawal were much less common. Drowsiness, dizziness
and disturbance in gait were the most common adverse
events.
The efficacy and safety profiles of the various anti-
convulsant agents addressed by this meta-analysis
showed them to be approximately equivalent in the
treatment of three pain syndromes ) trigeminal neur-
algia, painful diabetic neuropathy and migraine pro-
phylaxis ) and approximately equivalent to the
antidepressants as a class.
There are no published reports on the use of gabap-
entin in neuropathic pain before 1996 and consequently
the review of McQuay et al. (1995), described in detail
above, contains no mention of this relatively recently
introduced anticonvulsant analgesic agent (Mellegers
et al., 2001). A randomized, placebo-controlled clinical
trial of gabapentin in painful DPN, notable for its size
compared to other trials, was published by Backonja
et al. in 1998. These investigators treated 165 DPN
patients with pain that had been present 1–5 years prior
to enrollment; diagnosis was made clinically and was
16 P. Marchettini et al.

10 Gabapentin
Placebo
Placebo

Mean endpoint scores

Gabapentin 100
8 90 P = 0.03

Short Form-36
Mean pain score

6
4 a
2
0 Bodily pain
Screening 1
Figure 1 Effect of gabapentin on mean pain score. aP < 0.01;
b
P < 0.05. (Reprinted with permission from Backonja et al.,
1998).
not dependent on electrophysiologic examinations.
Dosing ranged from 900 to 3600 mg/day in three equal
divided doses. All patients were titrated to the maxi-
mum dose unless intolerable side-effects occurred, in
which case dosing was continued at the previously tol-
erated maximum dose for the remainder of the 8-week
study. In addition to evaluation of pain intensity, this
study included extensive measurements of sleep inter-
ference, mood and quality of life. These measures, while
intuitively relevant to the evaluation of painful condi-
tions, have not been routinely included in most trials of
neuropathic pain until recently.
Compared to placebo, the gabapentin-treated
patients experienced significant pain reduction by week
2 of treatment (Fig. 1), well before maximum tolerated
doses had been achieved in most patients. Significant
improvement in sleep interference occurred at week 1
(Fig. 2). Approximately 60% of patients receiving
gabapentin had at least a moderate improvement in
self-assessment compared with only 33% of placebo-
treated patients (P ¼ 0.001), and this was reflected in
5.5
Mean sleep interference score
80
70 P = 0.01 P = 0.01
60
50
40
a
b a b 30
b b
20
10
0
Mental health Vitality
2 3 4 5 6 7 8
Week
Figure 3 Effect of gabapentin on quality of life: Short Form-36.
(Reprinted with permission from Backonja et al., 1998).
significantly improved measures of quality of life
(Fig. 3). Dizziness and somnolence occurred in
approximately 20% of patients receiving gabapentin
compared with about 5% in those receiving placebo.
About the same number of patients withdrew from
active treatment (8%; n ¼ 7) compared to placebo (6%;
n ¼ 5). Since efficacy was noted prior to the completion
of upward titration, it is possible that many adverse
events could be avoided by titrating to pain relief rather
than maximum tolerated dose.
Another relatively recently introduced anticonvulsant,
lamotrigine, was also shown to be effective in DPN in a
randomized, placebo-controlled trial of 59 patients
(Eisenberg et al., 2001). In this study, the patients treated
with 200, 300 or 400 mg of lamotrigine showed significant
improvements on a 0–10 numerical pain scale compared
to placebo (P < 0.001). However, there were no
improvements in the McGill Pain Questionnaire or Beck
Depression Inventory in those treated with active drug.
Other therapies
Many of the established treatments for neuropathic
pain described above fail to provide complete pain
relief, thus creating a role for other therapeutic agents
Placebo that could provide pain relief and eventually be added

