Diabetologia (2005) 48: 817-823DO! 1O.1007/s00125-005-1734-2

ARTICLE

E. Bosi' M. Conti. C. Vermigli' G. Cazzetta . E. Peretti .M. C. Cordoni' G. Galimberti . L. Scionti

Effectivenessof frequency-modulatedelectromagneticneuralstimulationin the treatmentof painfuldiabeticneuropathy

Received: 7 December 2004/ Accepted: 23 February 2005/ Published online: 15 Aprii 2005cg Springer-Verlag 2005

Abstract Aims/hypothesis: The largely unsatisfactoryresults reported far the pharmacological treatment of dia-betic neuropathy has spurred the search for alternativetherapies. The aim ofthis study was to evaluate the efficacyof iTequency-modulated electromagnetic neural stimula-tiGli (FREMS) as a novel treatment for painful diabeticneuropathy. Methods: Patients (n=3l) with painful neu-ropathy associated with decreased nerve conduction veloc-ity «40 m/s) and increased vibration perception threshold(>25 V) were enrolled in a randomised, double-blind,crossover study designed to compare the effects ofFREMSwith those of placebo. Each patient received two series often treatments of either FREMS or placebo in random se-quence, with each series lasting no more than 3 weeks. Theprimary efficacy end point was the change in pain mea-sured by a visual analogue scale (VAS). Results: FREMSinduced a significant reduction in daytime and night-timeVAS pain score (alIp<0.02). Furthermore, FREMS induceda significant increase in sensory tactile perception, as as-sessed by monofilament; a decrease in foot vibration per-ception threshold, as measured by a biothesiometer; andan increase in molar nerve conduction velocity (alIp<O.Ol).No significant changes were observed after placebo. Com-parison of measurements at the 4-month folIow-up withthose at baseline revealed that a significant benefit persistedfor alI measures that showed an improvement at the endof treatment, with an additional improvement in qualityof life evaluated by the Short Form-36 questionnaire (alI

E. Bosi (~) . M. Conti. E. Peretti . G. GalimbertiDiabetes and Endocrinology Unit,Department of Generai Medicine,Vita-Salute San Raffaele University Hospital,Via Olgettina, 60,20132 Milan, Italye-mail: bosi.emanuele@hsr.itTe!.: +39-02-26432818Fax: +39-02-26432827

C. Vermigli. G. Cazzetta . M. C. Cordoni. L. SciantiDepartment of Internai Medicine, University of Perugia,Perugia, Italy

p< 0.05). No significant side effects were recorded duringthe study. Conclusions/interpretation: FREMS is a safeand effective therapy far neuropathic pain in patients withdiabetes and is able to modify some parameters of periph-eral nerve function.

Keywords Clinical trial . Diabetes . Electromagneticstimulation . Painful neuropathy

Abbreviations FREMS: frequency-modulatedelectromagnetic neural stimulation . MNCV: molar nerveconduction velocity . SF36: Short Form-36 questionnaire .SNCV: sensory nerve conduction velocity. TENS:transcutaneous electrical nerve stimulation . VAS: visualanalogue scale. VEGF: vascular endothelial growth factor

Introduction

Peripheral neuropathy is a frequent and disabling micro-vascular complication ofboth type l and type 2 diabetes [l].This condition may be prevented by good blood glucosecontrai [2]; however, it is at best halted, once established,even after long-term blood glucose normalisation, such asthat observed folIowing successful pancreas transplantation[3, 4]. The pathological halImarks of diabetic neuropathyafe microangiopathy of the vasa nervorum, loss of axonsand axonal atrophy, alI of which afe the result of a com-bination of different mechanisms of tissue damage that afecommon to alIlong-term complications of diabetes [5].

The pharmacological treatment of diabetic neuropathy islargely unsatisfactory, mainly due to a lack of drugs that acton the underlying pathogenetic mechanisms. Aldose reduc-tase inhibitors afe among the few compounds with thismode of action; however, the results of clinical trials per-formed to date bave been disappointing [6]. Consequent1y,current therapy is purely symptomatic, aiming to relieve thepain associated with neuropathy through the administrationof various analgesics, tricyclic antidepressants, anti-arrhyth-mics [7] and, more recent1y, the new anti-epileptic agentsgabapentin [8] and lamotrigine [9], and opioids [10].

