A Retrospective, Long-term, Third-Party

Follow-up of Patients Considered for

Spinal Cord Stimulation

Malvern S. May, FRCA, & Carol Banks, MSc, & Simon J. Thomson, FRCA

Basildon and Thurrock General Hospitals NHS Trust, Pain Management Services, OrsettHospital, Essex, United Kingdom

& ABSTRACTThe objective of this study was to follow up patientsconsidered for spinal cord stimulation and assessoutcomes and patient selection factors associatedwith outcome. A retrospective study of patientsconsidered for spinal cord stimulation was performed.This included three groups: A) those who did not havea temporary trial of stimulation, B) those who did notgo on to have long-term stimulation after a trial, andC) those who did go on to have long-term stimulationafter a trial. Patient notes were obtained, a structuredtelephone interview conducted, and patients returneda questionnaire. VAS scores, percentage pain reduc-tion reported, quality of life reports, Oswestry DisabilityIndex and Hospital Anxiety and Depression Indiceswere recorded. Fifty-two percent of patients who hadlong-term stimulation reported 50% or greater painreduction. In the subset of these patients with thediagnosis of failed back surgical syndrome, 51%reported 50% or greater pain reduction. Significantimprovements in VAS, Oswestry Disability Index, and

depression were reported pre and post long-termstimulation. There were significant differences in thequality-of-life reports between the patients who re-ceived long-term stimulation (who showed a positiveoutcome) and those who did not (who showed anegative outcome); 80% of patients receiving long-term stimulation reported an improvement in theirquality of life. Follow-up of a cohort of patientsreceiving long-term stimulation demonstrated somereduction of treatment efficacy. Use of a psychologi-cal interview aided patient selection, but no otherfactors showed a correlation with reported painreduction. No serious adverse effects were demon-strated. A revision rate of 23.5% occured in long-termstimulated patients. We conclude that spinal cordstimulation is an efficacious therapy which is notassociated with serious side effects. There is somereduction in therapeutic efficacy over time. Patientsdeteriorate without treatment. &

KEY WORDS: chronic pain, failed back surgicalsyndrome, spinal cord stimulation.

INTRODUCTION

Since first reports of its use in 1967 (1), spinalcord stimulation (SCS) has been used to treat avariety of chronic pain states. In general, the

# 2002 International Neuromodulation Society, 10974-7159/02/$15.00/0Neuromodulation, Volume 5, Number 3, 2002 137–144

This research was supported financially by Basildon and ThurrockGeneral Hospitals NHS Trust, Basildon, Essex, United Kingdom., Dr Maywas also funded in part by Medtronic Inc., Minneapolis, MN, USA.Address correspondence and reprint requests to: Dr. Malvern S. May,Clinical Fellow in Pain Management, Pain Management Services,Orsett Hospital, Rowley Road, Orsett, Essex, RM16 3EU, UK. E-mail:Morphmay@btinternet.com.

indications for SCS can be divided into two broadgroups: ischemic vascular pain (including refract-ory angina and peripheral vascular disease) andneuropathic pain (including failed back surgicalsyndrome, complex regional pain syndrome, andperipheral mononeuropathic pain). The mech-anism of action of spinal cord stimulation isthought to be different in these two groups (2).

Several series reporting the long-term efficacyof SCS in reducing pain and improving function-ing have been published (3–13).

One of the most important features of SCS ispatient selection. The pain states for which SCS iseffective are well recognized, however, selectionin an individual patient is also dependent onpsychosocial factors. We wanted to assess howfactors such as anxiety and depression affectedthe efficacy of SCS. Also we wanted to assess theprogress of patients who did not go on to receiveSCS and to compare these patients to those thatdid. It is not possible, by its very nature, toperform a double-blind trial of SCS. Althoughcross-over studies have been performed tocompare SCS against other treatments, we haveattempted to follow up groups of patients whocould be considered ‘‘control’’ groups.

