Motor cortex stimulation for refractory neuropathic pain: Four year outcome and predictors of efficacy

Pain 118 (2005) 43–52 www.elsevier.com/locate/pain

Motor cortex stimulation for refractory neuropathic pain: Four year outcome and predictors of efficacy Christophe Nutia,d,*, Roland Peyronb,d, Luis Garcia-Larread, Jacques Brunona, Bernard Laurentb,d, Marc Sindouc, Patrick Mertensc,d b

a Department of Neurosurgery, CHU Saint-Etienne, Bd Pasteur, 42055 Saint-Etienne cedex 2, France Department of Neurology & Pain center, CHU Saint-Etienne, Bd Pasteur, 42055 Saint-Etienne cedex 2, France c Department of Neurosurgery A, CHU Lyon, France d INSERM EMI 0342, UCBLyon1 & UJM Saint-Etienne, France

Received 28 September 2004; received in revised form 11 July 2005; accepted 25 July 2005

Abstract Thirty-one patients with medically refractory neuropathic pain were included in a prospective evaluation of motor cortex stimulation. The long-term outcome was evaluated using five variables: (a) rate (percentage) of pain relief, (b) pain scores as assessed on VAS, (c) postoperative decrease in VAS scores, (d) reduction in analgesic drug intake, (e) a dichotomic (yes/no) response to the question whether the patient would accept, under similar circumstances, to be operated on again. Pain relief was rated as excellent (O70 % pain relief) in 10 % of cases, good (40-69 %) in 42 %, poor (10-39 %) in 35 % and negligible (0-9 %) in 13 %. Intake of analgesic drugs was decreased in 52 % of patients and unchanged in 45 % (unavailable data in 3 %), with complete withdrawal of analgesic drugs in 36 % of patients. Twenty-one patients (70 %) declared themselves favourable to re-intervention if the same beneficial outcome could be guaranteed. Neither preoperative motor status, pain characteristics, type or localisation of lesions, quantitative sensory testing, Somatosensory Evoked Potentials, nor the interval between pain and surgery were found to predict the efficacy of MCS. The level of pain relief, as evaluated in the first month following implantation was a strong predictor of long-term relief (regression analysis, RZ0.744; p!0.0001). These results confirm that MCS can be a satisfactory and durable alternative to medical treatments in patients with refractory pain, and suggest that the efficacy of MCS may be predicted in the first month of therapy. q 2005 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. Keywords: Brain stimulation; Motor cortex; Chronic pain; Central pain; Analgesia

1. Introduction Since, the first reports by Tsubokawa et al. (1991,1993) showing the analgesic effects of stimulating (electrically and chronically) the primary motor cortex, this technique has been applied worldwide in an attempt to alleviate medically refractory neuropathic pain. However, the real analgesic effect of Motor Cortex Stimulation (MCS) has still

* Corresponding author. Corresponding author. Address: Department of Neurosurgery, CHU Saint-Etienne, Bd Pasteur, 42055 Saint-Etienne cedex 2, France. Tel.: C33 477 127 723; fax: C33 477 120 544. E-mail address: [email protected] (C. Nuti).

to be appraised and this is limiting the development of this technique. MCS is proposed as a ‘last chance’ therapy for patients in whom iterative trials with various drugs has failed to induce pain relief. This technique is generally proposed to patients with post-stroke or post-traumatic neuropathic pain and in trigeminal deafferentation pain (Katayama et al., 1998; Mertens et al., 1999; Meyerson et al., 1993; Nguyen et al., 1997; 1999; Tsubokawa et al., 1991; 1993). Since, the technique was elaborated from empirical experience (Tsubokawa et al., 1991; 1993) and since the mechanisms underlying the effects of MCS are not yet fully documented (Brown and Barbaro, 2003; Garcia-Larrea

0304-3959/$20.00 q 2005 International Association for the Study of Pain. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.pain.2005.07.020

44

C. Nuti et al. / Pain 118 (2005) 43–52

et al., 1999; Hanajima et al., 2002; Peyron et al., 1995), all adjustments to the stimulation parameters and more generally attempts to improve the technique still depend on iterative empirical testing. We postulated that observing the long-term outcome in patients on whom MCS had been used, combined with the gathering of extensive clinical and paraclinical data, could provide information on the clinical interest of the technique and help us to assess the actual analgesic effect of MCS in patients with neuropathic pain. To this end, we conducted a prospective study of 31 consecutive patients to assess, over a long period of time, the efficacy of the technique, with the assumption that a stable effect over time would tend to confirm the benefit of MCS on neuropathic pain. Analysis of clinical, radiological and electrophysiological correlations were made in an attempt to identify predictive factors for the efficacy of MCS.

