Double blinded placebo controlled crossover study of High

Report
Double blinded placebo controlled
crossover study of High Frequency
SCS
C. Perruchoud, E. Buchser, A Durrer, B. Rutschmann, M. Rosato1,2, N. Bovet, A.
Gulve3, M Brookes3, G. Madzinga3 A. Batterham4 and S . Eldabe3
1.Dept of Anaesthesiology and Pain Management Centre for Neuromodulation, Morges. Switzerland
2.Dept of Anaesthesiology and Pain Management University Hospital Center and University of Lausanne Switzerland
3. Dept of Anaesthesiology and Pain Management The James Cook University Hospital Middlesbrough UK
4. University of Teesside Middlesbrough UK
Disclosure
o Medtronic funded the study and provided the technical
support for IPG programming.
o No member of Medtronic personnel contributed to the
study design, data collection or data analysis.
o The study was sponsored by the author’s institutions
o E. Buchser, S. Eldabe and C. Perruchoud are
Medtronic Consultants
Background
o Double blinded controlled studies have not been feasible
in SCS
o In a randomised placebo controlled trial of 12 RA patients
Eddicks showed Sub-threshold stimulation (SS) to have an
analgesic effect1
o High Frequency (HF) stimulation and Burst stimulation
produce analgesia without paraesthesias2,3
o In this study we aim to compare HF stimulation and Sham
Stimulation in a double blind crossover design
1. Eddicks, S., et al., Thoracic spinal cord stimulation improves functional status and relieves symptoms in patients with refractory angina
pectoris: the first placebo-controlled randomised study. Heart, 2007. 93(5): p. 585-90.
2. De Ridder, D., et al., Burst spinal cord stimulation: toward paresthesia-free pain suppression. Neurosurgery, 2010. 66(5): p. 986-90.
3. Smet I, Van Buyten JP and Alkaisy A. Successful treatment of low back pain with a novel neurostimulation device. Proceeds of the North
American Neuromodulation Society Meeting 2011 Las Vegas
Patients and Methods
o 40 Patients with existing SCS systems (stable pain relief)
implanted for low back and or leg pain enrolled at Morges
Switzerland and Middlesbrough UK.


Patients Implanted with Restore Advanced/ Ultra/ Sensor and Prime
Advanced.
A special study programmer allowed any of the IPGs to deliver the
following settings
 Frequency of 5'000 Hz continuous stimulation
 Pulse width 60 μsec
 Amplitude increased gradually to threshold then reduced again to a point
where the patient is unable to feel paraesthesias regardless of the position
End of Study
NS is restarted
Randomisation
Patient Inclusion
Sequence 1
DATA
1
DATA
3
DATA
5
NS
NS to HF
HF to NS
DATA
7
NS to Sham
Sequence 2
DATA
2
NS
Visit 1
Visit 2
DATA
4
NS to Sham
Visit 3
Normal Stimulation
HF Stimulation 5Khz
Sham
2 week periods
DATA
6
DATA
8
Sham to NS
NS to HF
Visit 4
Visit 5
DATA
X
5 Day Data Collection
PGIC
Randomisation & Blinding
o Randomisation

Subjects were randomised to receive sequence 1 or 2 using
central randomisation service
o Blinding Precautions



Research teams were split into 2 groups of blinded and unblinded
personnel with no crossover in personnel
In patients with rechargeable devices we set an IPG current
leakage in the sham period to similar level of the current usage
during HFSCS
Patients were asked to guess the nature of the treatment they
received to test the blinding strength at visits 3 and 5
Primary Outcome measure: PGIC
o Patient Global Impression of Change (PGIC)

end of each 2-week period. "Since the last visit to the pain clinic my
overall pain control is:
 very much improved
 much improved
 minimally improved
 no change
 minimally worse
 much worse
 very much worse

