Content uploaded by Michael I Bennett
Author content
All content in this area was uploaded by Michael I Bennett on Oct 14, 2017
Content may be subject to copyright.
Feasibility Study of Transcutaneous Electrical Nerve Stimulation
(TENS) for Cancer Bone Pain
Michael I. Bennett,*Mark I. Johnson,
y
Sarah R. Brown,
z
Helen Radford,
{
Julia M. Brown,
#
and Robert D. Searle
x
*Professor of Palliative Medicine, Lancaster University, Lancaster, United Kingdom.
y
Professor of Pain and Analgesia, Leeds Metropolitan University, Leeds, United Kingdom.
z
Senior Medical Statistician, Clinical Trials Research Unit, University of Leeds, Leeds, United Kingdom.
{
Senior Research Nurse, Clinical Trials Research Unit, University of Leeds, Leeds, United Kingdom.
#
Professor and Director, Clinical Trials Research Unit, University of Leeds, Leeds, United Kingdom.
x
Specialist Registrar & Honorary Lecturer in Anaesthesia St James’s University Hospital, Leeds, Leeds, United Kingdom.
Abstract: This multicenter study assessed the feasibility of conducting a phase III trial of transcuta-
neous electrical nerve stimulation (TENS) in patients with cancer bone pain recruited from palliative
care services. Eligible patients received active and placebo TENS for 1 hour at site of pain in a random-
ized crossover design; median interval between applications 3 days. Responses assessed at 30 and 60
minutes included numerical and verbal ratings of pain at rest and on movement, and pain relief.
Recruitment, tolerability, adverse events, and effectiveness of blinding were also evaluated.
Twenty-four patients were randomised and 19 completed both applications. The intervention was
well tolerated. Five patients withdrew: 3 due to deteriorating performance status, and 2 due to in-
creased pain (1 each following active and placebo TENS). Confidence interval estimation around
the differences in outcomes between active and placebo TENS suggests that TENS has the potential
to decrease pain on movement more than pain on rest. Nine patients did not consider that a placebo
was used; the remaining 10 correctly identified placebo TENS. Feasibility studies are important in pal-
liative care prior to undertaking clinical trials. Our findings suggest that further work is required on
recruitment strategies and refining the control arm before evaluating TENS in cancer bone pain.
Perspective: Cancer bone pain is common and severe, and partly mediated by hyperexcitability.
Animal studies suggest that Transcutaneous Electrical Nerve Stimulation can reduce hyperalgesia.
This study examined the feasibility of evaluating TENS in patients with cancer bone pain in order
to optimize methods before a phase III trial.
ª2010 by the American Pain Society
Key words: Transcutaneous electric nerve stimulation (TENS), pain, cancer of bone, randomized
controlled trial, palliative care.
Bone metastases occur in around 40% of patients
with lung, renal and thyroid cancers and in around
70% of patients with breast and prostate cancer.
12
Approximately 50% of patients with bone metastases
suffer pain as a consequence, resulting in cancer bone
pain being the most common cause of cancer-related
pain
38
and one of the most painful cancer pain syn-
dromes that significantly reduces quality of life.
40
Less
than half of all patients with cancer bone pain experi-
ence significant relief of their pain with current
treatments such as radiotherapy, opioids, and bisphosph-
onates. Although these treatments can result in substan-
tial benefit for some patients, cancer bone pain remains
a significant therapeutic challenge with only 41% of pa-
tients achieving at least 50% pain relief after a course of
radiotherapy,
36
and only 42% of patients using oral
transmucosal fentanyl citrate experienced a 33% or
greater reduction in pain score.
13
A systematic review
of trials of bisphosphonates demonstrated a number
needed to treat (NNT) of 11 for analgesic benefit at 4
weeks, and NNT of 7 for analgesic benefit at 12 weeks,
45
and some patients do not seem to benefit at all.
43
Delivering electricity across the skin, termed transcuta-
neous electrical nerve stimulation (TENS), is an age-old
technique for managing pain.
24
TENS is used throughout
Received April 28, 2009; Revised July 8, 2009; Accepted August 4, 2009.
Supported by Cancer Research UK (C18324 / A7715) with additional
support from an unrestricted grant awarded by Cephalon.
Address reprint requests to: Michael I. Bennett, Professor of Palliative
Medicine, International Observatory on End of Life Care, School of
Health and Medicine, Bowland Tower East, Lancaster University, Lancas-
ter, United Kingdom. E-mail: m.i.bennett@lancaster.ac.uk
1526-5900/$36.00
ª2010 by the American Pain Society
doi:10.1016/j.jpain.2009.08.002
351
The Journal of Pain, Vol 11, No 4 (April), 2010: pp 351-359
Available online at www.sciencedirect.com
the world to manage painful conditions because it is in-
expensive, noninvasive, capable of self-administration,
and with no potential for toxicity or overdose. TENS is of-
ten used in combination with medication. The purpose
of TENS is to selectively activate low-threshold afferents
(A-beta) at the site of pain, and this is achieved in prac-
tice by delivering TENS at nonpainful intensities to pro-
duce strong but comfortable electrical paraesthesiae.
Electrophysiological studies have shown that TENS in-
hibits spontaneous and noxiously evoked spinal-cord
cell activity for short periods of time and with a rapid
onset.
21,31,34
When TENS is administered at intensities ac-
tivating high-threshold afferents, longer lasting depres-
sion of central nociceptor cells of at least 2 hours has
been observed.
33
Studies on healthy participants
exposed to experimentally induced pain have shown
that strong but nonpainful TENS produces rapid
onset, short-lasting reductions in pain-sensitivity re-
sponse.
1,10,11,27
The rapid onset of TENS suggests that it
may be particularly useful for breakthrough and incident
pain.