818

Non-pharmacological symptomatic treatments bave alsobeen proposed, including acupuncture [11], near-infraredphototherapy [12], low-intensity laser therapy [13], staticand pulsed magnetic field therapies [14, 15], and variouselectrotherapies, including transcutaneous electrical nervestimulation (TENS) [16, 17], percutaneous electrical nervestimulation [18] and spillai cord electrostimulation [19].The rationale for the lise of electrical nerve stimulation indiabetic neuropathy is based on its historical, though con-troversial, lise in various painful clinical conditions [20] andon some beneficial effects reported in the treatment of otherdiabetic complications, such as foot ulcers [21].

Frequency-modulated electromagnetic neural stimula-tion (FREMS) has recently been developed as a novel elec-trotherapy. This method is different frOlli TENS and otherknown electrotherapy systems, as it uses sequences of mod-ulated electrical stimuli that vary automatically in termsof pulse frequency, duration and voltage amplitude. TheFREMS method was designed on the basis of the hypoth-esis that the summation of sub-threshold electrical stimuli,conveyed through the skin proximal to a molar nerve in anon-invasive system, would induce composite molar actionpotentials in excitable tissues. A single impulse of low in-tensity and short duration, such as that used by conventionalelectrotherapies, is unable to overcome the dielectric skinbarrier to excite the underlying nervous or muscular tissue.However, FREMS achieves this effect through specificsequences of weak impulses, characterised by a rapid in-crease and decrease in pulse frequency and duration, whichresult in the graduaI recruitrnent of membrane potentials inthe stimulated tissues [22].

These characteristics prompted us to evaluate the thera-peutic potential of FREMS in human diabetic neuropathy.In this paper we report the results of a two-çentre, random-ised, double-blind, placebo-controlled, crossover clinicaltrial on FREMS treatrnent of patients with painful diabeticneuropathy.

Subjects and methods

Study design and end points The study had a randomised,double-blind, placebo-controlled, crossover designoThe pri-mary end point was the change in grading of daytime andnight-time pain, as assessed using a visual analogue scale(VAS). Secondary end points were changes in: sensitivity tomonofilament; vibration perception threshold, as measuredby a biothesiometer; quality oflife, as assessed by question-naire; molar nerve conduction velocity (MNCV); and sen-sory nerve conduction velocity (SNCV). The treatmentconsisted often sessions ofplacebo followed by ten sessionsofFREMS (sequence 1) or vice versa (sequence 2) at ran-dom, separated by a wash-out period of 1week. Each treat-meni session was administered at intervals of at least 24 h,and each ten-session series lasted no more than 3 weeks.Randomisation to sequence 1 or sequence 2 was performedcentrally at the lime of enrolment. Principal investigators,physicians, nurses, technicians and statisticians were un-aware of treatrnent assignment.

Characteristics oJ FREMS Treatment with FREMS wasperformed using sequences of monophase-compensatednegative potenti al electrical pulses that afe characterised bya sharp spike and an asymmetrical shape (peak amplitudevariable frOlli0-255 V, pulse frequency variable within therange 1-50 Hz, pulse duration variable within the range10-40 I-ls).

Administration oJFREMS and placebo Electrotherapy andplacebo were administered using the Physioflog ETS 501(Lorenz Therapy System; Lorenz Biotech, Medolla, Italy)via four electrodes applied to the lower extremities; theoriginaI device was modified by the addition of a switch toapply treatrnent A (later revealed to be placebo) or treatrnentB (later revealed to be FREMS). Each session of eitherplacebo or FREMS lasted far 30 min. Placebo consisted ofno electric current transmission. This placebo was chosenafter a preliminary study had shown that patients with avibration perception threshold higher than 25 Veffectivelyhad no perception of the electrical stimuli administered bythe FREMS device (data noi shown). These findings were inaccordance with those reported by two other studies show-ing a direct correlation between vibration perception thresh-old measured by a biothesiometer and current perceptionthreshold measured by a Neurometer (Neurotron, Baltimore,MD, USA) across the same range of frequencies used byFREMS [23, 24]. During sessions of either placebo orFREMS, patients were invited to modulate the delivery ofneurostimulation themselves, by progressively increasingthe voltage of electrical stimulation along a scale of 0--255V through a manually gradable remote contrai device thatincreased the voltage by 1 V per step up to the maximalallowed, which corresponded to the possible perception ofbuming at the site of the electrode.