5.0 Gabapentin to an anticonvulsant. Capsaicin and tramadol are two

4.5 such agents for which placebo-controlled clinical trials
4.0 provide supporting evidence.
a Capsaicin is a topical analgesic that binds to receptors
3.5
(VR1/TRPV1) on subpopulations of sensory nociceptive
3.0
a C or Ad fibers. Initially capsaicin causes pain by initiating
2.5 a nociceptor firing resulting in increased sensitivity to
a
2.0 b painful thermal and mechanical stimuli. An analgesic
a b
b
1.5 effect follows the painful experience as prolonged expo-
0 1 2 3 4 5 6 7 8
sure to capsaicin desensitizes nociceptive terminals
Week
through partial or complete degeneration of axon
Figure 2 Effect of gabapentin on sleep interference score. terminals or the neuron (Caterina and Julius, 2001).
a
P < 0.01; bP < 0.05. (Reprinted with permission from Backonja A large-scale placebo-controlled trial of capsaicin in
et al., 1998). 252 patients with DPN or radiculopathy who completed

Ó 2004 EFNS European Journal of Neurology 11 (Suppl. 1), 12–21


at least 2 weeks of treatment showed significantly more
patients with improved pain after capsaicin than with
placebo beginning at 2 weeks of treatment and persisting
through the end of the 8-week trial (Capsaicin Study
Group, 1991). Burning at the site of application was
experienced more frequently by patients applying
capsaicin than those applying vehicle cream, and more
patients discontinued treatment with capsaicin than
placebo because of increased burning sensation. The
sensation of burning diminished over time with repeated
applications and was thought to be generally well toler-
ated. However, in these authors’ experience, the burning
is often poorly tolerated.
Tramadol, a centrally acting analgesic unrelated
chemically to the opiates, was evaluated in the treat-
ment of DPN in two placebo-controlled clinical trials.
The larger trial of Harati et al. (1998), involving 131
patients, found a significantly reduced mean pain score
as well as improved physical and social functioning in
painful DPN patients receiving tramadol compared
with those treated with placebo. A smaller study in 28
patients conducted by Sindrup et al. found a significant
decrease in ongoing pain as well as touch-evoked pain
(Sindrup et al., 1999). The dose of tramadol in the
Harati et al. study was reported as a mean of 210 mg/
day, whereas Sindrup et al. employed a dose ranging
from 200 to 400 mg/day; pain relief was apparently not
correlated with blood levels.
The similarity of tramadol to the opioid class of
analgesics is still a source of concern in its use, for some
clinicians. Tramadol binds to l-opioid receptors and
inhibits re-uptake of norepinephrine and serotonin;
however, its precise mechanism is not known since the
analgesic effect is only partially antagonized by the
opioid antagonist naloxone. Tramadol causes signifi-
cantly less respiratory depression than morphine and, in
contrast to morphine, does not stimulate the release of
histamine. Tramadol does produce a l-opioid type
dependence like codeine or dextropropoxyphene; how-
ever, there appears to be little potential for tolerance
development or abuse. Thus tramadol may be consid-
ered another useful agent in painful DPN, particularly
effective to treat the nociceptive component of nerve
trunk pain such as that caused by ischemia. However,
its use may be limited in elderly patients or those with
impaired hepatic or renal function, and it should be
used with caution in patients taking other psychoactive
drugs (Physicians’ Desk Reference, 2003).
The use of dextromethorphan and memantine,
low-affinity N-methyl-D-aspartate antagonists, was
investigated in the treatment of neuropathic pain in
patients with DPN and postherpetic neuralgia (Sang
et al., 2002). In patients with DPN, the mean reduction
in pain intensity from baseline was 33% with dextro-
Ó 2004 EFNS European Journal of Neurology 11 (Suppl. 1), 12–21
Pain in diabetic neuropathy 17
methorphan, 17% with memantine and 16% with
lorazepam (active placebo). The corresponding reduc-
tions in pain intensity in patients with postherpetic
neuralgia were 6%, 2% and 0%. Although these results
were not statistically significant, 68% of patients with
DPN had moderate or better pain relief with dextro-
methorphan than did 47% with memantine. Adverse
events with dextromethorphan and memantine included
sedation, dry mouth and gastrointestinal distress (Sang
et al., 2002).
Optimizing treatment
Treatment should begin with an assessment of current
glycemic control, and if necessary prompt action should
be taken before implementing treatment for chronic
pain. Treatment for pain may start with a tricyclic
antidepressant or, if the side-effects of these drugs are a
consideration, as in the elderly, gabapentin may be tried
as a first-line therapy. The standard neurologic evalu-
ation should include an assessment of magnitude of
pain by means of pain diaries, visual analog scale
determinations or numeric rating scales. These will be
necessary to determine progress in alleviating specific
pain symptoms and in maintaining overall control of
the diabetes. Farrar et al. (2001) found that a reduction
of 2 points in a patient’s self-rated pain score (in both
placebo or active drug treatments) correlated with a
clinically important improvement (i.e. much improved)
in pain symptoms on the 0–11 point categorical Likert
scale.
Case resolution
Sensory examination of AC showed signs and symp-
toms representative of small fiber involvement. In
addition to the predominant complaint of burning,
squeezing leg pains with cramps, consistent with
unmyelinated fiber involvement, the patient reported a
Ôpins and needlesÕ sensation, consistent with degener-
ation of small myelinated fibers. Small fiber involve-
ment was further implicated by the presence of pain on
warming or stroking the affected skin.
Electromyography studies showed no spontaneous
activity and normal motor unit recruitment. However, a
minor increase in the prevalence of polyphasic units
indicated terminal collateral sprouting and possibly
muscle reinnervation. Nerve conduction studies of large
myelinated fibers revealed normal motor and sensory
function, normal amplitudes and normal conduction
velocities (Fig. 4). Quantitative sensory testing of the
foot confirmed an increased threshold to warm sensa-
tion and hyperalgesia to heat. A decreased threshold for
cold sensation was also noted (Fig. 5). The thermograph
18 P. Marchettini et al.