Evidence suggests that cancer bone pain is a unique
pain state mediated by mechanisms that include hyper-
excitability.
46
Animal studies have shown that TENS
reduces hyperalgesia through activation of receptor-
mediated pathways at the level of the spinal cord and
supraspinally.
35
Evidence from systematic reviews and meta-analyses is
conflicting for acute pain
3,7,44
but more positive for
chronic pain,
26
although many are inconclusive.
4,29,39
In-
adequate TENS technique and weak methodological
quality has contributed to negative findings in RCTs.
3
In
the context of cancer pain, evidence supporting TENS
therapy is lacking.
41
However, we have shown improve-
ments in cancer bone pain using TENS in an uncontrolled
pilot study,
42
but to further this research we examined
the feasibility of evaluating TENS within a future phase
III clinical trial.
Methods
The study was designed as a randomized controlled
crossover feasibility study (ISRCTN = 92118149). Ethical
approval was obtained from Leeds (West) Research Ethics
Committee (07/Q1205/5).
The trial was designed to assess the feasibility of run-
ning a large-scale phase III trial to investigate the efficacy
of active TENS in the control of cancer bone pain. Such
studies are recommended in the evaluation of complex
interventions when little information is available on
the application of the intervention, the study size, and
recruitment and retention, and are an essential step in
the development and testing of an intervention, prior
to a large-scale evaluation.
14
As discussed by Anderson
and Prentice,
2
feasibility studies play an important role
in health research, providing guidance on issues such as
choice of interventions, study populations, and sample
sizes for the development of a phase III trial.
We chose a crossover design because of the subjective
nature of pain assessment and because this design allows
patients to act as their own control and experience both
interventions. Crossover designs also ensure balanced
covariates between the intervention groups, and by ran-
domizing the order in which patients receive the treat-
ments, bias is minimized between the arms. The
endpoints for consideration in this study were pain
intensity and pain relief, on movement and at rest.
Eligibility
Patients aged over 18 referred to specialist palliative
care services in 2 UK cities (initially in Leeds and then
in Lancaster) with painful bone metastases of any pri-
mary cancer formed the study population. Patients
were required to have radiological evidence of bone me-
tastases, pain rated at least 3 out of 10 on a numerical
pain-intensity scale at rest or on movement at the first
visit, and an estimated survival of longer than 4 weeks.
Patients were excluded if they were unable to complete
patient-related information on entry, did not have ongo-
ing cancer, or were not suitable for TENS as described by
the UK Chartered Society of Physiotherapists Guidance
for the Clinical Use of Electrophysical Agents.
9
These in-
cluded pregnant patients, patients with pacemakers, ep-
ilepsy, and abnormal sensation over the site of pain (such
as allodynia).
Analgesic Medication
Patients on regular analgesic medication were eligible
for the study but were excluded if they had significant
changes to their medication within 48 hours prior to
baseline assessment, defined as an increase or decrease
in opioid dose of 30%, or the addition or removal of coa-
nalgesic medication such as an antidepressant or anti-
convulsant. We undertook TENS applications at similar
times of the day to ensure that time from last dose of
analgesia did not vary between the 2 visits.
TENS Application Protocol
Following informed consent, patients were random-
ized via the Clinical Trials Research Unit (University of
Leeds) central randomization system, using stratified
permuted block randomization to ensure treatment
groups were well balanced by age and gender. Baseline
data was collected before patients received either active
TENS at first application and placebo TENS at second ap-
plication, or vice versa. TENS was applied to the site of
bone pain by a medical researcher using a single-channel
TENS device (Ultima XL-A1; TensCare LTD, Epsom, UK)
and 2 self-adhering hypoallergenic gel pads. The TENS
pads were approximately 5 5 cm in size and placed
between 5 and 10 cm apart on an area of skin in good
condition and without signs of altered sensation. TENS
parameters were chosen in line with recommendations
from opinion leaders and the International Association
for the Study of Pain for conventional TENS. These pa-
rameters were continuous pulse pattern, pulse width of
200 microseconds, and a pulse frequency of 80 Hz.
8,25
Intensity was increased until the TENS sensation was
strong but comfortable.
Placebo TENS was delivered using devices that were
identical in appearance and digital display, but had no
352 TENS for Cancer Bone Pain
current output and were set to a fixed output based on
the display panel.
Pain-intensity measurements were taken at baseline
prior to TENS application, then repeated after 30 min-
utes and 60 minutes of TENS while the devices were
switched on. Pain-relief scales were completed at the
30- and 60-minute time points. TENS was switched off
and removed immediately after the 60-minute measures.
Patients returned for the second TENS application be-
tween 2 to 7 days later and followed an identical exper-
imental procedure. Any adverse reactions were
documented at the end of each TENS application using
research-nurse-assessed toxicity grades.
6
Patients were
also contacted by telephone within 48 hours of complet-
ing each TENS application to determine any nonimmedi-
ate side effects of the treatment.
At each recruitment site, a consistent pair of medical
researcher and research nurse-observer conducted the
research. The medical researcher applying the TENS de-
vice was not blind to the intervention but did not take
part in patient assessments. Patients and the research
nurse-observer were blind to the order of TENS applica-
tion. Patients were informed via a patient information
sheet that ‘‘some types of TENS produce a tingling sensa-
tion underneath the pads while others do not. It is possi-
ble that you could be allocated a placebo or dummy
TENS. It is possible that you may not feel anything from
some types of TENS, and neither application may provide
any relief of your pain.’’ Patients were instructed not to
reveal to the research nurse-observer any sensations
they were feeling at any time during the study period.
Outcome Measures
We chose numerical and verbal rating scales to assess
pain and pain relief because they have been extensively
validated as sensitive and reliable scales, and are recom-
mended as core outcome measures by international
guidelines.