Subjects Patients who met the following criteria wereinvited to participate in the study: (1) type 1 or type 2 dia-betes according to American Diabetes Association criteria[25]; (2) age between 18 and 70 years; (3) painful diabeticneuropathy with reduced sensory ami/or MNCV «40 m/sin at least one nerve trunk of lower limbs); and (4) vibrationperception at big toe >25 V. Exclusion criteria were: (1) thepresence of any other severe disease; (2) pregnancy; (3)renal disease with serum creatinine levels > 1.77 I-lmol/l;(4)a history or actual presence of foot ulcers; and (5) lowerlimb vasculopathy as indicated by an ankle-brachial index<0.9 or a transcutaneous partial pressure of oxygen <50mmHg. Any analgesic or other drug administered far thechronic treatrnent of painful neuropathy was discontinued atleast 3 weeks before randomisation. Patients were enrolledat two centres: Milan and Perugia. The study protocol wasapproved by the ethics committees of San Raffaele Uni-versity Hospital and Perugia University Hospital, and writ-ten informed consent was obtained frOlliaHpatients prior toenrolment.

Clinica l assessments Patients were evaluated four times: atbaseline, at the end o(each series, and 4 months after thecompletion of the study. Each patient saw the same phy-

819

sician far the clinical examination, the administration of locally by ionic exchange HPLC (Variant 2; Biorad, Milan,questionnaires and the assessment of side effects; two neu- Italy) at baseline and at the end of the two treatment series.rophysiologists, one in Milan and one in Perugia, performedelectroneurography examinations.

Outcome measures The parameters described below weremeasured on four occasions. Daytime and night-time painwere measured by a 0-100 VAS (results expressed as ab-salute numbers) [26]. Tactile sensation was assessed by the5.07 (lO g) Semmes-Weinstein monofilament test (thenumber of insensitive areas in the two feet out of nine stan-dard areas per foot-the first, third and fifth finger andrespective metatarsal heads, two mid-foot points and heel-afe assessed; only data far the right foot were considered foranalysis) [27]. Foot vibration perception threshold wasmeasured using a biothesiometer (Bio-MedicaI InstrumentCompany, Newbury, OH, USA) [28]. Electroneurographywas used to measure MNCVand SNCV in m/sec accordingto the standard procedure [29]; a Micromed System 98 wasused in Milan (Myoquick, Treviso, Italy), whereas a MedelecPremier Plus (TECA, NY, USA) was used in Perugia.MNCV was measured in the peroneal nerve and SNCV wasmeasured antidromically in the stirai nerve. Quality of lifewas assessed using the Short Form-36 questionnaire (SF36)[30]. The SF36, which was originally used to assess healthstatus in the MedicaI Outcomes Study, was calculated on atotal of 142 levels, where the highest score represents thehighest quality of life. The scale covers the following eightdomains: generaI health; physical functioning; role limita-tion due to physical and social functioning; social function-ing; bodily pain; generaI mental health; role limitation dueto emotional problems; and vitality. HbAjc was measured

Statistical analysis The ANOVA test for repeated measureswas used to analyse changes in alI variables. A Neuman-Keuls multiple comparison test was used in post-test anal-yses. A crossover model was used to evaluate both thecarryover and treatment effects [31]. A paired t-test wasused to analyse the changes in alI variables during FREMSor placebo. Descriptive statistics afe reported as means:1:SE.Comparison ofthe effects ofFREMS with those ofplacebowere made using alI FREMS and placebo series. In ad-dition, the effect of FREMS at 4-month follow-up wasanalysed by comparison with baseline values. Ap value lessthan 0.05 was considered statistically significant.

Results

Subjects Between October 2001 and December 2003, 38patients were screened and 31 were enrolled in the studyand randomised to sequence l or sequence 2. The charac-teristics of the two sequence assignment groups were notsignificant1y different at baseline (Table l). MNCV wasmeasured in 26 patients and SNCV was measured in 15patients, due to unrecordable nerve action potentials in theremainder, and the results ofthe Semmes-Weinstein mono-filament test were only recorded for the 12patients enrolledin Milan. Data analysis was based on 31 patients for alIother measures. Although no specific attempts were madeto improve diabetes contrai during the study, HbAjc sig-nificant1y decreased frOllibaseline to the end ofFREMS and

