value of tens for relief of chronic low back pain with or without radicular pain

ORIGINAL ARTICLE

Value of TENS for relief of chronic low back pain with orwithout radicular painA. Buchmuller1, M. Navez2, M. Milletre-Bernardin3, S. Pouplin4, E. Presles1, M. Lantéri-Minet3, B. Tardy1,B. Laurent2, J.P. Camdessanché2, on behalf of the Lombotens Trial Group*

1 Centre d’investigation Clinique – Epidémiologie Clinique/Essais Cliniques (CIE3), CHU de Saint-Etienne, 42055 Saint-Etienne Cedex 02, France

2 Centre d’Évaluation et de Traitement de la Douleur (CETD), CHU de Saint-Etienne, 42055 Saint-Etienne Cedex 02, France

3 Département d’Évaluation et de Traitement de la Douleur, Médecine Palliative, Pôle Neurosciences Cliniques, CHU de Nice, 06002 Nice

Cedex, France

4 Centre d’Évaluation et de Traitement de la Douleur, CHU de Rouen, 76031 Rouen Cedex, France

*Lombotens Trial GroupWriting committee – A. Buchmuller, E. Presles, B. Tardy (Inserm CIE3, CHU de Saint-Etienne, Saint-Etienne, France), M. Navez, B. Laurent, JP. Camdes-

sanché [Centre d’évaluation et de traitement de la douleur (CETD), CHU de Saint-Etienne, Saint-Etienne, France], M. Milletre-Bernardin, M. Lantéri-

Minet (Centre d’Evaluation et de traitement de la douleur, CHU de Nice, Nice, France), S. Pouplin (Centre d’Evaluation et de traitement de la douleur,

CHU de Rouen, Rouen, France).

Coordinating Centre – A. Buchmuller, J. Techer, G. El Asri (Inserm, CIE3, CHU de Saint-Etienne, Saint-Etienne, France).

Data Monitoring and Statistical Analysis – E. Presles, S. Laporte (Inserm, CIE3, CHU de Saint-Etienne, Saint-Etienne, France).

Investigators (number of patients recruited) – M. Navez, B. Pandraux, Hôpital Bellevue, St-Etienne (31), P. Nayme, Hôpital Nord, St-Etienne (23),

M. Lanteri Minet, M. Milletre-Bernardin, V. Lussiez, Nice (52), G. Ducable, S. Pouplin, Rouen (31), V. Dousset, Bordeaux (11), E. Touchard-Toulallan, Caen

(10), P. Roussel, Marseille La Timone (9), E. Serra, L. Douay, Amiens (9), G. Mick, G. Puech, Voiron (8), P. Mertens, Lyon (8), N. Cantagrel, B. Grandmottet,

Toulouse (7), B. Planchet Barraud, Marseille St-Joseph (7), S. Blond, N. Leleu, FX Derousseaux, Lille (5), P. Ginies, Dr, D. Kong A, Siou, Montpellier (5),

L. Boutault, I. Vannier, Tours (4), A. Lepeintre, Evreux (4), JP. Estebe, Rennes (3), P. Vergne-Salle, Limoges (3), F. Tiberghien, JL. Lajoie, V. Piccand,

Besançon (3), A. Serrie, Paris Lariboisière (1), M. Sorel, H. Vrolyk Vergeron, Nemours (1), C. Maindet-Dominici, C. Garin, Reseau Algo 38 de Grenoble (1).

Contributors: All authors are members of the LOMBOTENS Trial Group. AB represented the Coordinating Centre and was a member of the Writing

Committee; MN, MMB, SP and MLM were investigators and members of the Writing Committee; EP was responsible for data monitoring and statistical

analysis and was a member of the Writing Committee; BT, BL and JPC were members of the Writing Committee.

+Trial registration: Clinicaltrials.gov: NTC 00452010

Ethical approval: This study was approved by the Comité de Protection des Personnes (ethics committee) Sud-Est I, Saint-Etienne, France.

CorrespondenceAndréa Buchmuller

Tel.: +33 477127788; fax: +33 477127820.

E-mail: [email protected]

Funding sourcesThis study was funded by the Direction

Générale de la Santé (Ministère de la Santé et

des Sports), the Fondation CNP Assurances

and the Institut UPSA Douleurs. Sham TENS

devices were provided free of charge by

CEFAR France. The funding bodies played no

role in the research process, the writing of

the report or the decision to submit the

article for publication.

Conflicts of interestOne author (B.L.) is a member of a board

(Lilly).