17,23
Specifically, the use of both verbal and
numerical ratings scales is usually preferred by patients
over visual analogue scales, and the combination of mea-
sures reduces missing data particularly in older and frail
patient groups.
22
Baseline pain was measured at rest and on a specified
painful movement using a Numerical Rating Scale (NRS,
a standard 0 to 10 numbered scale anchored 0 = least
pain imaginable and 10 = worst pain imaginable) and
Verbal Rating Scale (VRS, with categories of no pain,
mild, moderate, or severe pain). The specified painful
movement was selected by the patient as one that was
encountered during their daily activities and included
standing up, walking, bending to tie a shoe lace, and
putting on a shirt. We measured pain quality before
and after TENS application using the Short-Form McGill
Pain Questionnaire (SF-MPQ).
37
This questionnaire is reli-
able and well validated, and its sensory and affective sub-
scales have demonstrated responsiveness in chronic-pain
clinical trials.
15
Because it assesses both specific sensory-
pain qualities and the affective component of pain, the
SF-MPQ is recommended for inclusion in clinical trials
as a secondary outcome measure to evaluate the effects
of pain treatment on both sensory and affective qualities
of pain.
16
The same pain-intensity scales were repeated at 30 and
60 minutes during TENS application. Primary endpoints
were pain intensity at rest and pain intensity on move-
ment 60 minutes after starting TENS. Pain relief at rest
or on movement at 30 and 60 minutes during TENS appli-
cation was measured using NRS (anchored 0% = no pain
relief and 100% = complete pain relief), and 4-point VRS
(no pain relief, some relief, good relief, excellent relief).
Pain-relief scales have demonstrated excellent validity
when compared with both pain-intensity scales and the
need for rescue analgesia in clinical trials of cancer
breakthrough pain.
18
At completion of the second TENS application, pa-
tients were asked which of the 2 applications provided
most benefit and which outcome scale best represented
their experience of pain intensity and pain relief. At the
end of the second application, they were also asked
whether they considered either treatment to be placebo.
Patients were asked (by means of a satisfaction question-
naire designed by the authors) if they found TENS ther-
apy beneficial, whether it was easy to use, and whether
TENS had most impact on pain at rest or on movement.
Statistical Methods
This feasibility study was designed to establish the
most appropriate primary outcome measure to use for
the phase III trial, and to determine initial data for this
measure following published guidance and recommen-
dations.
2,14,30
The choice of either pain on rest or pain on movement
as the primary outcome measure for a phase III trial was
made by considering the outcome which showed the
largest benefit for TENS (by assessing the difference in
change in pain intensity scores at 1 hour and pain relief
scores at 1 hour, and corresponding 95% confidence in-
tervals [CIs]). Verbal rating scores and the SF-MPQ were
summarized in terms of the number of patients with
each response, as well as an overall summary score for
the SF-MPQ. The choice of outcome measure (NRS, VRS,
SF-MPQ) was then decided by considering both patient
preference and largest potential benefit for TENS. These
choices were made by informal assessment of descriptive
summaries of the data, and not by formal statistical com-
parisons due to the lack of power to perform these. For-
mal statistical analysis is not necessary for this type of
study;
30
therefore, descriptive summaries were calcu-
lated only for each of the outcome measures. The differ-
ence in change in pain-intensity scores at 1 hour (1 hour
minus baseline) between active and placebo TENS was
calculated with corresponding 95% CIs. Similarly, the dif-
ferences in pain relief at 1 hour and SF-MPQ total score,
between active and placebo TENS, were also calculated
with corresponding 95% CIs.
We recruited a convenience sample of patients to al-
low estimation of parameters for a formal phase III trial,
and not to determine treatment effects, as is appropriate
for feasibility studies.
30
Recruitment rate was assessed in
terms of the number of patients recruited per month,
Bennett et al 353
and in total, within the 12-month study period. The
number of patients completing both visits was also
summarized. This allowed assessment of the potential
recruitment rates for a large-scale phase III study.
Results
Recruitment
In total, 24 patients were recruited to the study during
the 1-year study period (16 in Leeds, 8 in Lancaster; see
CONSORT flow chart, Fig 1). We encountered slower-
than-anticipated recruitment at the first site and so
began recruitment at a second site. Of the 24 patients
recruited, 19 received both applications. Results are
presented for those patients who received both
interventions.
Patient Characteristics
Patient characteristics are listed in Table 1. Cancers that
commonly metastasized to bone were well represented
and sites of painful metastases were largely pelvis, lum-
bosacral spine, and lower limbs. Of previous or current
treatments, 79% had received radiotherapy (median
193 days prior to randomization) and 33% had received
bisphosphonates.
Analgesic Medication
Of the 24 recruited patients, 21 (87%) were currently
treated with strong opioids (morphine, fentanyl, or
oxycodone), 15 with regular paracetamol, and 7 with
regular nonsteroidal anti-inflammatory drugs (NSAIDs).
Sixteen patients were treated with both a strong opioid
and either paracetamol or NSAID. No additional medica-
tion was taken by patients during the TENS applications.
Electrode Placement
In all patients, electrodes were placed at the site of
pain and this site was the same for both applications.
Treatment Outcome
Pain Intensity at Rest and on Movement at 1
Hour
Table 2 displays baseline and 1-hour mean NRS pain
scores, change in scores at 1 hour (1 hour score minus
baseline score), and differences in change in scores be-
tween active and placebo TENS (active–placebo) with
corresponding 95% CI.