2. Methods and patients 2.1. Patients After providing written informed consent, 31 consecutive patients with medically refractory neuropathic pain were included in a prospective evaluation of MCS between January 1992 and February 2003. The study consisted of a pre-operative clinical, psychological, radiological (MRI) and electrophysiological (somatosensory and Laser Evoked Potentials, quantitative sensory testing) assessment according to a protocol submitted to the local ethics committee of St Etienne University Hospital. The mean postoperative follow-up was four years (49 months, ranges: 2–104). The clinical data from these patients are summarized in Table 1. Their ages at the date of surgery ranged from 27 to 72 years. Patients had central pain caused by haemorrhagic (nZ11; 35.5%) or ischemic (nZ11; 35.5%) stroke, ruptured vascular

Table 1 Patients’ clinical and electrophysiological data P

Sex, age

Type of lesion

Type of pain

1 2 3 4 5 6 7 8 9 10 11

M, 41 M, 52 F, 48 M, 71 F, 56 M, 71 M, 44 F, 41 F, 72 M, 41 F, 62

Haem Avulsion Isch Haem Isch Isch Avulsion Isch Isch Isch Haem

C C C C C C C C C C C (–P)

12 13 14 15 16 17 18 19 20 21 22 23

M, 43 M, 56 F, 72 M, 45 M, 64 M, 71 F, 52 M, 63 M, 27 F, 39 F, 60 F, 58

Trauma Haem Isch Isch Haem Haem Trauma Haem Avulsion Haem Haem Isch

C–P C C C C C C-P C C C C C

24 25 26 27 28 29 30 31

M, 41 M, 54 F, 46 F, 65 M, 33 M, 45 F, 50 M, 50

Trauma Isch Isch Haem Avulsion AVM Haem Trauma

C -P C C C C C C C

Anatomic localization of lesion Th-midbrain Plexus Th (VPL) Capsulo-th Medulla Parietal Plexus Medulla Parietal Medulla Capsulo-lenticular-insula (Culnaris nerve) Cervical spine (CC6) Capsulo-th Th (VPL) Medulla Parietal Th (VPL) Cervical spine (CC6) Capsulo-th Plexus Capsulo-th Parietal Parietal-insula- anterior cingulate Conus and lumbar roots Th (VPL) Parietal-insula Parietal Plexus Conus Capsulo-th Fronto-parietal

A

Hyp

Motor

SEP

ST (LEP/ QST)

Delays

(years)

L–S

P–S

K K C C C C C C C C C

K K C C K K K C K C C

Im Im N Im N Im Amput N Im N Im

Im Im Im Im N Im Im N Im N Im

n.a. n.a. n.a. Im Im Im n.a. Im n.a. Im Im

2.8 17 1.6 1.9 5.5 6.5 24 2.7 4.7 1.8 7.5

1.6 16.9 1.5 1.7 5 3.2 23.9 2.7 2.5 1.8 7

C C C C C C C K C C C

C K C K C K C K K K C K

N Im Im N Im Im Im Im Amput Im Im Im

N Im Im N N Im Im N Im Im N Im

n.a. Im Im Im Im n.a. n.a. N n.a. n.a. Im Im

n.a. 8 2.8 1.7 4.4 1.5 6 3.6 9.6 3.75 20 3.6

n.a. 3 0.8 1.7 3.9 1.1 5 3.6 8.25 3.6 15 3.3

K C C C K K C C

K C C K K K C K

Im Im Im Im Im Im Im Im

Im Im Im Im Im Im Im Im

Im n.a. n.a. Im Im Im n.a. Im

6 1.6 7.9 n.a. 4.75 24 3 9

5.5 1.5 7.8 n.a. 4.6 14 1.8 6

Abbreviations: M, male; F, female; Haem, haemorrhagic stroke; Isch, ischemic stroke; AVM: arteriovenous malformation; C, central pain; P, neuropathic pain of peripheral origin; Th, thalamus; Medulla, medulla oblongata; A, allodyniae; Hyp, hyperpathia; C, presence; Amput, amputated limb; SEP, somatosensory evoked potentials; Im, impaired; N, normal; ST, spino-thalamic; LEP, laser-evoked potentials; QST, quantitative sensory testing; n.a., not available; L, lesion; P, pain; S, surgery.