administrations) and tramadol at 50 mg twice a day

41 µV
41 µV was added. Some additional benefit and further improve-
Index
Index to
to wrist
wrist 54
54 m/s
m/s
ment was attained by increasing the tramadol dose to
100 mg twice per day. Attempts to decrease the dose of
2.8
2.8 ms
ms gabapentin after tramadol was added caused worsening
20 µV
20 µV
of the pain. Thus the two drugs were continued at their
22 ms
ms maximum effective doses (i.e. 4200 mg/day gabapentin
19 µV
19 µV
and 100 mg/day tramadol). This regimen of gabapentin
was decided on because it was effective and the patient
7.1
7.1 ms
ms had no side-effects and a good although partial re-
sponse to lower doses.
wrist
wrist to
to elbow
elbow 60
60 m/s
m/s
At the time the patient was returned to the care of his
diabetologist, insulin therapy consisted of Humulin 30/
70, 16 units at noon and 12 units at night. Blood glu-
Figure 4 Case study: sensory nerve conduction results.
cose levels (mg/dl) were noted to be in the following
ranges: 75–80 (06.00 hours), 115–180 (11.00 hours),
41.5 (mean) 140–220 (15.00 hours), 130–180 (18.00 hours) and 130–
50
145 (21.00 hours).
43.0 42.1
41.6
39.3
Overall, the findings on examination and testing,
Temperature (˚C)

together with the symptoms reported by the patient, are

32
consistent with primary dysfunction of the unmyeli-
nated nerve fibers in conjunction with a loss of sym-
pathetic vasoconstrictor drive. Large fiber function was
23.1 22.9 preserved, and is consistent with the relatively recent
20.7
19.5
onset of symptoms.

21.5 (mean)
0
Figure 5 Case study: quantitative sensory testing results. Nor-
mative values for cold are 34 °C ± 2 °C; normative values for
warm are 30 °C ± 2 °C.
showed an inversion of the normal temperature gradi-
ent, with the patient’s toes being warmer than his
calves.
The patient was initially prescribed carbamazepine
slow release (Tegretol 400 CR twice per day) along with
amitriptyline (30 mg twice per day) because of their low
cost, even though gabapentin is the only drug with
broad approval for neuropathic pain in Italy. This
combination of therapies was discontinued when it
resulted in unpleasant sedation and dizziness and failed
to provide pain relief. Gabapentin was started at
300 mg three times per day and increased by 300 mg
every 3 days. Stimulus-induced pain (hyperalgesia) was
controlled at a relatively low dose of about 600 mg
three times per day, and sleep patterns began to
improve. Additional increases in the dose of gabapentin
to 3600 mg/day (maximal approved dose of gabapentin
in Italy) produced further improvement in spontaneous
pain as well as sleep patterns. Residual pain, which had
a squeezing/cramping quality, was not controlled by a
further increase to 4200 mg/day (divided into four
Summary and conclusions
Long-term studies have shown that poor glycemic
control is the most relevant etiologic factor in the
development of DPN. It has also been shown that poor
glycemic control may exacerbate painful symptoms on
an acute basis (Jensen and Larson, 2001). For these
reasons, all diabetic patients being considered for
chronic pain treatment should be evaluated for
improvement of glycemic control. The symptoms of
DPN at presentation can vary widely among patients
and can resemble symptoms of more serious conditions.
It is therefore important to eliminate other possible
causes of chronic pain before effective analgesic therapy
is begun. The possibility that an FN is the cause of a
painful symptom should also be considered, since these
conditions are not necessarily associated with long-term
nerve damage and tend to improve spontaneously. Pain
in FN for at least part of the history of symptoms is
thought to be a nerve trunk pain of nociceptive origin,
due to excitation of nervi nervorum. Our experience has
shown this pain responds well to tramadol. While
maintenance of glycemic control and prevention of
diabetic sequelae must remain the principal clinical
focus in treatment, an equally facilitating objective
must be to control chronic pain and its potential to
disrupt the mood and functioning of patients who have
a significant component of chronic painful DPN.