The mean change in NRS pain intensity score at rest at
1 hour for active TENS was –.84, and for placebo TENS
was –2.16 (within-patient mean difference = 1.32, 95%
CI = –.36 to 2.99). Similarly the mean change in NRS
pain-intensity score on movement at 1 hour was –2.32
for active TENS and –2 for placebo TENS (within-patient
mean difference = –.32, 95% CI = –1.85 to 1.22). Here,
a larger negative within-patient difference indicates bet-
ter pain improvement at 1 hour with active TENS. For
pain at rest, the confidence intervals around the mean
change in pain-intensity scores indicate that compared
with placebo TENS, active TENS may provide an
Assessed for eligibility (N = 62)
Excluded (n = 38)
Not meeting inclusion criteria (n = 28)
Refused to participate (n = 7)
Other reasons (n = 3)
Patients randomly assigned (n =24)
Discontinued intervention after first period
(n = 2)
Allocated to Active then Placebo (n =12)
Received allocated intervention (n = 12)
Primary analysis (n =10)
Excluded from primary analysis (n =2)
Discontinued intervention after first period
(n = 3)
Primary analysis (n =9)
Excluded from primary analysis (n =3)
Allocated to Placebo then Active (n =12)
Received allocated intervention (n = 12)
Figure 1. CONSORT flow chart of transcutaneous electrical nerve stimulation (TENS) feasibility trial.
354 TENS for Cancer Bone Pain
improvement in pain of up to .36 points, or may provide
deterioration in pain by up to 2.99 points. Similarly, for
pain on movement, active TENS may provide an improve-
ment in pain of up to 1.85 points or may provide deteri-
oration in pain by up to 1.22 points, compared with
placebo TENS.
We considered the within-patient difference in
change in 1-hour NRS pain-intensity scores according to
the sequence in which patients received the interven-
tion. For pain intensity at rest, the direction of the differ-
ences was consistent with that when sequence was
ignored. However, for pain intensity on movement, the
direction of the difference changed according to
sequence but the differences were minor (active then
placebo = –.7; placebo then active = .11).
At 1 hour,7 patients had no pain on rest with active
TENS, and 8 patients had no pain on rest with placebo
TENS, when assessed with the verbal rating scale. When
considering pain on movement, 10 patients had mild
pain and 7 had moderate pain with active TENS at 1
hour, compared to 7 patients with mild pain, 7 with mod-
erate pain and 1 with severe pain on placebo TENS. The
Short-Form McGill Pain Questionnaire (SF-MPQ) was
also used to assess patients’ pain. The mean total SF-MPQ
score at 1 hour on active TENS was 3.26 and on placebo
TENS was 4.58. The difference in mean scores between
the 2 interventions was -1.32 (95% CI = [–3.42, .79]),
where a negative difference implies less pain with active
TENS.
Pain Relief at Rest and on Movement at 1 Hour
Mean NRS pain-relief scores are displayed in Table 2.
When considering differences in pain relief, a positive
difference implies more pain relief with active TENS.
The mean pain relief on movement score for active
TENS was 52.6, compared to 38.4 with placebo TENS (dif-
ference = 14.2, 95% CI [–3.34, 31.76]). The confidence in-
tervals around the differences in pain relief on
movement at 1 hour reflect relative improvement in
pain relief of up to 83% and relative deterioration in
pain relief of up to 9%, with active TENS. Corresponding
relative improvement and deterioration in pain relief at
rest are 59% and 69% respectively, with active TENS. Ver-
bal pain-relief scores on movement indicated more pa-
tients experiencing good or very good pain relief with
active TENS (12/19, 63.2%) compared to placebo TENS
(5/19, 26.3%) (Table 3). The difference in the proportion
of patients experiencing good or very good pain relief on
movement with active TENS compared to placebo TENS is
36.8% (95% confidence interval [7.5%, 66.2%]), indicat-
ing that active TENS does have the potential to provide
more pain relief compared to placebo TENS.
Choice of Outcome and Sample Size
Calculation
As can be seen from the pain-intensity scores displayed
in Table 2, the confidence intervals around the differences
in change in pain-intensity score at 1 hour indicate that
TENS may have the ability to reduce pain on movement
more than pain at rest. This is confirmed by the difference
in pain relief on movement being greater than the differ-
ences in pain relief at rest. Pain on movement was there-
fore chosen as the most appropriate outcome for use in
a future phase III trial of TENS. Results of patient prefer-
ences indicated little difference in preference for numeri-
cal-ratingscores over verbal-rating scores. However, only 1
patient indicated a preference for the SF-MPQ for summa-
rizing pain relief, and only 3 patients indicated a prefer-
ence for this scale for summarizing pain intensity. The
summary tool to use based on these results was therefore
chosen to be the NRS scale, for pain on movement.
To perform a sample-size calculation, we derived the
within-patient standard deviation for change in pain in-
tensity on movement at 1 hour (within-patient standard
deviation = 3.18) and assumed that a 30% bigger change
in pain between 1 hour and baseline with active TENS
compared to placebo TENS would be clinically significant
within a crossover design.
19
One hundred and twelve pa-
tients per sequence would be required to detect a 30%
bigger change in pain intensity assuming a mean change
in NRS pain score at 1 hour of –2 on placebo TENS, with
Table 1. Patient Characteristics
ACTIVE TENS
THEN PLACEBO
TENS
PLACEBO TENS
THEN ACTIVE
TENS TOTAL
Total 12 12 24
Age
Mean (SD) 67.7 (12.1) 76.4 (8.2) 72.0 (11.1)
Median (range) 69 (40, 83) 79 (62, 91) 74 (40, 91)
Gender
Male 9 9 18
Female 3 3 6
Type of primary cancer
Prostate 5 7 12
Breast 3 2 5
Lung 1 2 3
Thyroid 1 0 1
Renal 1 0 1
Other 1 1 2
Site of TENS application
Lateral pelvis 2 4 6
Lower limb 2 3 5
Sacral spine 1 3 4
Lumbar spine 2 1 3
Ribs 2 1 3
Thoracic spine 2 0 2
Cervical spine 1 0 1
ECOG performance
status
(1) Restricted
strenuous activity
213
(2) Up and about 50%
of waking hours
7411
(3) Capable of limited
self care
279
(4) Completely
disabled
101
Radiotherapy 10 9 19
Bisphosphonates 5 3 8
Strong opioids 11 10 21
Abbreviation: TENS, transcutaneous electrical nerve stimulation.