A responder is defined as a subject reporting at least a “minimal
improvement”, with the point of reference being the last visit under
conventional stimulation
Study Flow
40 Patients
Consented
2 patient withdrew
consent
17 Completed Sequence 1
HF first
38 Patients
Randomized
5
1
1
2
1
Withdrawn
battery failure
lead fracture
withdrew consent
Battery flipping
33 Patients
Analysed
16 Completed Sequence 2
Sham First
Results: Sample Baseline
Characteristics
Data are mean (SD).
Results: Blinding
o None of the subjects reported a sensation of
paraesthesias in either group
o At visit 3, the proportion of patients guessing their
therapy group correctly was 15/33 = 45%
o At visit 5 the proportion guessed correctly was 18/33
= 55%.
o Neither is significantly different from chance (50%)
Results: Patient Global Impression of
Change (PGIC)
o In sequence 1 (HF first) 9/17 (52.9%)
responded to HF vs. 2/17 (11.7%) to 60%
sham; a difference of (41.2%)
50%
o In sequence 2 (sham first) 5/16
(31.2%) responded to HF vs. 8/16
(50%) to sham; a difference of
(18.8%, in favour of sham)
o The overall treatment effect (benefit
of HF vs. sham) is the average of
these two proportions: (0.412 and 0.188) = 0.112 (11.2%)1.
40%
30%
HF
20%
Sham
10%
0%
Seq 1 HF Seq 2 Sham
First
First
a responder is defined as a subject reporting at least a ‘minimal improvement’, with the point of reference being the
last visit under conventional stimulation)
1. Schouten H, Kester A: A simple analysis of a simple crossover trial with a dichotomous outcome measure. Stat
Med 2010; 29: 193-8.
The Period Effect
o We observed a substantial period effect, irrespective of treatment
received, with 17/33 (51.5%) patients benefiting at Visit 3
versus 7/33 (21.2%) at Visit 5; mean difference in proportions =
30% ( P=0.006).
18
16
14
12
10
8
6
4
2
0
HF
Sham
Total
Treatment Proposed First
Treatment Proposed Last
Conclusions I
o This is the first double blinded RCT in spinal cord
stimulation
o We showed a minimal advantage of HF over sham as a
proportion of 11.2% ( P=0.30).
o We showed a substantial period effect with a greater
proportion of responders at visit 3 versus visit 5
(difference = 30% P=0.006).
o Both VAS & EQ-5D show the same consistent pattern of
period effect (NS differences)
Conclusions II
o At settings used, HF offers small advantage over sham.
Only the order of the treatment in the sequence, dictates
a significant effect
o We believe this period effect to be related to patient
expectation, or a “study” effect.
o Study Limitation




These findings cannot be extrapolated to other stimulation
frequencies/modes applied for a longer duration of time on
stimulation naïve patients
We applied sub-threshold amplitudes; stimulation above sensory
threshold may have been unpleasant and resulted in ublinding
The short duration of HFSCS and sham may have allowed a
carry over effect of conventional SCS to play a significant role
Cathode position
Thank you
Middlesbrough Team
A Gulve
Morges Team
G. Madzinga
C. Perruchoud
F. Garner
A Durrer,
S. West
B. Rutschmann,
M. Brookes
M. Rosato
S. Eldabe
Prof. A. Batterham Teesside University
N. Bovet
E. Buchser
Thank You
o Morges Team





C. Perruchoud
A Durrer,
B. Rutschmann,
M. Rosato
E. Buchser
o Middlesbrough Team






A Gulve
G. Madzinga
F Garner
S. West
M. Brookes
S. Eldabe
• Prof. A. Batterham Teesside University
Sample Size and statistical
methods
o Sample size

Using methods proposed by Schouten and Kester1, we
estimated that a sample of 38 patients was required to have
90% power (2P=0.05) detect a difference in the proportion of
patients responding to treatment in HF vs. sham of 25%
o Statistical analysis


PGIC :methods of Schouten and Kester, in which the
treatment difference (HFSCS-sham) for each of the two
treatment sequences is computed and then averaged.
Pain VAS and EQ5-D index analysed using a conventional
within-subjects model accounting for the period effect and
utilizing the baseline scores before each treatment (visits 2
and 4) in an ANCOVA model
1. Schouten H, Kester A: A simple analysis of a simple crossover trial with a dichotomous outcome measure. Stat
Med 2010; 29: 193-8.
Secondary Outcome measures
o Pain VAS


mean pain VAS on the HF treatment was 4.35 cm vs. 4.26 cm
on the sham (P=0.82).
Irrespective of treatment received, the mean VAS in the first
treatment period (visit 3) was 3.99 cm vs. 4.63 cm in the
second period (Visit 5) ; difference (visit 3 minus visit 5) = 0.64 cm (95% CI, -1.41 to 0.14 cm; P=0.11).
Visit 3
VAS 3.99
Visit 5
VAS 4.63
Secondary Outcome Measures
o Mean EQ5D-index on the HF treatment was 0.480 vs.
0.463 on the sham; difference (HF minus sham) =
0.017 (95% CI, -0.101 to 0.135; P=0.78).
o Mean EQ5-D index was 0.510 at visit 3 vs. 0.432 at
Visit 5; difference (visit 3 minus visit 5) = 0.078 (95%
CI, -0.041 to 0.196 ; P=0.19).
Visit 3
EQ5D 0.51
Visit 5
EQ5D 0.432
The Period Effect
60%
50%
40%
HF
30%
Sham
20%
10%
0%
Sequence 1 HF First
Sequence 2 Sham First

similar documents