METHODS

Some patients were sent for a formal psycho-logical opinion prior to any trial of stimulation.All interventions were performed by one physi-cian (SJT). In patients likely to benefit from spinalcord stimulation, the practice was to insert thedefinitive epidural electrode(s) (Pisces Quad,Quad Plus or Quad Compact, Medtronic Inc.,Minneapolis, MN) and deliver a period of trialstimulation via a temporary extension. The trial inthis population of patients lasted from one day toa few weeks. Because of the accepted highsuccess rate of SCS in refractory angina, thesepatients did not receive a formal trial. Electrodeswere placed under local anesthesia to obtainadequate topographical mapping followed bygenerator insertion. After the first 20 patientswere implanted, the remainder of the patientsreceived a trial of at least 6 days. If reduction inpain (defined as 50% or greater pain reduction)and improvement in function were demon-

strated, patients were converted to a permanentlyimplanted system. This involved reopening theoriginal wound and connecting the electrode viaa lead extension to an implanted pulse generator(Itrel 2, Itrel 3 or Synergy, Medtronic Inc.,Minneapolis, MN) or implanted radiofrequencyreceiver (Extrel or Mattrix, Medtronic Inc., Min-neapolis, MN).

From our database we obtained the notes ofpatients who had been considered for spinal cordstimulation. These were divided into threegroups: A) ‘‘Permanent Implant’’: Eighty-threepatients had permanent systems implantedalthough some subsequently had their systemsremoved; these patients were still included in thisgroup, even if, for example, the system had beenremoved for failure to remove pain. B) ‘‘FailedTrial of SCS’’: Eighteen patients had unsuccessfultrials of stimulation where permanent implanta-tion did not follow. C) ‘‘No Trial of SCS’’:Fourteen patients did not have an SCS trialalthough this had been initially consideredbecause of the patient’s pain etiology. In somepatients, unresolved psychological issues meant atemporary lead was not thought appropriate atthe time. Other patients simply wished to put offtheir trial of stimulation and possible permanentimplantation temporarily for reasons such as theyfelt they were too young.

At their initial assessment by the pain man-agement department, patients had been asked torate their pain on a Visual Analog Score (VAS)(0 = no pain and 100 = worse pain imaginable)and had completed Part 1 (anxiety) and Part 2(depression) of the Hospital Anxiety and Depres-sion (HAD) Index and an Oswestry DisabilityIndex.

Patients were phoned by an independent,noninvolved third party and completed a struc-tured telephone interview. Patients were askedidentical questions by the interviewer from aprepared document. If acceptable to the patient,a written questionnaire was sent, which includeda VAS score, HAD Index, and Oswestry DisabilityIndex.

We noted whether patients had been sent for aformal psychological opinion prior to their trial ofstimulation. Patients would be sent for this if itwas felt that there were significant psychosocialfactors which needed to be explored prior to a

138 & MAY ET AL.

trial of stimulation. Three groups emerged fromthese patients:

1. Those who it was felt were not suitable for anSCS trial (these were included in the ‘‘no trialof SCS’’ group).

2. Those who it was felt were suitable for an SCStrial and had no psychological contraindica-tions.

3. Those where it was felt an SCS trial wasindicated but caution was expressed by thepsychologist who had assessed the patient.

In 1997 a similar audit had been conducted on23 patients who had permanent implants. It waspossible to follow up this cohort as a separategroup.

Statistical analysis was by methods stated andsignificance was defined as p < 0.05.

RESULTS

A total of 101 patients had SCS trials orpermanent implants. Two sets of notes wereunavailable and are not included in the study.Fourteen patients did not have SCS trials. Table 1shows a breakdown of patients’ characteristics.Table 2 lists the questionnaire results.

In the 81 patients who had permanent implantsinserted there appeared no correlation betweenduration of pain prior to implantation, age orHAD scores and reported levels of pain relief. Of

these 81 patients, nine patients had their systemssubsequently permanently removed.

Telephone Interview

Patients were asked to report if they had any painreduction, and to try to score it as a percentagereduction (0% = no pain reduction, 100% =complete pain abolition).

Only two of 14 patients from the ‘‘no trial ofSCS’’ group who took part in a telephoneinterview reported any pain reduction. Thesetwo patients were receiving regular epiduralsteroid and local anesthetic injections for FBSS.Overall one patient reported 50% or greater painrelief (7% of all patients) with a mean percentagepain reduction of all patients of 8%.