Bennett et al 355
80% power at the 5% 2-sided significance level using
Student’s t-test for difference of means. Allowing for
an approximate 25% dropout rate, a total of 298
patients would need to be recruited (149 patients per
sequence).
As a result of the difference in pain relief on move-
ment, we also calculated sample sizes based on the pro-
portion of patients with good or very good pain relief on
movement. Based on the results of this feasibility study,
a total sample size of 112 patients would be required
to detect an absolute increase of 30% in the proportion
of patients experiencing good or very good pain relief
with active TENS, compared with placebo TENS (we ob-
served a 36.8% difference in this feasibility study). This
assumes 26.3% of patients already experience this level
of pain relief with placebo TENS, and is based on a chi-
squared test of equal proportions with 80% power and
a 2-sided significance level of 5%.
Patient Preference
At the conclusion of the study, 10 patients preferred
active TENS, 4 preferred placebo TENS, and 5 expressed
no preference. A small majority of patients favored nu-
merical rather than verbal rating scales for assessing
pain intensity (9 patients vs 7 patients, respectively). In
contrast, patients favored verbal rating scales for assess-
ing pain relief (11 patients vs 7 patients). Based on
satisfaction questionnaires, patients found TENS devices
easy to use and comfortable.
Toxicity and Adverse Effects
Five patients withdrew from the study after first treat-
ment. Three were withdrawn by their clinician due to de-
teriorating performance status (deemed unrelated to
the TENS application) and 2 patients withdrew because
of increased pain during TENS application: 1 following
active TENS and 1 following placebo TENS. Overall, 9 pa-
tients experienced adverse events and median number
of adverse events per patient was 2 (range 1, 6). Distribu-
tion of adverse events was similar following active or
placebo TENS applications (Table 4).
Feasibility of Blinding and Crossover
The median interval between each intervention period
was 3 days (range 2 to 5 days). At the second visit, 1 pa-
tient scheduled to receive active TENS had no pain on
movement or at rest at baseline, and 1 patient had a base-
line NRS pain score of 2 on movement. Similarly, 2 pa-
tients scheduled to receive placebo TENS at the second
visit had baseline NRS pain scores of less than 3 (1 and
2). The washout period for these 4 patients with pain
on movement and on rest less than 3/10 at second visit
was 3 days for 3 patients, and 4 days for the other, and
were not different from the median washout for the
whole sample. With regard to pain levels at rest and on
Table 2. Numerical Pain Scores; Mean (SD)
BASELINE
ACTIVE
TENS
1HOUR
ACTIVE
TENS
CHANGE AT
1HOUR
ACTIVE TENS
BASELINE
PLACEBO
TENS
1HOUR
PLACEBO
TENS
CHANGE AT
1HOUR
PLACEBO TENS
DIFFERENCE IN CHANGE AT
1HOUR
(ACTIVE-PLACEBO)
95% CI
FOR
DIFFERENCE
Pain intensity
at rest*
2.95 (2.95) 2.11 (2.42) –.84 ( 2.17) 3.95 (2.59) 1.79 (2.18) –2.16 ( 3.2) 1.32 (3.48) –.36, 2.99
Pain intensity
on movement*
5.16 (2.17) 2.84 (2.17) –2.32 ( 1.8) 5.05 (2.53) 3.05 (2.46) –2 ( 3.18) –.32 (3.18) –1.85, 1.22
Difference at 1 hour
(Active-Placebo)
95% CI
for difference
Pain relief
at rest**
— 33.2 (34.3) — 34.7 (38.5) –1.57 (46.3) –23.91, 20.76
Pain relief
on movement**
— 52.6 (33.8) — 38.4 (36.3) 14.2 (36.4) –3.34, 31.76
Abbreviation: TENS, transcutaneous electrical nerve stimulation; CI, confidence interval.
*A negative difference implies more pain reduction at 1 hour on Active TENS.
**A positive difference implies more pain relief on Active TENS.
Table 3. Verbal Pain Relief Scores at 1 Hour
NUMBER OF PATIENTS
WITH EACH RESPONSE
OUTCOME
TENS
TYPE
VERY
GOOD GOOD MODERATE MILD NONE
Pain at rest Active 3 3 5 3 5
Placebo 5 2 2 3 7
Pain on
movement
Active 4 8 1 4 2
Placebo 2 3 4 4 6
Table 4. Adverse Events by Occurrence and
Relationship to TENS Application
SAME DAY
AS ACTIVE
TENS
SAME DAY
AS PLACEBO
TENS
AFTER
ACTIVE
TENS*
AFTER
PLACEBO
TENS*TOTAL
Relationship
Not related 1 2 6 8 17
Possibly related 1 1 2
Probably related 1 1 2
Related 1 1
*In washout period following TENS application.
356 TENS for Cancer Bone Pain
movement, 79% (15/19) patients had pain levels >3 on
movement at their second visit, compared to 53% (10/
19) patients with this level of pain at rest. Considering
baseline pain levels at both visits, 16% (4/19) patients
had pain at rest levels <3 at both visits, and only 5%
(1/19) patients had pain on movement <3 at both visits.
After both TENS applications had been completed, 11
out of 19 patients thought that placebo TENS was used in
the study and of these 11 patients, 10 correctly identified
the placebo. Blinding was judged by the research nurse-
observer to have been successfully concealed to them in
15 of 19 patients (ie, the patient did not reveal the sensa-
tion that they were experiencing).