C. Nuti et al. / Pain 118 (2005) 43–52

malformation in the conus medullaris (nZ1; 3.2%), myelopathy complicating herniated discs (nZ3; 9.7%) and brain trauma with frontal and parietal haematoma (nZ1; 3.2%). Four of these patients (no 11, 12, 18, 24) had additional pain related to peripheral nerve injury within their central pain area (cervical (nZ2) and lumbar (nZ1) radicular pain; ulnaris nerve pain (nZ1)). The last four patients (12.9%) had pain after complete brachial plexus avulsion and were considered as patients with central pain since the avulsion was proximal to the dorsal root ganglia. Lesions were right-sided in 13 patients (42%), leftsided in 15 (48.4%) and bilateral in three patients who had unilateral pain but bilateral lesions of the spinal cord. Thirty patients (97%) had spontaneous pain, which was continuous in 13 cases (42%), both continuous and paroxysmal in 15 (48.4%) and strictly paroxysmal in two (6.5%). In addition to spontaneous pain, 24 patients (77.4%) had hyperalgesia (allodynia alone in 11 cases (35.5%), both allodynia and hyperpathia in 13 cases (42%)). Pain was described as ‘burning’ in 25 patients (80.6%), ‘electrical discharge’ in 17 patients (55%), ‘pressure’ in 13 (42%), ‘cramps’ in four (13%), ‘tearing’ in four (13%), ‘stinging’ in two (6.5%) and ‘biting’ in one patient (3.2%). Precipitating factors included effort (nZ14; 45%), anxiety or emotion (nZ2; 6.5%), cold (nZ8; 26%) or heat (nZ1; 3.2%) and alcohol consumption (nZ1; 3.2%). Patients were systematically treated before surgery by oral and/or parenteral pharmacological treatments including non-opioid analgesic drugs, non-steroidal anti-inflammatory drugs, benzodiazepines, anti-convulsive and anti-depressant (tricyclic) agents. Morphine treatments had been used in nine patients before their inclusion in the present study (no 7, 12, 18, 19, 21, 22, 24, 28, 29), including intrathecal administration in one patient (no 24) with no significant effect. Trans-cutaneous stimulation had been used in 13 patients with no significant analgesic effect. Before their inclusion in the present study, five patients had had previous neurosurgical treatments. Two patients (no 20, 28) with plexus avulsion, and one with conus and lumbar roots lesions (no 24) had had microsurgical DREZotomies, which were beneficial only on paroxysmal pain. One patient (no 2) had a dorsal rhizotomy followed by a DREZotomy and then by spinal cord stimulation, both of which gave poor pain relief. Anterolateral cordotomy followed by a microsurgical DREZotomy were performed on one patient with no effect on his pain (no 7). Psychological testing (Beck questionnaire (depression or anxiety status) and pain and illness questionnaire) was conducted by an independent physician in the preoperative stage to verify the patient’s ability (i) to correctly understand the objectives and the details of the surgical treatment and (ii) to evaluate his/her pain level in the postoperative stage. The psychological interview was also used to exclude any subjects in search of financial compensation and to uncover any psychiatric comorbidity that could interfere with the results of motor cortex stimulation. 2.2. Clinical and neurophysiological investigations of the sensory systems Position sense, sensitivity to touch, vibration, cold, heat and pinprick stimuli were initially tested as part of a conventional neurological examination. Then a battery of Somatosensory Evoked Potentials (SEPs), Quantitative Somatosensory Testing (QST) and Laser Evoked Potentials (LEPs) was performed to