None of the 10 patients from the ‘‘failed trial ofSCS’’ group who took part in a telephoneinterview reported any pain reduction after theirfailed trial of stimulation with a variety oftherapies, including surgery, injection therapies,medication, pain management programs, orspontaneous improvement.

In the 81 ‘‘permanently implanted’’ patients(including nine who had their implants subse-quently removed) of the 69 patients who took partin a telephone interview, 52% of patients reported50% or greater pain relief, with a mean percentagepain reduction of all the patients of 42%.

Table 1. Breakdown of the Number of Patients Auditedand their Diagnoses

DiagnosisNo trial

SCSFailed trial

SCSPermanent

implant

Failed backsurgical syndrome

8 10 50

Peripheralvascular disease

0 2 9

Complex regionalpain syndrome

2 0 8

Refractory angina 0 0 4Peripheral

neuropathic pain3 4 8

Phantom limb pain 0 2 0Others 1 0 1Unknown 0 0 1Total number

of patients14 18 81

Table 2. Patient Demographicsa

No trial SCS Failed trial SCSPermanent

implant

Mean age(years)

50.1 55.2 55.7

Range, SD 30–73, 13.6 29–74, 14.2 28–81, 13.8Male/female 5/9 10/8 43/38Mean duration

pain (months)63.6 76.4 75.3

Range, SD 24–260, 40.4 12–250, 63.5 3–300, 73.8Mean major

surgical episodes1.4 1.8 1.6

Range, SDb 0–4, 1.1 0–9, 2.2 0–9, 1.7Mean follow-up

(months)44.4 42.5 37.5

Range, SD 10–82, 21.2 21–64, 16.6 6–83, 21.9

a

There were no significant differences (in all cases p > 0.1) between thethree groups (Chi squared tests for gender, Student’s t test for others).b These were defined as operative procedures undertaken to treat thesame pain as the SCS, eg, discectomies and stabilization procedures infailed back surgical syndrome (FBSS) patients.

Retrospective Third-Party Follow-up for Patients Considered for SCS & 139

From the permanently implanted group, it waspossible to separate out the 47 patients whosediagnosis was FBSS. In this group, 51% reported50% or greater pain reduction, with a meanpercentage pain reduction of 43%.

In their telephone interviews, patients wereasked to comment on whether their quality oflife, level of physical activity and quality of sleephad improved, stayed the same, or worsened.These results are presented in Figs. 1–3 below.

Fifty-nine of the 69 patients (87% of total) whohad permanent systems inserted would have animplant inserted again and would recommend itto another patient.

Written Questionnaire

Eleven of the ‘‘no trial of SCS’’, eight of the‘‘failed trial of SCS’’, and 53 of the ‘‘permanentlyimplanted’’ patients returned written question-naires sent out after their telephone interviews.Table 3 lists the results of these questionnaires.

In the permanently implanted group therewere significant reductions in the OswestryDisability Index (0.001 < p < 0.01), the HADPart 2 score (0.001 < p < 0.01) and the VAS score( p < 0.001) (Student’s t-test). There were nosignificant differences between the pre and posttreatment scores in the other two patient groups.

In the permanent implant group, those whohad a formal psychological opinion which indi-

cated no contraindications to a trial of stimu-lation (n = 50) showed little difference in meanreported levels of pain relief vs. patients (n = 23)where the implanting physician had not sought apsychologist’s opinion (45.0% vs. 44.6% meanreported pain reduction).

However, in patients in whom a caution wasattached to their psychological opinion (n = 8), asmaller mean reported pain reduction (31.9%)was demonstrated. This difference did not reachstatistical significance.

Figure 1. Patients who had permanent implants inserted

reported much improved quality of life compared to the other

two groups (p < 0.001) (Chi squared test).

Figure 2. Patients who had permanent implants inserted

reported much improved levels of physical activity; 0.02 < p

< 0.05 for No Trial of SCS vs. Permanent Implant, 0.001 < p < 0.01

for Failed Trial of SCS vs. Permanent Implant (Chi squared test).

Figure 3. Patients with permanent implants inserted reported

much improved quality of sleep compared to the other two

groups (p < 0.001) (Chi squared test).

140 & MAY ET AL.

Follow-up of a Cohort of Patients withPermanent Implants

In 1997, a similar audit had been conducted andit was possible to compare reported painreduction.