Accepted for publication27 September 2011

doi:10.1002/j.1532-2149.2011.00061.x

Abstract

Objectives: To evaluate the efficacy of transcutaneous electricalneurostimulation (TENS) in patients with chronic low back pain (LBP).Methods: Design: Prospective, randomized, multicentre, single-blindstudy. Setting: Twenty-one French pain centres. Participants: Two hundredthirty-six consecutive adult patients consulting for chronic LBP, with orwithout radicular pain (mean age � standard deviation: 53 � 13 years;range: 28–86 years). Intervention: Patients were randomly assigned toreceive either active (n = 117) or sham (n = 119) TENS in four 1-h dailytreatment sessions for 3 months. Main outcome measures: The primaryoutcome measured was improvement of functional status at 6 weeks(Roland–Morris Disability Questionnaire). Secondary outcome measureswere improvement of functional status at 3 months, pain relief (weeklyvisual analogue scale assessments), positive functional repercussions ofpain levels on quality of life, a diminution of the use of analgesic andanti-inflammatory medication, satisfaction with the overall treatmentstrategy and compliance.Results: Functional status did not differ between the groups, whether at6 weeks or 3 months (p = 0.351 at 6 weeks). A significant improvementbetween the first and last visual analogue scale assessments was observed

656 Eur J Pain 16 (2012) 656–665 © 2011 European Federation of International Association for the Study of Pain Chapters

in patients with either lumbar pain alone or lumbar and radicular paintreated with active TENS. Other outcome measures did not differsignificantly between the two groups.Conclusion: There was no functional benefit of TENS in the treatment ofpatients with chronic LBP.

1. Introduction

Low back pain (LBP) is a leading cause of work absen-teeism and visits to healthcare professionals (Anders-son, 1999; Deveraux, 2004). Around 60–90% of theadult population is at risk of developing LBP at somepoint in their lifetime. It is estimated that 10–20% ofaffected adults develop symptoms of chronic LBP,defined as persistent pain occurring on most days andlasting longer than three consecutive months (VonKorff and Saunders, 1996; Hildebrandt et al., 2004;Maher, 2004). Chronic LBP has a significant impact onfunctional status, restricting occupational activitieswith marked socio-economic repercussions (VanTulder et al., 1999; Becker et al., 2011). The use ofmultiple complementary treatments is frequent andactive therapies encouraging individuals to participatein the treatment process are increasingly advocated(Dogan et al., 2008). One of these methods is transcu-taneous electrical nerve stimulation (TENS), a non-invasive treatment that can be self-administered bypatients and is generally associated with few safetyconcerns, adverse effects being principally limited totransient skin irritation. TENS is currently widely usedas an add-on therapy for chronic LBP patients(Buchmuller-Cordier et al., 2008; Carey et al., 2009).

Despite wide use of this treatment, its efficacy inrelieving chronic LBP has not been established: in thelast review of the Cochrane Database (Khadilkar et al.,2008), only four studies meeting the criteria for highmethodological quality were included in the qualita-tive analysis (Deyo et al., 1990; Cheing and Hui-Chan,1999; Topuz et al., 2004; Jarzem et al., 2005). Thisreview presented conflicting evidence as to whetherTENS was beneficial in reducing back pain intensity asassessed in 235 randomized patients (three studies)using a visual analogue scale (VAS), and consistentevidence that it did not improve back-specific func-tional status in 410 randomized patients (two studies)using different but validated scales.

We therefore conducted a randomized, single-blindstudy in adult patients consulting French pain centreswith chronic LBP, to compare the efficacy of active andsham TENS in terms of functional disability, short- andlong-term pain relief, quality of life (QOL) and anal-

gesic use. To better define the appropriate clinicalcontext for the use of TENS, in terms of neuropathic ornociceptive pain, both patients presenting LBP with orwithout radicular pain were enrolled and the efficacyof TENS was assessed in these two subgroups (Deyoet al., 1990; Carroll et al., 2000; Jarzem et al., 2005;Khadilkar et al., 2008).

2. Methods

Patients aged over 18 years, covered by a nationalhealth insurance policy, and who had consulted a paincentre during the previous week for chronic LBPgraded on average � 40 on a VAS, with or withoutradicular pain, were potentially eligible for inclusion.Diagnosis of a radicular pathology was confirmedbased on the clinical examination. Patients were ineli-gible for the study if they had been previously treatedwith TENS; if they had undergone surgery for radicu-lopathy within the last 3 months; if they had experi-enced LBP for less than 3 months; if LBP wasassociated with bilateral radiculopathy; if they hadacute or non-stabilized radicular syndrome (sciatalgia,cruralgia); if any surgery was planned within the next6 months; if they had a pacemaker; if other non-pharmacological treatments were planned, includingphysiotherapy, acupuncture, mesotherapy, manipula-tions, wearing a corset, or psychological support; iftheir LBP was symptomatic of another condition (i.e.,compression fractures or progressive inflammatory,neoplastic or infectious conditions); if the physicianhad estimated their life expectancy to be less than 3months; if articular or foraminal infiltrations wereplanned during the study period; or if the patient wasinvolved in an ongoing medico-legal dispute.