45

achieve clinical examination. These investigations are standardized and routinely used in our laboratory. The data collected are synthesized in Table 1. Scalp SEPs were obtained after bilateral stimulation of both the median and the posterior tibial nerves. The detailed procedures for these recordings have been described elsewhere (Garcia-Larrea et al., 1999). The results were divided into two groups i.e. normal or impaired. On this basis, 23 (74%) patients showed evidence of lemniscal dysfunction. Quantitative sensory testing was recorded using a Peltier probe applied on the skin and piloted with a MEDOCw-TSA-2001 device. Sensory thresholds for innocuous cold and heat stimuli were determined by the staircase method, and thresholds for noxious thermal stimuli by the method of limits. Measurements were taken within but also outside the painful area, both on the symptomatic side and on the other (normal) side. QST measurements were taken on the face (V2 territory), shoulder, hand, thigh, and foot. Thresholds to innocuous thermal stimuli were compared to the performance of normal subjects and were used to define spino-thalamic involvement. QST data was obtained from 20 patients, four of whom had diseases which confounded interpretation of the data. Perceptive and pain thresholds to CO2 laser were first determined by using the methods of limits. LEPs were then recorded for an energy of stimulus set to the pain threshold. Recordings were performed on the painful side and on the normal side, in painful and non-painful areas, according to the methods described elsewhere (Garcia-Larrea et al., 2002). LEPs could be recorded in 14 patients, three of whom had diseases which confounded interpretation of the data. Since, both QST and LEPs illustrate the spino-thalamic function and produced concordant results in our patients, the data from these investigations have been regrouped in Table 1. 2.3. Motor function To improve the statistical power of the tests in this sample of patients, the motor function in the painful limbs was classified in two categories segregating patients with either normal (nZ6) or impaired (nZ23) motor function. 2.4. Surgical procedure An electrode (Resume) was placed over the dura through a frontoparietal craniotomy (40!50 mm) made over the central area and under general anaesthesia. This technique was preferred to the alternative of a simple burr hole under local anaesthesia in order both to minimize the risk of epidural haematoma and to facilitate the use of electrophysiological techniques to localize the central sulcus. The motor area was localized with the SEP phase reversal technique (Wood et al., 1988). Somatosensory evoked potentials (N20/P20) were obtained in 28/31 (90.3%) patients. They could not be obtained in three patients because of either complete lemniscal deafferentation (two cases (6.5%) with brachial plexus avulsion (no 2, 20)) or technical failure (one patient (3.2%) (no 6)). Concomitant with SEPs, optimised guided surgery was performed by MRI ‘neuronavigation’ (Stealth Station, Medtronic Sofamor Danekw) in 10 patients (no 6, 10, 11, 12, 13, 15, 19, 23, 27, 31). One or two quadripolar electrodes (Resume Medtronicw in all but 1 case (Symix Medtronicw)) were implanted over the motor

46

C. Nuti et al. / Pain 118 (2005) 43–52

representation of the painful area: the suprasylvian region of the convexity was stimulated if the pain was located in the face and/or the upper limb (electrode immediately anterior and parallel to the central sulcus), while the paramedian region was stimulated if the pain was in the lower limb (parasagital anteroposterior electrode). All electrodes but 2 (patients no 26, 30) were implanted extradurally. In these two cases, one electrode was introduced sub-durally in the interhemispheric fissure over the medial part of the precentral gyrus to treat pain in the lower limb. Only one electrode was needed in 13 patients (no 1, 2, 4, 5, 7, 9, 20, 21, 24, 25, 28, 29, 31) because their pain was restricted to a limited area. Two electrodes were needed for 18 patients, because of the extent of the pain, including both the upper and lower limb and the face. In these cases, the aim was to stimulate the whole motor cortex to include somatotopic representations of the painful areas. The electrode was connected to a subcutaneously implanted stimulator (Medtronicw): radiofrequency Xtrel (three patients), Matrix (six patients), Itrel II (17 patients), or Synergy (five patients). The whole stimulation device including the pulse generator was implanted in a single session in all the patients in our series. During follow-up, the stimulating device was changed in nine patients: for battery failure in seven cases and for internalisation of the radiofrequency device (Matrix) in two patients. The stimulation parameters were adapted empirically in the postoperative stage with the aim of optimising the analgesic effect. The bipolar stimulation intensity was initially very low and was gradually increased over one week in order to prevent adverse effects. After adaptation, the intensity ranged from 0.5 to 5 V (mean 1.5) and was always maintained under the motor threshold. The frequency ranged from 30 to 80 Hz (mean 45.5), the pulse width from 60 to 330 ms (mean 140). The stimulation was set on a cyclic mode in all patients (‘On’ periods: from 30 to 120 min; ‘Off’ periods: from 15 min to 24 h).