Ó 2004 EFNS European Journal of Neurology 11 (Suppl. 1), 12–21


Methods of neurologic evaluation in DPN
clinical examination
Traditionally, assessing diabetic patients for nerve
damage has involved a Ôhands-onÕ clinical examination
in the search for sensory-motor deficits (Arezzo,
1999). While highly efficient in the hands of an
experienced clinician, the results are necessarily ex-
pressed in subjective, relativistic terms such as
ÔdiminishedÕ tendon reflexes, ÔloweredÕ pain threshold,
pinpricks, heat or cold, etc., and are therefore asso-
ciated with a considerable degree of variability when
carried out at different times or by different examiners
(Arezzo, 1999). Another drawback to the standard
clinical evaluation is that observable deficits are lim-
ited to the effects of established neuropathology. More
sensitive objective measurements are needed to track
the early stages of onset and effects of various inter-
ventions.
Traditional electrophysiology
Electrophysiologic methods for evaluating DPN offer
greater objectivity than a clinical examination, espe-
cially since neurologic decline in long-standing dia-
betes is slow and insidious in its onset. A broad
battery of tests would generally include measurements
of the conduction velocity of both sensory and motor
nerves and the amplitude of the signals generated
(Arezzo and Zotova, 2002). Each component of a
multitest battery has a particular value in assessing
neural function as well as limits to its interpretation
(Arezzo, 1999; Arezzo and Zotova, 2002). Testing is
commonly limited to maximal nerve conduction
velocity (maxNCV) (Fig. 6), sensory action potential
and amplitude of the motor action potential as well
as motor nerve conduction velocity (Arezzo, 1999;
Arezzo and Zotova, 2002).
A narrow focus on electrophysiology alone creates
a limited estimate of disease status for a number of
reasons. Whole nerve electrophysiology measures
function exclusively in large-diameter neurons, but the
spectrum of nerve fibers involved in DPN includes
both large-diameter myelinated fibers (Ab) that con-
vey vibratory and tactile information and small-
diameter, thin myelinated (Ad) or unmyelinated (C)
fibers that convey information about pain and tem-
perature. Measurement of maxNCV is generally not
considered adequate for tracking an individual patient
as disease progresses, since nerve fiber types may not
be uniformly affected in DPN and some patients may
experience earlier damage in a fiber class not well
detected by maxNCV. Even if the large-diameter
Ó 2004 EFNS European Journal of Neurology 11 (Suppl. 1), 12–21
Pain in diabetic neuropathy 19
Diabetic
54 Normal
Maximal nerve conduction
52
velocity (m/sec)
50
48
46
44
42
1 2 3 4 5 6 7 8 9 10
Years
Figure 6 Decrease in maximal nerve conduction velocity as a
function of years after diagnosis of diabetes compared to non-
diabetics. (Reprinted with permission from Arezzo, 1999).
fibers are representative of overall disease progression
as some clinicians think, the rate of decline is too slow
to be a sensitive measure of disease progression in
individual patients, and it is relatively insensitive to
changes in axonal function that may occur earlier and
in the absence of demyelination (Arezzo, 1999). The
amplitude of whole nerve action potentials also
changes dramatically with disease progression (Arezzo
and Zotova, 2002). This change may be related to a
number of factors that correlate with disease pro-
gression, including, and in addition to maxNCV, the
amount of myelin remaining, mean cross-sectional
diameter, internodal distances in the nerve segment