2.1 Randomization and data collection

This study was approved by the Comité de Protection desPersonnes (ethics committee) Sud-Est I, Saint-Etienne,France. Eligible patients were included after havingread an information sheet detailing the TENS proce-dure and noting its potential benefits with regard topain relief even when no tingling was perceived (as inthe sham TENS procedure). After written informed

A. Buchmuller et al. Value of TENS for relief of chronic low back pain

657Eur J Pain 16 (2012) 656–665 © 2011 European Federation of International Association for the Study of Pain Chapters

consent had been obtained, randomization was per-formed at a central location with stratification bycentre and by type of chronic LBP (with or withoutradicular pain). To determine the treatment group towhich each patient should be assigned, the investiga-tor called the study coordination centre, whichassigned the patient according to the randomizationlist established for each centre before the start of thestudy, based on a computer-generated pseudo-randomnumber sequence. The data were collected andrecorded by the Centre d’Investigation Clinique de Saint-Etienne (CIE3).

2.2 Interventions

2.2.1 TENS apparatus and procedure

Active TENS treatment combined two types of electri-cal stimulation:(1) Conventional TENS (‘gate control’), characterizedby continuous stimulation at high frequencies (80–100 Hz), with wave durations of 50–100 ms and lowintensities, potentially achieving painless paraesthesiain the part of the body concerned.(2) ‘burst’ TENS (acupuncture-like TENS), character-ized by discontinuous stimulation at low frequencies(1–4 Hz), with wave durations of 100–400 ms and highintensities, inducing weak muscle twitches.

The treatment was self-administered using a CEFARPrimo Pro device and program P5, which associatescontinuous high-frequency (80 Hz) stimulation withbursts of low-frequency (2 Hz) stimulation every 3 s.For patients suffering from chronic LBP withoutradicular pain, two of the four electrodes were placedon healthy skin on each side of the painful area. Forthose experiencing LBP associated with radicular pain,two electrodes were placed spanning the painful areaof the back and two electrodes on the trajectory of thetroncular nerve involved in the radiculopathy. At thestart of stimulation, the intensity was increased pro-gressively until the onset of painless tingling, or up to25 mA if no sensation was perceived by the patient.After the device’s screen read zero, the intensity wasfurther increased until tolerable impulses wereperceived or up to 30 mA in the absence of any sen-sation. Stimulation was then continued for 1 h at thisintensity.

The treatment protocol was the same for patientsassigned to sham TENS treatment, except that theTENS devices provided to this group did not deliver anelectrical current. The four electrodes were placed asdescribed above, depending on whether LBP was asso-ciated with radicular pain. The sham treatment was

self-administered, using a CEFAR Primo Pro deviceand program P5. The term ‘placebo’ was not used inthe information leaflet. The leaflet explained that thestudy aimed to evaluate currents of much more vari-able intensity than usual TENS, and that both TENStechniques could trigger sensations such as tingling.The leaflet also stated that this did not occur in allcases and that there was no relation between tinglingand TENS efficacy.

Irrespective of the treatment group, patients wereinstructed to complete four 1-h treatment sessions perday for a total of 3 months. Both active and shamTENS devices incorporated a mechanism which auto-matically recorded stimulation frequency and dura-tion, which were checked at each study visit to verifycompliance.

2.2.2 General study procedure and follow-up

On the day of inclusion, the patient’s demographicdata, clinical history, and ongoing analgesic,co-analgesic and other medication were recorded andthe patient independently completed a self-evaluationquestionnaire including the Roland–Morris DisabilityQuestionnaire (RDQ), the Dallas pain questionnaireand the SF-36 QOL questionnaire. The DN4 Neuro-pathic Pain Diagnostic Questionnaire (Bouhassiraet al., 2004) and the Quebec Task Force classification(Spitzer, 1987) were also completed on the day ofinclusion, as well as the mean and maximum levels ofpain during the week preceding the consultation,using a VAS ranging from 0 (absence of pain) to100 mm (maximum imaginable pain), were recorded.The patient was then given a personal diary. This con-tained detailed instructions on self-administering theTENS treatment, including a silhouette showing thecorrect placement of the electrodes, as well as formsfor noting daily analgesic medication, weekly VASassessments of LBP and, if relevant, radicular pain, thenumber of TENS sessions accomplished per day, theanalgesic effect noted during and after each session,and any adverse event experienced after each session.Follow-up visits at the pain centre were scheduled at15 days (�2 days), 6 weeks (�1 week) and 3 months(�1 week) after the start of active or sham TENStreatment.