Table 1 Characteristics of the Characteristic Sequence 1 Sequence 2 p valuepatients at baseline

Age (years) 63.1:1:3.1 59.2:1:3.1 0.1863

Duration of diabetes (years) 15.9:1:3.0 16.6:1:2.7 0.9038

Type of diabetes (type 1/type 2) 3/12 5/11

Diabetes management (insulin/oral agents/diet) 5/1010 8/7/1n 15 16

HbAjc (%) 8.3:1:0.4 8.2:1:0.3 0.9555

n 15 16

VAS daytime pain score 32.3:1:6.8 41.4:1:8.0 0.2034n 15 16

VAS night-time pain score 36.3:1:6.3 45.5:1:8.2 0.727

n 15 16

VPT (V) 35.1:1:2.3 36.0:1:2.3 0.7836n 15 16

Monofilament (out ofnine standard areas) 5.9:1:1.4 5.7:1:1.1 0.7566

n 6 6

MNCV (m/s) 36.1:1:1.4 35.0:1:2.0 0.2031

VPT vibration perceptionn 13 13

threshold; VAS visual analogue SNCV (mls) 26.7:1:3.7 29.2:1:4.6 0.93

scale; MNCV motor nerve con- n 7 8

duction velocity; SNCV sensory SF36 103.5:1:2.1 103.8:1:2.2 0.3755nerve conduction velocity; SF36 n 15 16Short Form-36 questionnaire

820

Table 2 Effeet of treatment with FREMS or plaeebo

n Plaeebo

Pre-treatmentp value FREMS p value

Post-treatment Pre-treatment Post-treatment

VPT vibration pereeption threshold; VASVigilaianalogue scale; MNCV motor nerve eonduetion veloeity; SNCV sensory nerve eonduetionveloeity; SF36 Short Form-36 questionnaire

plaeebo sequenees (8.5::1:0.3vs 7.9::1:0.2%, p=O.OOl), mogilikely as a trial effeet.

Efficacy Within-treatment analysis showed that, followingFREMS treatment, there was a signifieant deerease in day-time paio seore (p=0.0025) and night-time paio seore (p=0.0107), a signifieant deerease in the number of pointsinsensitive to the Semmes-Weinstein monofilament (p=0.0077), a signifieant deerease in the vibration pereeptionthreshold (p=0.0001), and a signifieant inerease in MNCV(p=0.0019). Non-signifieant trends towards improvementsin quality oflife and SNCV were also observed after FREMS.None ofthe outeome measures ehanged signifieant1yduringtreatment with plaeebo (Table 2). No earryover effeet wasevident within the erossover analysis.

When eompared with baseline measurements, the resultsof the 4-month follow-up showed the persistenee of sta-tistieally signifieant ehanges in alI the parameters modifiedby FREMS during treatment, including: daytime paio seore(p<0.01); night-time paio seore (p<0.01); vibration percep-tion threshold (p<0.05); sensory perception, as assessed bymonofilament (p<0.00 l); and MNCV (p<0.05). In addition,significant improvements were observed far overall qualityof lire (p<0.001); the domains far which significant im-

provements were observed were bodily paio, social fune-tioning, physical functioning, role limitation to physical andsocial functioning, and generai mental health (all p<0.05),while no significant improvement was shown in generaihealth pereeption, role limitation to emotional problems, orvitality (Table 3). A non-significant trend towards improve-ment of SNCV was alg~ observed.

Safety No systemic side effects were recorded during thestudy. Patients reported only a very slight buming sensa-tion at the gite of electrode plaeement during the series oftreatments later revealed as FREMS, with no residual skinsigns. No particular perception was recorded during placeboseSSlons.

Discussion

We bave demonstrated that FREMS is a gare and effectivetherapy far neuropathic paio in diabetic patients with pe-ripheral neuropathy and that it is able to modifY some oftheparameters of peripheral nerve function. The results of ourtwo-centre, randomised, double-blind, placebo-controlled,

VAS daytime pain seore 31 31.2:1:3.9 31.9:1:4.2 ns 37.1:1:5.3 26.2:1:3.9 0.0025

VAS night-time pain seore 31 33.3:1:3.8 30.4:1:4.2 ns 38.1:1:5.5 28.5:1:3.8 0.0107