Seventeen of the 23 patients (74% of allpatients) had implants for FBSS, a larger propor-tion than other permanently implanted patients(in which 62% were FBSS patients).

Mean follow-up in 2000 was 57.6 months(range 41–82, SD 11.8); this is significantly longer(p < 0.001 (Student’s t-test)) than other perma-nently implanted patients. There were no sig-nificant differences in age, number of previoussurgical procedures, or duration of pain prior toimplant between this group and the otherpermanently implanted patients.

At the first audit, a mean of 16.4 months afterimplantation, mean reported pain reduction was64.3%, and 96% of patients reported greater than50% pain reduction.

In 2000, a mean of 57.6 months after implanta-tion, mean reported pain reduction was 52.3%,and 61% of patients reported greater than 50%pain reduction.

If only patients with a diagnosis of FBSS areincluded. a pain reduction of 50% or greater isdemonstrated in all patients in the initial audit,and in 59% of the patients in 2000. Mean pain

reduction was 67.2% in the initial audit, and52.6% in 2000.

Adverse Events

Nine patients had their systems removed. In twopatients with FBSS, after several years of stimu-lation, the patients reported a gradual reductionin pain and need for stimulation. This resulted inthe removal of their lead and generator as theyfelt they longer needed it.

Two patients reported early (less than 6months), and two reported later (greater than6 months) failure to reduce pain in a fullyfunctioning implanted system with adequateparesthesia mapping and stimulation amplitude.In these cases failure to produce pain reliefoccurred despite optimal spinal cord stimulationtherapy.

Two patients had their systems removed due toinfection and one due to lead migration and didnot elect to have a new system inserted. Thesewere early cases in the series.

The most common complication among allpatients was infection during the trial period; thisoccurred in 11 of the first 59 patients, resulting indelays in permanent system insertion whereappropriate, and antibiotic therapy. This com-plication of infection during the trial period wasreduced to three episodes in the next 40 patients,

Table 3. Results of the Written Questionnairea

No Trial SCS (n = 11) Failed trial SCS (n = 8) Permanent implant (n = 53)

Pre Post Pre Post Pre Post

Mean Oswestry Disability Indexb 27.7 25.9 28.0 28.0 28.0 24.3e

Range 9–38 10–35 8–39 9–44 9–40 4–38SD 10.1 9.2 11.7 10.3 5.8 7.5

Mean HAD Part 1 (anxiety)c 13.6 11.0 10.6 11.4 10.1 9.2Range 7–21 8–18 4–18 3–19 3–20 1–19SD 5.7 3.4 4.6 6.4 4.3 4.5

Mean HAD Part 2 (depression)c 11.3 9.4 11.1 12.3 10.0 7.5e

Range 5–21 4–15 4–16 4–19 3–17 1–17SD 4.9 4.0 4.0 4.7 3.6 4.0

Mean Pain VASd 68.3 66.4 60.0 65.0 71.6 51.7e

Range 50–100 50–90 30–100 50–100 30–100 0–100SD 19.4 16.3 22.7 18.9 21.8 22.5

a

The mean scores for the data is tabulated, with the first figure being the score recorded at the initial consultation prior to any treatment (pretreatment) and the second that of the written questionnaire at the time of the audit (post treatment).b Oswestry Disability Index scored from 0 – 50.c HAD scored 0 –21.d VAS scored from 0–100.e Pre and post treatment scores significantly different.

Retrospective Third-Party Follow-up for Patients Considered for SCS & 141

on the introduction of a revised dressing tech-nique of the exit site of the lead extension, whichreduced lead movement at the exit site. All theinfections were superficial, with no episodes ofepidural or intrathecal spread.

Nineteen ‘‘permanent implant’’ patientsneeded revisions, a rate of 23.5%.

Ten patients needed revision due to leadbreakage, including one patient in which thisoccurred twice. Lead breakage was diagnosed byimpedance changes noted on telemetric inter-rogation of the implanted pulse generator andconfirmed by the inability of an external pulsegenerator to produce paresthesia during a revi-sion procedure. In some cases the breakage wasvisible to the naked eye, in others confirmationwas by returning the lead to the manufacturer fordetailed examination. The mean period afterimplantation at which lead breakage firstoccurred was 18.0 months (range 3–54 months).The majority of the lead fractures occured in theearlier implants (seven of the 10 patients were inthe first 25 patients permanently implanted).