2.3 Outcome measures

The primary study objective was to evaluate the effi-cacy of TENS in terms of functional status at 6 weeks,assessed using the RDQ. A 4-point decrease was con-sidered an improvement in functional status.

Value of TENS for relief of chronic low back pain A. Buchmuller et al.


The secondary objective was to compare the twogroups with respect to pain on a weekly basis, andwith respect to functional status, global improvementand QOL at 3 months. Secondary outcome measurescomprised:(1) Pain relief evaluated on the basis of weekly VASscores for the lumbar and radicular components ofpain (both overall pain and grading LBP and radicularpain separately); the criterion for improvement was a50% decrease in the VAS score (Moore et al., 2008).(2) Functional status at 3 months, assessed using theRDQ.(3) Daily activities, professional/leisure activities,anxiety/depression, and sociability at 3 months, deter-mined by means of the Dallas questionnaire.(4) Overall satisfaction with the pain managementstrategy at 6 weeks and at 3 months (percentage ofpatients with scoring 50% or more on the satisfactionscale).(5) QOL at 3 months relative to D0, assessed using theSF-36 scale.(6) Use of analgesic and anti-inflammatory medica-tion between D0 and 3 months.(7) Compliance with TENS treatment.

2.4 Sample size

We hypothesized an improvement of the primaryoutcome at 6 weeks in 30% of patients with LBPtreated in a pain centre (Flor et al., 1992) and that50% of the patients in the active TENS group wouldachieve a decrease of at least four points on the RDQat 6 weeks compared with baseline. On this basis, withan alpha risk of 5%, it was calculated that 103 evalu-able patients per treatment group would be needed toachieve an 80% power to detect a difference betweenthe groups.

2.5 Statistical analysis

Quantitative variables were summarized by means,standard deviations, medians, minimums and maxi-mums. Categorical variables were expressed as abso-lute and relative frequencies. Comparability of thegroups at inclusion was verified using Student’s t-testfor quantitative variables (or Wilcoxon’s test if distri-bution was abnormal) and the chi-squared test orFisher’s exact test for qualitative variables. Theprimary endpoint analysis was performed on allpatients included in the study, i.e., all those havingundergone centralized randomization, according tothe intention-to-treat (ITT) principle. For primaryendpoint analysis, patients who did not complete the

6-week follow-up visit were considered as showing nofunctional improvement on the basis of the RDQ. Withregard to the secondary outcomes, patients wereincluded in the analyses of all variables for which theirdata were available. The percentages of patients withimproved functional capacity in the two treatmentgroups were compared on the basis of the estimatedrelative risk and the corresponding 95% confidenceinterval. Differences between the two treatmentgroups were considered statistically significant atp < 0.05 (two sided). Subgroup analyses according topatient initial pathology (LBP with and withoutradicular pain, DN4 score <4 and �4, Quebec TaskForce type 1 or 2 and type 3 or 4) were descriptiveonly. A heterogeneity test was performed only for theprimary endpoint in order to compare the differencesobserved in the two subgroups. The statistical analyseswere accomplished using SAS-Windows® software,version 9.1 (SAS Institute Inc, Cary, NC, USA).

3. Results

A total of 236 consecutive patients were included inthe study between 14 September 2006 and 11 June2008, in 21 pain centres located throughout France,with 117 patients being randomized to the activeTENS and 119 to the sham TENS. The majority ofpatients (139; 58.9%) were suffering from LBP asso-ciated with radicular pain. The two treatment groupswere comparable in terms of demographic and clinicalcharacteristics at inclusion, with no statistically signifi-cant difference between the groups in any of the vari-ables analysed (Table 1). Most patients (88.0%) weretaking at least one type of analgesic medication, pre-dominantly paracetamol (71.6%), with 64% takingvarious co-analgesic medications, mainly antidepres-sants. Follow-up with respect to the primary efficacyoutcome was incomplete for 28 patients (11.9%). At 3months, follow-up was incomplete for an additional34 patients (Figure 1).

3.1 Efficacy of active versus sham TENS in thestudy population as a whole

The results of the primary and secondary outcomeanalyses performed on the total study population arepresented in Table 2.