VPT (V) 31 34.7:1:1.6 34.2:1:1.6 ns 35.5:1:1.6 33.4:1:1.6 0.0001

Monofilament (out ofnine standard areas) 12 5.1:1:0.9 5.2:1:0.9 ns 5.8:1:0.8 4.6:1:0.9 0.0077

MNCV (m/s) 26 37.2:1:1.2 37.4:1:1.4 ns 35.7:1:1.3 40.5:1:1.8 0.0019

SNCV (m/s) 15 31.9:1:2.1 30.2:1:2.7 ns 30.4:1:2.9 33.0:1:2.3 nsSF36 31 104.4:1:1.5 105.9:1:1.5 ns 103.7:1:1.5 105.6:1:1.3 ns

Table 3 Comparison of results n Baseline 4-month follow-up p valueat 4 months with baseline values

VAS daytime pain seore 31 37.0:1:5.3 25.1:1:4.2 <0.01

VAS night-time pain seore 31 41.1:1:5.2 26.5:1:3.9 <0.01VPT 31 35.6:1:1.6 31.7:1:1.8 <0.05Monofilament 12 5.8:1:0.8 4.7:1:0.9 <0.01

MNCV (m/s) 26 35.5:1:1.2 38.7:1:1.4 <0.05

SNCV (m/s) 15 26.6:1:2.8 33.1:1:3.1 nsSF36 31 103.6:1:1.5 107.9:1:1.2 <0.001Generai health 31 4.9:1:0.3 4.9:1:0.2 ns

Physieal funetioning 31 23.1:1:0.9 25.0:1:0.7 <0.05

Role limitation due to physieal and soeial 31 6.1:1:0.3 6.6:1:0.3 <0.01

funetioning

VPT vibration pereeptlonSoeial funetioning 31 8.5:1:0.4 9.2:1:0.3 <0.05

threshold; VASVigilaianalogue Bodily pain 31 6.2:1:0.4 6.8:1:0.3 <0.05

scale;MNCV motor nerve eon- Generai mental health 31 37.9:1:0.7 39.0:1:0.5 <0.05duetion veloeity; SNCV sensory Role limitation due to emotional problems 31 3.5:1:0.2 3.4:1:0.1 nsnerve eonduetion veloeity; SF36 Vitality and health pereeption 31 13.4:1:0.3 13.0:1:0.4 nsShort Form-36 questionnaire

crossover clinical trial show that the beneficial effect ofFREMS is superior to any placebo influence.

Therapy far neuropathic pain is generally based on anal-gesic drugs, bui the frequency of side effects and the lack ofefficacy of these agents in a significant proportion of caseshas spurred the search for non-pharrnacological treatments.FREMS is significant1y different from other known elec-trotherapy systems because of the modulation of the ITe-quency, amplitude and duration ofthe electrical stimuli. Atthe end of FREMS sessions, we observed a significant re-duction ofpain (up to -29% for daytime pain and -25% fornight-time pain as measured by a VAS). As expected, somepain reduction was also observed after placebo, althoughthe reduction was only statistically significant after FREMS.This non-invasive treatment was demonstrated to be safeand was not associated with any side effects. Although thestudy was blind, patients reported some perceptions at the siteof electrode placement during several sessions ofthe treat-meni later revealed as FREMS; this dicinoi happen duringsessions later revealed as placebo. However, it is unlikelythat these subjective perceptions influenced the study, sinceneither the patients nor the investigators were aware ofwhether the placebo was a nonsense electrical stimulationor a lack of stimulation. At most, only a marginaI influenceon some of the less objective measures, such as tactile orvibration perceptions, might be hypothesised. Our resultsalso demonstrate that the efficacy ofFREMS is maintainedfor at least 4 months, as both daytime and night-time painremained significant1y reduced compared with baseline atthe 4-month follow-up. This finding could be specific toFREMS, since none ofthe non-pharrnacological treatmentsfar painful diabetic neuropathy investigated to date havereported a beneficial effect lasting formare than a few weeks.Our results add to increasing evidence suggesting that non-pharrnacological tools may be useful in the treatment ofneuropathic pain. Of these modalities, FREMS could beparticularly advantageous in terrns of effect duration.