Discussion
In this randomized controlled crossover feasibility
study, we found that TENS was acceptable to patients
and well tolerated with few adverse effects. However,
we recruited fewer than expected patients and 3 patients
were withdrawn because of deteriorating performance
status. This recruitment and attrition reflects the rela-
tively poor health of the population in which we were
researching and highlights the need for multiple recruit-
ment centres in a phase III study if based within palliative
care services. Two patients withdrew because of in-
creased pain though it is likely that natural variations in
baseline pain intensity might have accounted for this
observation as it occurred in both treatment arms.
The sample-size calculation is based on the results of the
19 patients who completed both interventions and
should, as a stand-alone estimate, be regarded with cau-
tion. Browne
5
suggests that 30 patients are required to es-
timate a parameter for future sample-size calculation;
however, other authors have detailed that as few as 12 pa-
tients per arm is sufficient to estimate data for a future
study.
28
The data generated from the 19 patients receiving
both applications is therefore deemed sufficient to power
a sample-size calculation for a future phase III study.
The confidence intervals around the difference in
change in pain levels on movement at 1 hour suggest
that TENS has the potential to decrease pain on move-
ment more than pain at rest, which is reflected by both
pain-intensity and pain-relief scales. This might reflect
the fact that mean pain intensity at rest was less than
that on movement and therefore it will be easier to dem-
onstrate change in scores on movement than at rest.
However, when we considered the verbal pain-relief
scale, we found that there were potentially large differ-
ences to be seen between the proportion of patients ex-
periencing good or very good pain relief on movement
on active TENS as compared to placebo TENS. These dif-
ferences were more evident than for the change in NRS
pain-intensity scores on movement.
Although international guidance recommends the use
of pain-intensity numerical-rating scales,
17
the results of
this feasibility study have shown that the pain relief scale
may better reflect the activity and clinical effectiveness
of TENS. We identified potentially larger improvements
in pain relief than reduction in pain intensity with active
TENS and further feasibility work on this outcome using
qualitative methods is warranted. Using pain intensity as
an outcome measure only allows patients to report their
experience at fixed points in time. It is the researcher
(and not the patient) who infers pain relief after sub-
tracting one value from another. In contrast, using pain
relief as an outcome measure allows patients to compare
their pain experience at 2 time points and report
whether they are better or worse with the TENS applica-
tion. Since the patient is the best judge of their perceived
pain relief, assessing this outcome directly is preferential
to assessing this indirectly by comparing numerical-
rating scores. Further work is currently being carried
out to consider the feasibility of using verbal-rating
scales to assess pain relief on movement.
It is important to reiterate that the aim of this feasibil-
ity study was not to investigate an intervention effect,
but to obtain data for a sample-size calculation for
a phase III study in which the intervention effect can be
formally addressed and sequence effects would be inves-
tigated in detail. Within this feasibility study, however,
sequence effects were informally investigated and we
did not consider that the conclusions of the study—in
particular, the choice of primary outcome to use in future
trials—were affected by the sequence order due to the
crossover design. We did not assess expectations of treat-
ment prior to the study and this may have accounted for
some of our findings. For example, Linde et al
32
have
shown that high expectations of treatment strongly pre-
dicts better outcomes, independent of other con-
founders such as treatment allocation. It is possible
that the tingling sensation that accompanies active
TENS resulted in an expectation by patients that this de-
vice would help their movement-related pain (as op-
posed to a direct effect on nociceptive pathways) and
this might explain the observed differences in verbal
pain relief between the 2 arms.
We were not convinced that blinding was maintained
for patients, although blinding of the research nurse-ob-
server was reasonably maintained. We will explore the
feasibility of using a more authentic control arm, for ex-
ample, using a placebo TENS device that delivers an initial
stimulus for approximately 40 seconds after which time
the current amplitude resets to zero for the remainder
of the study duration, or applying active TENS to a distant
nonpainful site. Despite a relatively short interval be-
tween TENS applications (2–5 days), 4 patients (2 per se-
quence) had pain intensity of less than 3/10 on either
rest or movement at second visit. There is no physiological
evidence that TENS has analgesic effects that last for
many days once removed,
11,20
and so this finding proba-
bly reflects natural variations in pain intensity because it
occurred following active and placebo TENS. This would
support using pain on movement as a primary outcome
rather than pain at rest in a future clinical trial. The wash-
out period of 2 to 5 days appears to be adequate.
The strengths of this study include the novel ap-
proach of using TENS for cancer bone pain, the ran-
domized crossover design which allowed patients to
act as their own controls, the assessment of blinding
adequacy and assessment of preference for outcome
measures for pain at rest and on movement. Our study
Bennett et al 357
was designed to examine the feasibility of using this
methodology to test the efficacy of TENS in cancer
bone pain. We recruited a small sample of patients
with advanced disease but 20% of patients did not
complete both trial arms because of deteriorating per-
formance status or pain. We were not able to maintain
adequate blinding in this clinical context and this may
have impacted the observed difference in verbal ratings
of pain relief on movement. We were able to identify
the most appropriate outcome measure, confirm toler-
ability and safety of this technique, and generate suffi-
cient data to calculate the sample size required in
a future trial. Further feasibility work is required to im-
prove recruitment rate, examine differences perceived
by patients in the meaning of pain intensity and pain
relief when using TENS, and investigate a more authen-
tic control arm.
Patients with movement-related cancer bone pain re-
main poorly served by conventional treatment strate-
gies, and better approaches are urgently needed. This
feasibility study suggests that TENS may have benefits
on movement related pain but that further improve-
ments in study design are needed before testing this
assumption formally in a large clinical trial.