being tested, and the distribution of ion channels in
the nodal regions (Arezzo and Zotova, 2002). As with
maxNCV, the rate of change in whole nerve amplitude
is uneven, which, together with the multitude of fac-
tors determining its size and form, limits its value in
tracking disease progression in individual patients
(Arezzo and Zotova, 2002).
Quantitative sensory testing
Quantitative sensory testing represents a refinement in
the use of standard electrophysiologic equipment that
permits an earlier detection of sensory dysfunction
(Arezzo, 1999). The procedure measures sensory
thresholds in a variety of anatomical districts by
delivering appropriate predefined stimuli (heat, cold,
vibratory) at designated intensities (Arezzo, 1999).
The changes thus detectable with these instruments
have revealed subclinical neuropathy in children and
adolescents with type 1 diabetes, and may eventually
be used to predict the onset of foot ulcers in adult
20 P. Marchettini et al.
patients with DPN (Arezzo, 1999). Quantitative sen-
sory testing must be considered only a semiobjective
measure of nerve function, as it is dependent to some
degree on the attention and motivation of the patient
(Arezzo, 1999). Abnormalities detected by quantita-
tive sensory testing are also not specific to peripheral
nerve dysfunction, and other causes may need to be
ruled out (Arezzo, 1999).
Developing technologies
There are several methods for assessing neuropathol-
ogy that are still in the early stages of development.
References
Arezzo JC (1999). New developments in the diagnosis of
diabetic neuropathy. Am J Med 107:9S–16S.
Arezzo JC, Zotova E (2002). Electrophysiologic measures of
diabetic neuropathy: mechanism and meaning. Int Rev
Neurobiol 50:229–255.
Backonja M (1999). Managing painful diabetic neuropathy.
Hospital Practice 79–88.
Backonja M, Beydoun A, Edwards KR et al. (1998). Gabap-
entin for the symptomatic treatment of painful neuropathy
in patients with diabetes mellitus. A randomized controlled
trial. J Am Med Assoc 280:1831–1836.
Benbow SJ, Wallymahmed ME, MacFarlane A (1998).
Diabetic peripheral neuropathy and quality of life. Q J
Med 91:733–737.
Benbow SJ, Cossins L, MacFarlane A (1999). Painful diabetic
neuropathy. Diabetic Med 16:632–644.
Boulton AJ, Knight G, Drury J, Ward JD (1985). The
prevalence of symptomatic, diabetic neuropathy in an
insulin-treated population. Diabetes Care 8:125–128.
Brownlee M. (1990). Advanced products of nonenzymatic
glycosylation and the pathogenesis of diabetic complica-
tions. In: Rifkin H, Porte D, eds. Diabetes Mellitus: Theory
and Practice, 4th edn. Elsevier Science Publishing Co., Inc,
New York, NY, pp. 279–291.
Capsaicin Study Group (1991). Treatment of painful diabetic
neuropathy with topical capsaicin. A multicenter, double-
blind, vehicle-controlled study. Arch Intern Med 151:2225–
2229.
Caterina MJ, Julius D (2001). The vanilloid receptor: a
molecular gateway to the pain pathway. Annu Rev Neurosci
24:487–517.
Chan AW, MacFarlane IA, Bowsher DR, Wells JC, Bessex C,
Griffiths K (1990). Chronic pain in patients with diabetes
mellitus; comparison with a non-diabetic population. Pain
Clinic 3:147–159.
Dyck PJ (1989). Hypoxic neuropathy: does hypoxia play a
role in diabetic neuropathy? The 1988 Robert Wartenberg
lecture. Neurology 39:111–118.
Eisenberg E, Lurie Y, Braker C, Daoud D, Ishay A (2001).
Lamotrigine reduces painful diabetic neuropathy: a rando-
mized, controlled study. Neurology 57:505–509.
Farrar JT, Young JP Jr, LaMoreaux L, Werth JL, Poole RM
(2001). Clinical importance of changes in chronic pain