3.1.1 Primary outcome measure – improvement infunctional status at 6 weeks

At 6 weeks, the median score achieved on the RDQwas 12 in the active TENS group and 13 in the sham



TENS group, compared with a median baseline scoreof 15 in both groups. Improvement was achieved by29.9% of patients in the active TENS group and 24.3%in the sham TENS group, and the difference betweenthe groups did not reach the threshold of statisticalsignificance [RR = 1.23 (0.80; 1.89), p = 0.351].

3.1.2 Secondary outcome measures

Improvement in functional status at 3 months wasachieved by 26.4% and 25.0% of patients in the activeand sham TENS groups, respectively, with no statisti-cally significant difference between the groups[RR = 1.05 (0.67; 1.65), p = 0.816].

No statistically significant difference between thegroups was evident with regard to scores for thefour dimensions of the Dallas score at 3 months(Table 2).

An improvement of at least 50% in lumbar painbetween the first and last assessments was achieved by25.0% of patients in the active TENS group comparedwith 6.7% in the sham TENS group (p = 0.0003),

corresponding to a number needed to treat (NNT) of 5.An improvement of at least 50% in radicular painbetween the first and last assessments was achieved by33.8% of patients in the active TENS group and 15.0%in the sham TENS group (p = 0.0148), also corre-sponding to a NNT of 5. The median interval betweenthe first and last VAS pain assessments was 12 weeks(Q1–Q3: 9–13) for both patients with lumbar painalone and those with lumbar and radicular pain.

There was no statistically significant differencebetween the two groups with regard to overall sat-isfaction with the pain management strategy mea-sured at 6 weeks and 3 months. More than half ofthe patients were very satisfied, with a score exceed-ing 50% on the satisfaction scale (61.6% in theactive TENS group vs. 57.3% in the sham TENSgroup at 3 months).

No statistically significant change in QOL betweenbaseline and 3 months was detected on the basis of theSF-36 questionnaire and there was no statistically sig-nificant difference between the two treatment groupsconcerning this parameter.

Table 1 Patient characteristics at inclusion.

Parameter

Active TENS Sham TENS Total population

n = 117 n = 119 n = 236

Men/Women, n/n (%/%) 45/72 (38.5/61.5) 43/76 (36.1/63.9) 88/148 (37.3/62.7)

Age (year), mean � SD 52.9 � 13.0 53.4 � 12.9 53.1 � 12.9

Body mass index � 30 kg/m2 22 (19.3) 15 (13.2) 37 (16.2)

Description of pathology

LBP/LBP + RP 46/71 (39.3/60.7) 51/68 (42.9/57.1) 97/139 (41.1/58.9)

Interval between symptom onset and randomization

Median [range] – months 40 [1–622]a 34.5 [3–423] 36.5 [1–622]

Surgery for disk hernia 44 (37.9) 41 (34.7) 85 (36.3)

Surgery for lumbar arthrodesis 8 (6.9) 14 (11.8) 22 (9.4)

Schober index (cm), mean � SD 12.8 � 2.4 13.0 � 1.5 12.9 � 2.0

Baseline VAS (mm), mean � SD 63 � 15 66 � 17 65 � 16

Maximum VAS (mm), mean � SD 85 � 12 85 � 12 85 � 12

DN4 � 4 points 69 (61.6) 69 (61.1) 138 (61.3)

QTF classification

Type 1 27 (23.7) 32 (27.8) 59 (25.8)

Type 2 32 (28.1) 27 (23.5) 59 (25.8)

Type 3 18 (15.8) 27 (23.5) 45 (19.7)

Type 4 37 (32.5) 29 (25.2) 66 (28.8)

Roland–Morris disability scale score

Median [range] – months 15 [3–24] 15 [5–22] 15 [3–24]

Description of treatment

At least one analgesic medication per dayb 91 (90.1) 93 (86.1) 184 (88.0)

At least one co-analgesic medication per dayb 68 (66.7) 69 (62.2) 137 (64.3)

Use of class 2 analgesic drug 52 (64.2) 49 (61.2) 101 (62.7)

Non-pharmacological treatment 41 (42.7) 56 (53.8) 97 (48.5)

LBP, low back pain; QTF, Québec Task Force; RP, radicular pain; SD, standard deviation; TENS, transcutaneous electrical nerve stimulation; VAS, visual

analogue scale.aOne patient who had suffered from pain for only 1 month was included by error and randomized to the active TENS group.bThis included paracetamol, non-steroidal anti-inflammatory drugs, antidepressants, antiepileptic agents, myorelaxant agents and weak opioid drugs.