In our study, in addition to its analgesic effect, FREMSwas shown to improve several other functional peripheralnerve parameters. On average, at the end of active treat-meni, MNCV was increased by almost 5 mIs; vibrationperception threshold, as measured by a biothesiometer, wasreduced by more than 2 V; and the number of foot pointsinsensitive to the Semmes-Weinstein monofilament wasdecreased by 1.2. Although the monofilament observationis limited by the small number of patients assessed, thesefindings demonstrate that peripheral nerve function is ame-liorated after treatment with FREMS. It is unlikely that thesefindings afe the result ofthe observed improvement in bloodglucose contrai, indicated by the significant decrease inHbA lc during the study. As demonstrated by the DiabetesContrai and Complications Trial [2] and in studies on pan-creas transplant patients [3,4], strict glucose contrai is ableto prevent or halt the progression of an already establishedperipheral neuropathy. However, such effects afe seen avera long period oftime, whereas in this study the symptomaticimprovement induced by FREMS was measurable duringthe 3 weeks of active treatment.

821

The extent to which the MNCV increased during thestudy was remarkable: a ~5-mIs increase was observed atthe end of FREMS treatment and a >3-mIs increase wasmaintained at the 4-month follow-up. A meta-analysis ofclinical studies using aldose reductase inhibitors, which in-cluded 19 clinical trials using four different drugs, reporteda reduction of 0.53 mIs in the rate of decrease of peronealnerve conduction velocity aver a median period of 6 months[32]. According to our results, FREMS appears to be supe-rior to these pharrnacological agents. Furtherrnore, the im-provement obtained after FREMS is greater than 2.2 mIs,the value indicated by the Peripheral Nerve Society as theminimal increase in peroneal nerve conduction velocity atwhich clinical amelioration can be appreciated [33].

The mechanism of action of FREMS was not investi-gated in this study, and can therefore only be speculatedupon. Based on a number of assumptions, an attractive hy-pothesis is that FREMS stimulates the release of vasoactivefactors, which results in an increase in endoneural bloodflow, thus reversing the endoneural microvessel disease andassociated nerve hypoxia. Indirect findings that may berelated to the effects ofFREMS include: (1) the induction ofthe synthesis of vascular endothelial growth factor (VEGF)and other angiogenic factors and angiogenesis promotionby different electrical [34-36] or electromagnetic [37] stim-uli; (2) the reversal of diabetic neuropathy in an experimentalmodel by VEGF gene transfer [38]; (3) increased conduc-tion velocity mediated by an increase in endoneural bloodflow after electrical stimulation of peripheral nerves in amodel of experimental diabetic neuropathy [39]; and (4)increased nerve conduction velocity after an improvementin blood flow in the lower limbs, achieved through eitherrevascularisation [40] or physical exercise [41]. Altema-tively, FREMS might act on neuron sodium channels. Re-ceni evidence suggests that sodium channel expression inprimary sensory neurons is altered in diabetic neuropathy[42], indicating a possible molecular basis far neuropathicpain. Given the evidence that exogenous electric fields in-duce cellular responses that involve the redistribution ofintegraI membrane proteins, including calcium channels[43], a similar effect on sodium channels may be hypoth-esised. It is noteworthy that the subjects enrolled in ourstudy were affected by relatively severe neuropathy, whichis norrnally associated with a decrease in the number ofmyelin fibres and altered endoneural vessels. Thus, it islikely that the observed improvements in MNCVana vibra-tion perception threshold at the end of FREMS treatmentsimply reflect some functional changes of the nerve; struc-turaI changes may only occur at a later stage, possibly pro-viding the basis for the long-terrn effects of this therapy.Nonetheless, studies on experimental models ofneuropathyafe needed in order to clarify the mechanisms that underliethe effects ofFREMS.

In conclusion, the results of this controlled clinical trialdemonstrate the analgesic efficacy of FREMS, and showadditional beneficial effects of this novel electrotherapy onperipheral nerve function in patients with peripheral diabeticneuropathy. If confirrned in a larger series of cases, and

822

possibly in a multicentre study, these findings may offernew perspectives for the treatment of diabetic neuropathy.

Acknowledgements This study was supported in part by a researchgrani from Lorenz Biotech (Medolla, Italy). The funding companyhad no Talein the design and conduci of the study, the analysis andinterpretation of the data, or in the preparation and review of themanuscript.