Eleven patients needed revision due to leadmigration, including two patients who had tworevisions due to migration. The mean period afterimplantation at which this first occurred was13.4 months (range 3–31 months). Migrationoccurred in three of the first 27, two of the second27, and six of the final 27 patients permanentlyimplanted.

DISCUSSION

In this series we have demonstrated a markedimprovement in many indices of pain andfunction in those patients who received long-termstimulation compared to a worsening in thosewho did not. Placebo-controlled studies are notfeasible with SCS, by its very nature. The follow-upof patients who did not go on to permanentimplantation demonstrates minimal, if any,improvement in outcome measures. Other studieshave compared two patient groups who receivedSCS or an active surgical treatment, whether thisbe for FBSS (6) or angina (8,9). We believe thatthis study is the first to have followed up patientswho did not receive permanent implantation anddemonstrated their worsening outcome.

This series is comparable to other studies inthe reported levels of pain reduction. In thesestudies, the mean number of operations prior toimplantation is greater than our series, usuallyabout three (5,6,10,11). However this is the firstseries published in the United Kingdom, wherethe incidence of spinal surgery is accepted asbeing lower than in North America.

Kumar (4) reported that, at a mean follow upof 66 months, there was a 50% or greater painreduction in 59% of patients. At the relativelyshort follow-up of one year, Burchiel et al. (11),in a prospective study, reported 50% or greaterpain reduction in 55% of patients. In a literaturereview of SCS in FBSS patients by Turner et al.(10), in 29 studies, the mean percentage ofpatients implanted who reported 50% or greaterpain reduction across all studies included was59%, but there is no indication of follow-up time.In Turner’s review, greater than 50% painreduction at 1, 2, and 5 years follow-up waspresent in 62%, 64%, and 53% of patients,respectively. In our follow-up cohort, the resultsfor 16.4 months (1 year) and 57.6 months(5 years) were 100% and 59%, respectively.

These good initial results may be explained byour patients’ relatively low number of previousspinal surgeries. North (5) noted a tendency to abetter outcome in those who had fewer prioroperations.

It is of interest that the difference in VAS scoresis indicative of a 28% reduction in pain scores, incomparison with the direct ‘‘please could youscore your pain reduction. . .’’, where a meanreduction of 42% was reported; a not uncommonfinding in any chronic pain study.

It was noticed after the completion of the auditthat there was some discrepancy between thepatients’ reports of pain reduction in the auditand the patients’ pain reduction when seen bythe authors of this article subsequently in anoutpatient setting. At the structured telephoneinterview by an uninvolved third party, somepatients who reported poor or no pain reductionhad in fact been getting excellent long-term painreduction. However at the time of the interviewthese patients had temporary technical problemsin the process of being dealt with (eg, implantedpulse generator battery exhaustion or recent leadmovement needing reprogramming). The tele-

142 & MAY ET AL.

phone interview was not vague in its questioningbut we have made no allowances for this in theresults, and thus there is underreporting of painreduction.

In other series improvement in other func-tional outcomes, eg, Oswestry Disability Index,perceived quality of life, as well as pain reduction,were recorded (3–7,11). These have been shownto correlate with reported pain reduction (10),and the presence of significant differencesbetween the outcomes of the three groups inthis study is in agreement with this.

We, like others (11), have demonstrated areduction in depression in patients treated withSCS, but contrary to other studies did not find acorrelation between depression scores at initialassessment and subsequent levels of pain reduc-tion (11). We did not find any correlation withany factors other than the opinion of a psycho-logist and the level of pain reduction the patientreported.

We did not use any scales or indices topsychologically assess our patients. In patientswhom it was felt were ‘‘psychologically uncom-plicated’’ by the treating physician, a formalpsychologist’s opinion had not been sought atthe physician’s discretion. These patients’ resultsdiffered little from those who were felt suitablefor a trial of stimulation by a psychologist, butboth groups had better pain reduction thanpatients where caution had been attached.