More than 50% of patients did not modify theirmedication use from inclusion to the end of the studyperiod, regardless of their assignment to active orsham TENS group. Irrespective of the medication con-sidered, the variation in the median daily dosagebetween the first and last doses did not differ signifi-cantly between the two treatment groups. The medianvariations for each medication were zero in bothgroups, signifying that half of the patients did notdecrease the dose of medication taken.

Compliance with TENS treatment was good in bothtreatment groups, with no statistically significant dif-ference between the groups.

3.1.3 Safety

Twelve patients presented a serious adverse eventduring the study: five in the active TENS group andseven in the sham TENS group. None of these eventswas considered to be attributable to the treatmentstudied. Skin irritation was observed in 11 patients in

the active TENS group (leading to study discontinua-tion in one patient) and in three patients in the shamTENS group.

3.2 Subgroup analyses according to theexistence of a neuropathic component of pain

Whatever the method used to define the presence ofneuropathic pain (clinical assessment, DN4 or QuebecTask Force scores), a trend in favour of active TENSwas observed in patients with radicular or neuropathicpain in terms of functional status (Figure 2). A similartrend was observed with regard to VAS scores for bothlumbar and radicular pain intensity (Figures 3 and 4).

4. Discussion

This multicentre single-blind study was designed toevaluate the efficacy of TENS in improving functionaldisability in patients with chronic LBP. At 6 weeks, nostatistically significant difference in functional status

Figure 1 Treatment and follow-up of study patients. Of the 22 and 31 patients in the active and sham transcutaneous electrical nerve stimulation (TENS)

groups, respectively, who dropped out of the study prematurely, all declared that the TENS procedure was too constraining. VAS, visual analogue scale.



improvement was observed between patients receiv-ing active and sham TENS and a similar result wasobtained at 3 months. In contrast, the weekly evalu-ation of pain intensity recorded in patients’ diariesshowed a highly significant difference in favour ofactive TENS in both the lumbar and radicular compo-nents of pain. This effect on pain intensity was par-ticularly marked in the subgroup of patients with LBPassociated with radicular pain. In the population as awhole, this positive effect of active TENS did not leadto either decreased medication intake or improvedQOL, despite clear overall satisfaction with the painmanagement strategy.

Our choice to include patients with and withoutradicular pain could be a matter for debate. However,there is some evidence that radiculopathy may be amixed pain syndrome including a neuropathic com-ponent (Flor et al., 1992; Baron and Binder, 2004;Moore et al., 2008), and so patients may experience a

mixture of neuropathic and nociceptive pain. In thisrespect, it is important to note that patients with andwithout radicular pain were also included in two ofthe four studies considered in the Cochrane Databasefor TENS evaluation in chronic LBP (Deyo et al., 1990;Jarzem et al., 2005; Khadilkar et al., 2008).

The relevance of the primary outcome measureused to judge the effect of TENS on functional status,a decrease of four points in the RDQ score, may bedebatable. However, a decrease of 3.5 points in theRDQ score was previously reported to be theminimum clinically important change (Ostelo and deVet, 2005), and a decrease of five points has recentlybeen proposed by an international expert panel(Ostelo et al., 2008). The inclusion of patients withand without radicular pain, as specified in the studyprotocol, permitted assessment of the efficacy of TENSin relieving both neuropathic and nociceptive pain.Interestingly, this improvement of pain intensity in

Table 2 Outcome measures.

Parameter

Active TENS

n = 117

Sham TENS

n = 119 p-value

Functional status

Improvement on Roland–Morris disability questionnaire, n/N (%)

At 6 weeks 32/107a (29.9) 28/115a (24.3) 0.351

At 3 months 29/110 (26.4)b 28/112 (25.0)b 0.816

Pain relief – VAS

Improvement of lumbar VAS between first and last VAS

�50%, n/N (%) 26/104 (25.0) 7/104 (6.7) 0.0003

Improvement of radicular VAS between first and last VAS

�50%, n/N (%) 22/65 (33.8) 9/60 (15.0) 0.015

Satisfaction with pain management strategyc

Score on satisfaction scale > 50%, n/N (%)

At 6 weeks 51/96 (53.1) 55/96 (57.3) 0.562

At 3 months 53/86 (61.6) 43/75 (57.3) 0.580

Quality of life – SF-36 (mean � SD)

At 3 months

Aggregate score for physical dimensions 35.3 � 7.3 34.3 � 7.8 0.218

Aggregate score for psychological dimensions 39.3 � 12.4 39.1 � 11.1 0.959

Functional repercussion of pain – Dallas score

At 3 months

Score for everyday activities, median [range] 69 [0–100] 69 [12–96] 0.843

Score for professional and leisure activities, median [range] 70 [0–100] 70 [10–95] 0.970