Acknowledgments
We are very grateful to Helen Radford and Robert
Searle in Leeds, and Gail Wiley and Claire Capewell in
Lancaster, for conducting TENS applications.
References
1. Aarskog R, Johnson MI, Demmink JH, Lofthus A, Iversen V,
Lopes-Martins R, Joensen J, Bjordal JM: Is mechanical pain
threshold after transcutaneous electrical nerve stimulation
(TENS) increased locally and unilaterally? A randomized pla-
cebo controlled trial in healthy subjects. Physiother Res Int
12:251-263, 2007
2. Anderson GL, Prentice RL: Individually randomized inter-
vention trials for disease prevention and control. Statistical
Methods in Medical Research 8:287-309, 1999
3. Bjordal JM, Johnson MI, Ljunggreen AE: Transcutaneous
electrical nerve stimulation (TENS) can reduce postoperative
analgesic consumption. A meta-analysis with assessment of
optimal treatment parameters for postoperative pain. Eur
J Pain 7:181-188, 2003
4. Bjordal JM, Johnson MI, Lopes-Martins RA, Bogen B,
Chow R, Ljunggren AE: Short term efficacy of physical inter-
ventions in osteoarthritic knee pain. A systematic review
and meta-analysis of randomised placebo-controlled trials.
BMC Musculoskelet Disord 8:51, 2007
5. Browne RH: On the use of a pilot sample for sample size
determination. Statistics in Medicine 14:1933-1940, 1995
6. Cancer Therapy Evaluation Program: Common Terminol-
ogy Criteria for Adverse Events v3.0 (CTCAE). National Can-
cer Institute. Available at: www.cancer.gov; 2006
7. Carroll D, Tramer M, McQuay H, Nye B, Moore A: Transcu-
taneous electrical nerve stimulation in labour pain: A sys-
tematic review. Br J Obstet Gynaecol 104:169-175, 1997
8. Charlton J: Core Curriculum for Professional Education in
Pain, 3rd Edition. Seattle, IASP Press, 2005, pp 93-96
9. Chartered Society of Physiotherapy: Guidance for the
Clinical Use of Electrophysical Agents, 1st Edition. London,
Chartered Society of Physiotherapy, 2006
10. Chesterton LS, Barlas P, Foster NE, Lundeberg T,
Wright CC, Baxter GD: Sensory stimulation (TENS): Effects
of parameter manipulation on mechanical pain thresholds
in healthy human subjects. Pain 99:253-262, 2002
11. Chesterton LS, Foster NE, Wright CC, Baxter DG, Barlas P:
Effects of TENS frequency, intensity and stimulation site pa-
rameter manipulation on pressure pain thresholds in
healthy human subjects. Pain 106:73-80, 2003
12. Chow E, Hoskin P, van der Linden Y, Bottomley A,
Velikova G: Quality of life and symptom end points in palli-
ative bone metastases trials. Clinical Oncology 18:67-69,
2006
13. Coluzzi PH, Schwartzberg L, Conroy JD, Charapata S,
Gay M, Busch MA, Chavez J, Ashley J, Lebo D,
McCracken M, Portenoy RK: Breakthrough cancer pain: A
randomized controlled trial comparing oral transmucosal
fentanyl citrate and morphine sulphate immediate release.
Pain 91:123-130, 2001
14. Craig P, Dieppe P, Macintyre S, Michie S, Nazareth I,
Petticrew M: Medical Research Council Guidance. Develop-
ing and evaluating complex interventions: The new Medical
Research Council guidance. BMJ 337:1655, 2008
15. Dworkin RH, Corbin AE, Young JP, Sharma U,
LaMoreaux L, Bockbrader H, Garofalo EA, Poole RM:
Pregabalin for the treatment of postherpetic neuralgia:
A randomized, placebo-controlled trial. Neurology 60:
1274-1283, 2003
16. Dworkin RH, Turk DC, Farrar JT, Haythornthwaite JA,
Jensen MP, Katz NP, Kerns RD, Stucki G, Allen RR,
Bellamy N, Carr DB, Chandler J, Cowan P, Dionne R,
Galer BS, Hertz S, Jadad AR, Kramer LD, Manning DC,
Martin S, McCormick CG, McDermott MP, McGrath P,
Quessy S, Rappaport BA, Robbins W, Robinson JP,
Rothman M, Royal MA, Simon L, Stauffer JW, Stein W,
Tollett J, Wernicke J, Witter J: Core outcome measures for
chronic pain clinical trials: IMMPACT recommendations.