The use of formal tests does not aid patientselection (14), and we are in agreement with thosewho rely on broad categories in this respect (7).

In the early part of this series, technicalproblems were greater, particularly with respectto lead fracture. This particular complication isexplained by the institution of an improvedepidural lead fixator device, in approximately1995. Three early patients in the series electedto have their systems removed as a consequence oftechnical complications, and it is possible that ifimplanted in 2001, would not have taken this step.

It was our practice not to perform non-incisionpercutaneous trials. Trial leads were insertedunder local anesthesia via a small incision andfixed to appropriate structures. The use of amodified dressing technique and increasedpatient education with respect to the dressinglead to a reduction in infection rates during our

trials. Prophylactic antibiosis was used at all leadand generator insertions or revisions.

Overall, the infection rate compared favorablywith other series (4,5,10), considering theinherent increased infective risk with our tech-nique of using the trial electrode for permanentimplantation.

CONCLUSIONS

Our study shows that spinal cord stimulation isan efficacious therapy, improving analgesia andfunction in patients with chronic pain due toFBSS and vascular disorders such as angina andperipheral vascular disease. Our study also showsa significant drop off in pain reduction withlonger term follow-up. The patient group withouttreatment worsened in lifestyle and pain indicesover time. The incidence of serious side effectswith spinal cord stimulation is low, but minorside effects are common. Selection appears aidedby a formal psychological opinion.

REFERENCES

1. Shealy CN, Mortimer JT, Reswick JB. Electricalinhibition of pain by stimulation of the dorsal column:Preliminary clinical reports. Anesth Analg 1967;46:489–491.

2. Linderoth B, Foreman RD. Physiology of spinalcord stimulation. Review and Update. Neuromodula-tion 1999;2:150–164.

3. Bennett D, Alo KM, Oakley J, Feler C. Spinalcord stimulation for complex regional pain syndrome(RSD): a retrospective multicentre experience from1995 to 1998 in 101 patients. Neuromodulation 1999;2:202–210.

4. Kumar K, Toth C, Nath RK, Laing P. Epiduralspinal cord stimulation for treatment of chronicpain—some predictors of success. A fifteen yearexperience. Surg Neurol 1998;50:110–121.

5. North RB, Kidd DH, Zahurak M, James CS, LongDM. Spinal cord stimulation for chronic intractablepain: Experience over two decades. Neurosurgery1993;32:384–395.

6. North RB, Kidd DH, Lee MS, Piantodosi S. Aprospective, randomised study of spinal cord stimula-tion versus reoperation for failed back surgery syn-drome. Stereotact Funct Neurosurg 1994;62:267–272.

7. Kupers Rc, van den Oever R, Van Houdenhove B

Retrospective Third-Party Follow-up for Patients Considered for SCS & 143

et al. Spinal cord stimulation in Belgium. a nation-widesurvey on the incidence, indications and therapeuticefficacy by the health insurer. Pain 1994;56:211–216.

8. Hautvast RW, DeJongste MJ, Staal MJ, van GilstWH, Lie K. Spinal cord stimulation in chronicintractable angina pectoris: a randomized controlledefficacy study. Am Heart J 1998;136:1114–1120.

9. Mannheimer C, Eliasson T, Augustinsson LEet al. Electrical stimulation versus coronary bypasssurgery in severe angina pectoris. The ESBY Study.Circulation 1998;97:1157–1163.

10. Turner JA, Loeser JD, Bell KG. Spinal cordstimulation for chronic low back pain: a systematicliterature synthesis. Neurosurgery 1999;37:1088–1095.

11. Burchiel KJ, Anderson VC, Brown FD et al.Prospective, multicenter study of spinal cord stimula-tion for relief of chronic back and extremity pain.Spine 1996;21:2786–2794.

12. De La Porte C, Van de Kelft E. Spinal cordstimulation in failed back surgery syndrome. Pain1993;52:55–61.

13. LeDoux MS, Langford KH. Spinal cord stimula-tion for the failed back syndrome. Spine 1993;18:191–194.

14. North RB, Kidd DH, Wimberly RL, Edwin D.Prognostic value of psychological testing in patientsundergoing spinal cord stimulation. Neurosurgery1997;39:301–310.

144 & MAY ET AL.