Score for anxiety and depression, median [range] 42.5 [0–95] 42.5 [0–100] 0.950

Score for sociability, median [range] 30 [0–100] 35 [0–80] 0.796

Compliance with TENS treatment

Mean number of TENS sessions per day, median [range] 3.6 [0.8–9.7] 3.7 [0.7–4.8] 0.203

SD, standard deviation; TENS, transcutaneous electrical nerve stimulation; VAS, visual analogue scale.aA complete Roland–Morris Disability Questionnaire (RDQ) performed both at inclusion and at 6 weeks was not evaluable for 10 patients in the active TENS

group and four patients in the sham TENS group.bA complete RDQ performed both at inclusion and at 3 months was not evaluable for seven patients in the active TENS group and seven patients in the

sham TENS group.cThe outcome was measured at the 6-week and 3-month visits on a scale ranging from 0% to 100%.



favour of active TENS treatment in the group ofpatients with radicular pain was observed for bothcomponents of pain: lumbar and radicular.

Compared with the other four high-quality ran-domized studies evaluating TENS efficacy (Deyo et al.,1990; Cheing and Hui-Chan, 1999, Topuz et al., 2004;Jarzem et al., 2005), our study is the first in which theresults are reported using an ITT analysis and in which

long-term follow-up outcomes are reported. Com-pared with the four other studies evaluating TENSefficacy, our study population was larger and pre-sented more severe LBP: only one of the previousstudies included more than 200 patients (Jarzem et al.,2005). In that study, 18% of the patients included hadundergone prior back surgery and the mean RDQscore was 10.5. In our study, in contrast, 46% of the

Figure 2 Improvement in Roland–Morris Disability Questionnaire score at 6 weeks, subgroup analysis. CI, confidence interval; QTF, Québec Task Force;

TENS, transcutaneous electrical nerve stimulation.

Figure 3 Lumbar pain relief: visual analogue scale improvement between first and last assessments, subgroup analysis. CI, confidence interval; QTF,

Québec Task Force; TENS, transcutaneous electrical nerve stimulation.



patients had previously undergone back surgery andthe median RDQ score was 15. In addition, our TENStreatment strategy, with sessions of 240 min/day for 3months, appears to be more aggressive than thoseemployed in the other studies (ranging from five20-min sessions for 2 weeks to 3 h/day for 4 weeks).Despite the inclusion of more patients with severe LBPin our study, and a more intensive schedule of TENStreatment, the 26% premature discontinuation rate at3 months was similar to the 30% rate reported in alarge study in which the duration of TENS treatmentwas 4 weeks (Jarzem et al., 2005). The high-compliance rates seen in both treatment groups in ourstudy also confirm that self-administered TENS is atreatment option that patients readily submit to.

TENS may conceivably modify not only the nocicep-tive component of lumbar pain, but also the neuro-pathic component. As the treatment protocol com-prised a combined program of conventional (gatecontrol) and acupuncture-like TENS, the improve-ment of lumbar pain could not be attributed toacupuncture-like TENS alone. As previously reported(Freynhagen and Baron, 2009), part of this lumbarpain could therefore be related to a neuropathic com-ponent. Interestingly, a third of the 97 patients in ourpure LBP group had a baseline DN4 score > 4 in thelumbar pain description. Finally, participants reporteda high level of overall satisfaction with the pain man-agement strategy; even those treated with sham TENS,demonstrating that a valid placebo control is possiblein a LBP trial (Machado et al., 2008). However, as themain effect of pain intensity improvement was notedin the subgroup of patients with LBP and radicularpain, it might be worth performing a further random-ized study focusing on that population.

5. Conclusion

The overall results of this study do not support theuse of TENS in the treatment of patients with chronicLBP.

Acknowledgements

We thank the Direction Générale de la Santé (Ministère de laSanté et des Sports), the Fondation CNP Assurances and theInstitut UPSA Douleurs for their financial support. We alsothank CEFAR France, which provided the sham TENSdevices free of charge, and Noélie Buisson-Descombes forreviewing the English manuscript.

References

Andersson GBJ. Epidemiologic features of chronic low-backpain. Lancet 1999;354:581–5.

Baron R, Binder A. How neuropathic is sciatica? The mixedpain concept. Orthopade 2004;33:568–75.

Becker A, Held H, Redaelli M, Chenot JF, Leonhardt C,Keller S, et al. Implementation of a guideline for low backpain management in primary care – a cost-effectivenessanalysis. Spine (Phila Pa 1976) 2011. [Epub ahead ofprint].

Bouhassira D, Attal N, Fermanian J, Alchaar H, Gautron M,Masquelier E, et al. Development and validation of theNeuropathic Pain Symptom Inventory. Pain 2004;108:248–57.