References

I. VinikAI, Park TS, Stansberry KB, Pittenger GL (2000) Diabeticneuropathies. Diabetologia 43:957-973

2. DCCTIEDICResearch Group (2002) Effect ofintensive therapyon the microvascular complications of type I diabetes mellitus.lAMA 287:2563-2569

3. Martinenghi S, Corni G, Galardi G, Di Carlo V,Pozza G, SecchiA (1997) Amelioration of nerve conduction velocity followingsimultaneous kidney/pancreas transplantation is due to the gly-caemic contrai provided by the pancreas. Diabetologia 40:1110-1112

4. Navarro X, Sutherland DER, Kennedy WR (1997) Long-terrneffectsofpancreatic transplantation on diabetic neuropathy. AnnNeuroI42:727-736

5. Brownlee M (2001) Biochemistry and molecular celIbiology ofdiabetic complications. Nature 414:813-820

6. Pfeifer MA, Schumer MP, Gelber DA (1997) Aldose reductaseinhibitors: the end of an eraor the need far different trial designs?Diabetes 46(SuppI2):S82-S89

7. Benbow Sl, Cossins L, MacFarlane lA (1999) Painful diabeticneuropathy. Diabet Med 16:632-644

8. Backonja M, Beydoun A, Edwards KR et al (1998) Gabapentinfar the symptomatic treatment of painful neuropathy in patientswith diabetes mellitus: a randomized controlled trial. lAMA280:1831-1836

9. Eisenberg E, Lurie Y, Braker C, Daoud D, Ishay A (2001)Lamotrigine reduces painful diabetic neuropathy: a randomized,controlled study. Neurology 57:505-509

IO. Watson CP, Moulin D, Watt-Watson1, Gordon A, Eisenhoffer 1(2003) Controlled-release oxycodone relieves neuropathic pain:a randomized controlledtrial in painful diabetic neuropathy.Pain105:71-78

Il. Abuaisha BB, Costanzi JB, Boulton Al (1998) Acupuncture farthe treatmentof chronic painful peripheral diabetic neuropathy: along-terrnstudy. Diabetes Res Clin Pract 39:115-121

12. Leonard DR, Farooqi MH, Myers S (2004) Restoration ofsensation, reduced pain, and improved balance in subjects withdiabeticperipheralneuropathy.A double-blind,randomized, pla-cebo-controlled study with monochromatic near-infrared treat-ment. Diabetes Care 27:168-172

13. Zinman L, Ngo M, Ng ET, Nwe KT, Gogov S, Brii V (2004)Low-intensity laser therapy far painful symptoms of diabeticsensorimotor polyneuropathy. A controlled trial. Diabetes Care27:921-924

14. WeintraubMI, WolfeGl, Barohn RA et al (2003) Staticmagneticfield therapy far symptomatic diabetic neuropathy: a random-ized, double-blind, placebo-controlled trial. Arch Phys MedRehabiI84:736-746

15. WeintraubMI, Cole SP (2004) Pulsed magnetic field therapy inrefractory neuropathic pain secondary to peripheral neuropathy:electrodiagnostic parameters-pilot study. Neurorehabil NeuralRepair 18:42-46

16. Kumar D, Marshall Hl (1997) Diabetic peripheral neuropathy:ameliorationof pain with transcutaneouse1ectrostimulation.Dia-betes Care 20:1702-1705

17. Kumar D, Alvaro MS, lulka IS, Marshall Hl (1998) Diabet-ic peripheral neuropathy. Effectiveness of electrotherapy andamitriptyline far symptomatic relief. Diabetes Care 21:1322-1325

18. Hamza MA, White PF, Craig WF et al (2000) Percutaneouselectricalnerve stimulation:a novel analgesictherapy fordiabeticneuropathic pain. Diabetes Care 23:365-370

19. Tesfaye S, Watt1,Benbow Sl, Pang KA, Miles 1,MacFarlane lA(1996) Electrical spinal-cord stimulation far painful diabeticperipheral neuropathy. Lancet 348:1698-1701

20. Neumann V (1993) Electrotherapy. Br 1 RheumatoI32:1-221. Baker LL, DeMuth SK, Chambers R, Villar F (1997) Effects of

electrical stimulation on wound healing in patients with diabeticulcers. Diabetes Care 20:405-412

22. Bevilacqua M, Barrella M, Toscano R et al (2004) Disturbancesof vasomotion in diabetic (type 2) neuropathy: increase ofvascular endothelial growth factor, elicitation of sympatheticefflux and synchronization ofvascular flow (vasomotion) duringfrequency modulated neural stimulation (FREMS). 86th AnnualMeeting ofthe Endocrine Society, p 321, P 2-61 (abstract)