by clinical visits at a mean interval of six months, for a period that reached 104 months in the patient with the longest follow-up (mean: four years). At the end of the patients’ follow-up, the percentages of pain relief were averaged over time for each patient to obtain an individual ‘global score of pain relief’. Patients were then separated into four groups according to their mean individual percentage of pain relief, as follows: S70% (excellent result, group 1), 40–69% (good result, group 2), 10–39% (poor results, group 3) and 0–9% (failure, group 4). Long-term outcome was evaluated using these same classifications. To reach the second goal of the study, the following parameters were tested by parametric statistics as possible prognostic factors of MCS efficacy: (i) presence/absence of allodynia/hyperpathia, (ii) motor status (normal/impaired), (iii) SEPs, QST and LEPs data (2 conditions: normal/impaired), (iv) anatomic localization (2 conditions; ‘brain’/‘spine or brachial plexus’)), (v) the type of stroke (Ischemic/Haemorrhagic), (vi) the interval between pain and surgery. These parameters were compared to the second postoperative %R (obtained at the end of the first month of followup), the mean %R, the mean VAS, the postoperative percentage decrease in VAS, the postoperative drug intake (2 conditions: identical/modified (decreased or withdrawn)) and the ‘yes/no’ response. (vii) The second %R, obtained at the end of the first postoperative month, was compared to the mean %R (to test whether this first evaluation could be predictive of the follow-up period). The tests (unpaired t-test, paired t-test, association tests (contingency table-summary data: Chi-2), regression test) used to compare variables are described in Table 3.

3. Results 3.1. Overall results

2.5. Assessment of pain and postoperative outcome Preoperative pain was evaluated using Visual Analog Scaling (VAS, ranging between 0: no pain and 10: maximal pain). Before surgery, VAS ranged from 6 to 10 (mean 8.5). Postoperative outcome was evaluated by using five variables of interest, namely (i) the rate (percentage) of pain relief (%R) scored between 0% (no pain relief) and 100% (complete pain relief), (ii) the pain score as assessed on VAS, (iii) the percentage decrease in the postoperative VAS, (iv) the level of analgesic drug intake compared to the preoperative stage (3 conditions: identical, decreased, withdrawn), (v) the yes/no response to the question of whether the patient would accept, assuming similar efficacy levels, to be operated on again. 2.6. Evaluation A first systematic assessment of the postoperative outcome was made in the first 10 days (D10). A second evaluation was made at the end of the first postoperative month. This latter evaluation (instead of the D10 evaluation) was used for the statistical comparisons to exclude from the analysis the period of adaptation of the stimulation parameters and the confounding effects of surgical wound pain. Then, evaluations were made monthly for a period of four months. The postoperative outcome was evaluated

The overall results are presented in Table 2a. Three patients (9.7%) belonged to group 1 (excellent results), 13 patients (42%) to group 2, 11 patients (35.4%) to group 3 and four patients (12.9%) were in the failure group. For each group, a graph representing the evolution of the percentage of pain relief over time is presented in Fig. 1. Medical treatment could be withdrawn altogether at the end of the follow-up in 11 patients (35.5%) (two in group 1, five in group 2, four in group 3), was decreased over time in five (16.1%) (two in group 2, three in group 3) and remained stable in 14 patients (45.2%) (six in group 2, four in group 3, four in group 4). Longitudinal data were not available for one patient. To the question ‘would you accept surgery again?’, the answer was positive in 21 cases (70%) (two in group 1, 11 in group 2, eight in group 3) and negative in nine (30%) (two in group 2, three in group 3, four in group 4). There was no response for one patient. The parameters %R, medical treatment and question: ‘would you accept surgery again?’ were consistent and showed no discrepancy in groups 1, 2 and 4 (Table 2b). Conversely, 10 of the 11 patients in group 3 had

C. Nuti et al. / Pain 118 (2005) 43–52

contradictory responses. Despite a percentage of pain relief (10–39%) theoretically in favour of a poor result, a postoperative decrease/withdrawal of analgesic drugs intake and/or a positive answer to the question about iterative surgery were observed in these 10 patients, indicating a possible clinical benefit. In these patients, the evaluation of MCS effect may therefore be underestimated by the %R criteria. In the present series, the motor cortex stimulation did not induce any subjective sensation in our patients, contrary to other studies (Katayama et al., 1998; Tsubokawa et al., 1993). This was mentioned in the early papers reporting use