Buchmuller-Cordier A, Navez N, Presles E, Venet E, DupontO, Laurent B. Première enquête nationale sur les tech-niques non médicamenteuses utilisées en Centresd’Evaluation et de Traitement de la Douleur (CETD) pourla prise en charge de la douleur chronique de l’adulte.Douleurs 2008;9:315–19.

Figure 4 Radicular pain relief: visual analogue scale improvement between first and last assessments, subgroup analysis. CI, confidence interval; QTF,

Québec Task Force; TENS, transcutaneous electrical nerve stimulation.



Carey TS, Freburger JK, Homes GM, Castel L, Darter J, AgansR, et al. A long way to go: practice patterns and evidencein chronic low back pain care. Spine 2009;34:718–24.

Carroll D, Moore RA, McQuay HJ, Fairman F, Tramer M,Leijon G. Transcutaneous electrical nerve stimulation(TENS) for chronic pain. Cochrane Database Syst Rev2000;(4):CD003222.

Cheing GL, Hui-Chan CW. Transcutaneous electrical nervestimulation: nonparallel antinociceptive effects on chronicclinical pain and acute experimental pain. Arch Phys MedRehabil 1999;80:305–12.

Deveraux MW. Low back pain. Prim Care Clin Office Pract2004;31:33–51.

Deyo RA, Walsh NE, Martin DC, Schoenfeld LS, Ramamur-thy S. A controlled trial of transcutaneous electrical stimu-lation (TENS) and exercise for chronic low-back pain. NEngl J Med 1990;322:1627–34.

Dogan SK, Tur BS, Kurtais Y, Atay MB. Comparison of threedifferent approaches in the treatment of chronic low backpain. Clin Rheumatol 2008;27:873–81.

Flor H, Fydirch T, Turk DC. Efficacy of multidisciplinary paintreatment centers: a meta-analytic review. Pain1992;49:221–30.

Freynhagen R, Baron R. The evaluation of neuropathic com-ponents in low back pain. Curr Pain Headache Rep2009;13:185–90.

Hildebrandt J, Ursin H, Mannion AF, Airaksinen O, Brox JI,Cedraschi C, et al. European guidelines for the management ofchronic non specific low back pain. 2004. http://www.backpaineurope.org/web/files/WG2_Guidelines.pdf,accessed 11 December 2011.

Jarzem P, Harvey EJ, Arcaro N, Kazarowski J. Transcutane-ous electrical nerve stimulation (TENS) for chronic lowback pain. J Musculoskeletal Pain 2005;13:3–8.

Khadilkar A, Odebyi DO, Brosseau L, Wells GA. Transcuta-neous electrical nerve stimulation (TENS) versus placebofor chronic low-back pain. Cochrane Database Syst Rev2008;(4):CD003008.

Machado LAC, Kamper SJ, Herbert RD, Maher CG, McAuleyJH. Imperfect placebos are common in low back pain trials:a systematic review of the literature. Eur Spine J2008;17:889–904.

Maher CG. Effective physical treatment for chronic low backpain. Orthop Clin North Am 2004;35:57–64.

Moore RA, Moore OA, Derry S, McQuay HJ. Numbersneeded to treat calculated from responder rates give abetter indication of efficacy in osteoarthritis trials thanmean pain scores. Arthritis Res Ther 2008;10:R39.

Ostelo RW, de Vet HC. Clinically important outcomes in lowback pain. Best Pract Res Clin Rheumatol 2005;19:593–607.

Ostelo RW, Deyo RA, Stratford G, Waddell G, Croft P, VonKorff M, et al. Interpreting change scores for pain andfunctional status in low back pain: towards internationalconsensus regarding minimal important change. Spine2008;33:90–4.

Spitzer WO. Scientific approach to the assessment and man-agement of activity-related spinal disorders. A monographfor clinicians. Report of the Quebec Task Force on SpinalDisorders. Spine 1987;12:S1–59.

Topuz O, Ozfidan E, Ozgen M, Ardic F. Efficacy of transcu-taneous electrical nerve stipulation and percutaneous elec-trical neuromodulation therapy in chronic low back pain.J Back Musculoskelet Rehabil 2004;17:127–33.

Van Tulder MW, Koes BW, Assendelft WJJ, Bouter LM. Theeffectiveness of conservative treatment of acute andchronic low back pain. Amsterdam: EMGO Institute; 1999.

Von Korff M, Saunders K. The course of back pain in primarycare. Spine 1996;21:2883–7.



value of tens for relief of chronic low back pain with or without radicular pain

Documents