23. Donague VM, Giurini 1M, Rosenblum BI, WeissmanPN, VevesA (1995) Variability in function measurements ofthree sensoryfoci nerves in neuropathic diabetic patients. Diabetes Res ClinPract 29:37-42

24. Pitei DL, Watkins Pl, Stevens Ml, Edmonds ME (1994) Thevalue of the Neurometer in assessing diabetic neuropathy bymeasurement of the current perception threshold. Diabet Med11:872-876

25. Genuth S, Alberti KG, Bennett P et al (2003) Expert committeeon the diagnosis and classification of diabetes mellitus. Follow-up report on the diagnosis of diabetes mellitus. Diabetes Care26:3160-3167

26. Ohnhaus EE, Adler R (1975) Methodological problems in themeasurement of pain: a comparison between the verbal ratingscale and the visual analogue scale. Pain 1:379-384

27. Kumar S, Fernando Dl, Veves A, Knowles EA, Young Ml,Boulton Al (1991) Semmes-Weinstein monofilaments: a sim-pie, effective and inexpensive screening device far identifyingdiabetic patients at risk of foci ulceration. Diabetes Res ClinPract 13:63-67

28. Bloom S, Till S, Sonksen P, Smith S (1984) Use of a bio-thesiometer to measure individuai vibration thresholds and theirvariation in 519 non-diabetic subjects. Br Med 1288:1793-1795

29. Preston DC, Shapiro BE (1998) Electromyography and neuro-muscular disorders. Butterworth-Heinemann, Boston, pp 103-142

30. Ware JE lr, Sherbourne CD (1992) The MOS 36-item short-forrnhealth survey (SF-36). I. Conceptual framework and itemselection. Med Care 30:473-483

31. Koch GG (1972) The lise of non-parametric methods in thestatistical analysis of the two-period change-over designoBio-metrics 28:577-584

32. Airey M, Bennett C, Nicolucci A, Williams R (2000) Aldosereductase inhibitors far the prevention and treatment of diabeticperipheral neuropathy. Cochrane Database Syst Rev CDO02182

33. Peripheral Nerve Society (1995) Diabetic polyneuropathy incontrolled clinical trials: consensus report of the PeripheralNerve Society. Ann Neurol 38:478-482

34. Zhao M, Bai H, Wang E, Forrester N, McCaig CD (2004)Electrical stimulation directly induces pre-angiogeniclesponsesin vascular endothelial cells by signaling through VEGF recep-torso1 CelI Sci 117:397-405

35. Hang 1, Kong L, Gu JW, Adair TH (1995) VEGF geneexpression is upregulated in electrically stimulated rat skeletalmuscle. Am 1 PhysioI269:HI827-HI831

36. Kanno S, Oda N, Abe M et al (1999) Establishment ora simpleand practical procedure applicable to therapeutic angiogenesis.Circulation 99:2682-2687

37. Tepper OM, Callaghan Ml, Chang EI et al (2004) Electro-magnetic fields increase in vitto and in vivo angiogenesisthrough endothelial release ofFGF-2. FASEB 118:1231-1233

38. Schratzberger P, Walter DH, Rittig K et al (2001) Reversal ofexperimental diabetic neuropathy by VEGF gene transfer.1ClinInvest 107:1083-1092

39. Cameron NE, Cotter MA, Robertson S, Maxfield EK (1993)Nerve function in experimental diabetes in rats: effects ofelectrical stimulation. Am J Physio1264:E16l-E166

40. Young MJ, Veves A, Smith JV, Walker MG, Boulton AJM(1995) Restoring lower limb blood flow improves conductionvelocity in diabetic patients. Diabetologia 38:1051-1054

41. Tesfaye S, Harris ND, Wilson RM, Ward JD (1992) Exercise-induced conduction velocity increment: a marker of impairedperipheral nerve blood flow in diabetic neuropathy.Diabetologia35:155-159

823

42. Craner MJ, Klein JP, Renganathan M, Black JA, Waxman SG(2002) Changes of sodium channel expression in experimentalpainful diabetic neuropathy. Ann Neurol 52:786-792

43. Cho MR, Thatte HS, Silvia MT, Golan DE (1999) Transmem-brane calcium influx induced by ac electric fie1ds. FASEB J13:677-683