These include peripheral nerve visualization with
magnetic resonance imaging, composite parameters
based on clinical and electrophysiologic measurements
(i.e. information from quantitative sensory testing,
electrophysiology, and signs and symptoms combined
into a single score), and minimally invasive skin punch
biopsies in conjunction with specific immunohisto-
chemical staining (Arezzo, 1999). Each of these
methods will have its limitations even when fully
documented, but they will all have some merit in being
able to evaluate disease progression and therapeutic
benefit in individual patients.
intensity measured on an 11-point numerical pain rating
scale. Pain 94:149–158.
Galer BS, Gianas A, Jensen MP (2000). Painful diabetic
polyneuropathy: epidemiology, pain description, and qual-
ity of life. Diabetes Res Clin Prac 47:123–128.
Giannini C, Dyck PJ (1995). Basement membrane reduplica-
tion and pericyte degeneration precede development of
diabetic polyneuropathy and are associated with its severity.
Ann Neurol 37:498–504.
Greene DA, Lattimer SA, Sima AA (1987). Sorbitol,
phosphoinositides, and sodium-potassium-ATPase in the
pathogenesis of diabetic complications. N Engl J Med
316:599–606.
Harati Y, Gooch C, Swenson M et al. (1998). Double-blind
randomized trial of tramadol for the treatment of the pain
of diabetic neuropathy. Neurology 50:1842–1846.
Haythornthwaite JA, Benrud-Larson LM (2000). Psy-
chological aspects of neuropathic pain. Clin J Pain 16:
S101–S105.
Jamal GA (1990). Pathogenesis of diabetic neuropathy: the
role of the n-6 essential fatty acids and their eicosanoid
derivatives. Diabetic Med 7:574–579.
Jensen PG, Larson JR (2001). Management of painful diabetic
neuropathy. Drugs Aging 18:737–749.
McQuay H, Carroll D, Jadad AR, Wiffen P, Moore A (1995).
Anticonvulsant drugs for management of pain: a systematic
review. Br Med J 311:1047–1052.
McQuay HJ, Tramer M, Nye BA, Carroll D, Wiffen PJ,
Moore RA (1996). A systematic review of antidepressants in
neuropathic pain. Pain 68:217–227.
Mellegers MA, Furlan AD, Mailis A (2001). Gabapentin for
neuropathic pain: systematic review of controlled and
uncontrolled literature. Clin J Pain 17:284–295.
National Diabetes Information Clearinghouse (2002). Dia-
betic Neuropathies: The Nerve Damage of Diabetes. Avail-
able at: http://www.niddk.nih.gov/health/diabetes/pubs/
neuro/neuro.htm. Accessed 10 January 2003.
Partanen J, Niskanen L, Lehtinen J, Mervaala E, Siitonen O,
Uusitupa M (1995). Natural history of peripheral neuro-
pathy in patients with non-insulin-dependent diabetes
mellitus. N Engl J Med 333:89–94.
Physicians’ Desk Reference (2003). Ultracet prescribing
information. Thompson PDR, Montvale, NJ, pp. 2508–
2510.
Ó 2004 EFNS European Journal of Neurology 11 (Suppl. 1), 12–21

Sang CN, Booher S, Gilron I, Parada S, Max MB (2002).
Dextromethorphan and memantine in painful diabetic
neuropathy and postherpetic neuralgia: efficacy and dose–
response trials. Anesthesiology 96:1053–1061.
Sindrup SH, Andersen G, Madsen C, Smith T, Brosen K,
Jensen TS (1999). Tramadol relieves pain and allodynia in
polyneuropathy: a randomised, double-blind, controlled
trial. Pain 83:85–90.
Ó 2004 EFNS European Journal of Neurology 11 (Suppl. 1), 12–21
Pain in diabetic neuropathy 21
Tomlinson DR, Fernyhough P, Diemel LT (1996). Neuro-
trophins and peripheral neuropathy. Philos Trans R Soc
Lond B Biol Sci 351:455–462.
Ziegler D, Gries FA, Spuler M, Lessmann F (1992). The
epidemiology of diabetic neuropathy. Diabetic Cardiovas-
cular Autonomic Neuropathy Multicenter Study Group.
J Diabetes Complications 6:49–57.