Pain 113:9-19, 2005
17. Dworkin RH, Turk DC, Wyrwich KW, Beaton D,
Cleeland CS, Farrar JT, Haythornthwaite JA, Jensen MP,
Kerns RD, Ader DN, Brandenburg N, Burke LB, Cella D,
Chandler J, Cowan P, Dimitrova R, Dionne R, Hertz S,
Jadad AR, Katz NP, Kehlet H, Kramer LD, Manning DC,
McCormick C, McDermott MP, McQuay HJ, Patel S,
Porter L, Quessy S, Rappaport BA, Rauschkolb C,
Revicki DA, Rothman M, Schmader KE, Stacey BR,
Stauffer JW, von Stein T, White RE, Witter J, Zavisic S: Inter-
preting the clinical importance of treatment outcomes in
chronic pain clinical trials: IMMPACT recommendations. J
Pain 9:105-121, 2008
18. Farrar JT, Portenoy RK, Berlin JA, Kinman JL, Strom BL:
Defining the clinically important difference in pain outcome
measures. Pain 88:287-294, 2000
19. Farrar JT, Young JP, LaMoreaux L, Werth JL, Poole RM:
Clinical importance of changes in chronic pain intensity
358 TENS for Cancer Bone Pain
measured on an 11- point numerical pain rating scale. Pain
94:149-158, 2001
20. Garrison DW, Foreman RD: Decreased activity of sponta-
neous and noxiously evoked dorsal horn cells during trans-
cutaneous electrical nerve stimulation (TENS). Pain 58:
309-315, 1994
21. Garrison D, Foreman R: Effects of transcutaneous elec-
trical nerve stimulation (TENS) electrode placement on
spontaneous and noxiously evoked dorsal horn cell activity
in the cat. Neuromodulation 5:231-237, 2002
22. Jensen MP, Karoly P: Self-report scales and procedures
for assessing pain in adults, in Turk DC, Melzack R (eds):
Handbook of Pain Assessment, 2nd Edition. New York,
Guilford Press, 2001, pp 15-34
23. Jensen MP, Turner JA, Romano JM, Fisher LD: Compara-
tive reliability and validity of chronic pain intensity mea-
sures. Pain 83:157-162, 1999
24. Johnson MJ: Transcutaneous electrical nerve stimula-
tion, in Watson T (ed): Electrotherapy: Evidence Based
Practice, 12th Edition. London, UK, Churchill Livingstone,
2008, pp 253-296
25. Johnson MI, Oxberry SG, Robb K: Stimulation-induced
analgesia, in Sykes N, Bennett MI, Yuan C-S (eds): Cancer
Pain. Clinical Pain Management, 2nd Edition. London, UK,
Hodder Arnold, 2008, pp 235-250
26. Johnson M, Martinson M: Efficacy of electrical nerve
stimulation for chronic musculoskeletal pain: a meta-anal-
ysis of randomized controlled trials. Pain 130:157-165,
2007
27. Johnson MI, Tabasam G: An investigation into the anal-
gesic effects of interferential and transcutaneous electrical
nerve stimulation on experimentally induced ischemic pain
in otherwise pain-free volunteers. Phys Ther 83:208-223,
2003
28. Julious S: Sample size of 12 per group rule of thumb for
a pilot study. Pharmaceutical Statistics 4:287-291, 2005
29. Khadilkar A, Odebiyi DO, Brosseau L, Wells GA: Transcu-
taneous electrical nerve stimulation (TENS) versus placebo
for chronic low-back pain. Cochrane Database Syst Rev
(4):CD003008, 2008
30. Lancaster GA, Dodd S, Williamson PR: Design and
analysis of pilot studies: Recommendations for good prac-
tice. Journal of Evaluation in Clinical Practice 10:307-312,
2004
31. Lee KH, Chung JM, Willis WD Jr: Inhibition of primate
spinothalamic tract cells by TENS. J Neurosurg 62:276-287,
1985
32. Linde K, Witt CM, Streng A, Weidenhammer W,
Wagenpfeil S, Brinkhaus B, Willich SN, Melchart D: The
impact of patient expectations on outcomes in four random-
ized controlled trials of acupuncture in patients with
chronic pain. Pain 128:264-271, 2007
33. Liu XG, Morton CR, Azkue JJ, Zimmermann M,
Sandkuhler J: Long-term depression of C-fibre-evoked spi-
nal field potentials by stimulation of primary afferent A del-
tafibres in the adult rat. Eur J Neurosci 10:3069-3075, 1998
34. Ma YT, Sluka KA: Reduction in inflammation-induced
sensitization of dorsal horn neurons by transcutaneous elec-
trical nerve stimulation in anesthetized rats. Exp Brain Res
13:94-102, 2001
35. Maeda Y, Lisi TL, Vance CGT, Sluka KA: Release of GABA
and the activation of GABAA in the spinal cord mediates the
effects of TENS in rats. Brain Res 1136:43-50, 2007
36. McQuay HJ, Collins SL, Carroll D, Moore RA: Radiother-
apy for the palliation of painful bone metastases. Cochrane
Database Syst Rev (3):CD001793, 2000
37. Melzack R: The short-form McGill Pain Questionnaire.
Pain 30:191-197, 1987
38. Mercadante S: Malignant bone pain: Pathophysiology
and treatment. Pain 69:1-18, 1997
39. Nnoaham KE, Kumbang J: Transcutaneous electrical
nerve stimulation (TENS) for chronic pain. Cochrane Data-
base Syst Rev (3):CD003222, 2008
40. Portenoy RK, Payne D, Jacobsen P: Breakthrough pain:
Characteristics and impact in patients with cancer pain.
Pain 81:129-134, 1999
41. Robb K, Bennett MI, Johnson MI, Simpson KH, Oxberry
SG: Transcutaneous Electrical Nerve Stimulation for cancer
pain in adults. Cochrane Database Syst Rev (3):CD006276,
2008
42. Searle RD, Bennett MI, Johnson MI, Callin S, Radford H:
Transcutaneous electrical nerve stimulation (TENS) for can-
cer bone pain (letter). Palliat Med 22:878-879, 2008
43. Small EJ, Smith MR, Seaman JJ, Petrone S, Kowalski MO:
Combined analysis of two multicentre randomized placebo
controlled studies of pamidronate disodium for the pallia-
tion of bone pain in men with metastatic prostate cancer.
Journal of Clinical Oncology 21:4277-4284, 2003
44. Walsh DM, Howe TE, Johnson MI, Sluka KA: Transcuta-
neous electrical nerve stimulation for acute pain. Cochrane
Database Syst Rev (2):CD006142, 2009
45. Wong R & Wiffen PJ. Bisphosphonates for the relief of
pain secondary to bone metastases. Cochrane Database
Syst Rev (2):CD002068, 2002
46. Urch CE, Donovan-Rodriguez T, Dickenson AH: Alter-
ations in dorsal horn neurones in a rat model of cancer
induced bone pain. Pain 106:347-356, 2003
Bennett et al 359