47

of the technique but was not confirmed in later studies (Nguyen et al., 1999). Side effects were found in eight patients (26%). Neither of these led to permanent dysfunction, nor to removal of the stimulation device. Partial epileptic seizures occurred in three patients in the early postoperative stage, or during trials to increase stimulation intensity. They rapidly stopped without treatment when the stimulation intensity was decreased. Two transient episodes of postoperative dysfunction (one speech disorder and one motor deficit) were also reported, and resolved spontaneously. Finally, two delays in surgical wound healing and one local sepsis were observed.

Table 2a Patients postoperative data

Mean %R

Mean VAS

Double choice test Decrease in postoperative VAS percentage

Medical analgesic treatment

1

30

%R of the end of the first postoperative month 25

6.4

11

Y

=

2

60

60

4.4

46

N

=

3

59

60

7

30

Y

4

0

0

9

10

N

↓↓ =

5

57

100

3.4

56

Y

↓

6

41

80

2.6

44

Y

=

7

41

100

7.1

9

Y

=

8

33

0

6.1

9

Y

↓↓

N˚

9

24

60

7.3

17

N

10

79

75

0.6

74

Y

↓↓ ↓↓

11

52

30

4.2

38

Y

↓↓

12

14

20

6.6

30

Y

=

13

25

20

5.6

24

Y

=

14

30

50

6.7

33

Y

↓

15

20

5

5.4

26

N

16

41

30

2.8

52

Y

↓ ↓↓

17

43

45

4.5

35

Y

↓↓

18

36

40

6.1

14

Y

↓

19

48

40

5

35

N

=

20

40

40

7

10

Y

↓

21

93

100

3.5

35

n.a.

n.a.

22

0

0

9.1

9

N

=

23

73

50

2.8

62

Y

↓↓

24

45

50

5.3

22

Y

25

40

40

5

40

Y

= ↓↓

26

15

20

5.7

23

N

=

27

15

30

7.6

4

Y

↓↓

28

30

5

6.8

17

Y

↓↓

29

44

80

5.1

44

Y

=

30

0

0

10

0

N

=

31

0

0

9.3

0

N

=

0

20

40

60

80

100

Abbreviations: % R, percentage of pain relief; Mean %R: four columns (indicated in grey) show the four sub-groups of subjects, from left to right, group 4 (0–9%), group 3 (10–39%), group 2 (40–69%), group 1 (70–100%); VAS, Visual Analog Scale; Double choice test: ‘would you accept surgery again?’, Y, yes; N: no; Medical analgesic treatment: Z, unchanged; Y, decrease; k, withdrawal; n.a., not available.

48

C. Nuti et al. / Pain 118 (2005) 43–52

100

Gp 1: n = 3

80 70 Gp 2: n = 13

Gp 1 Gp 2 Gp 3 Gp 4

%R

60

40

Gp 3: n = 11 20 10 Gp 4: n = 4

0

1

3

5

7

9

11

13

15

17

19

21

23

25

27

29

31

33

36

37

39

Follow-up (months) Fig. 1. Patients’ postoperative percentage of pain relief (%R). (Gp 1: 100–70%; Gp 2: 69–40%; Gp 3: 39–10%; Gp 4: 9–0%).

In the four patients with peripheral nerve lesions associated with central pain (no 11, 12, 18, 24), only the central component of the pain was modified by MCS. During follow-up, eight patients asked for their stimulators to be checked due to loss of the analgesic effect (no 6, 8, 10, 11, 16, 20, 24, 29). In all cases, a technical failure of the stimulating device or battery emptying was evidenced, and repair or battery replacement returned the analgesic effect to former levels. 3.2. Predictive factors The comparison between the second %R (end of the first month of follow-up) and the mean value on the total follow-up did not show any statistical difference (paired t-test). This argues for a strong correlation between these two elements, and is confirmed by the regression test (RZ 0.744; P!0.0001; Tables 3 and 4). There was a tendency for allodynic patients to decrease or withdraw their post-operative analgesic drugs more

frequently than other patients. There was a similar tendency for non-hyperalgesic patients to give better scores immediately after surgery. The patients with ischemic stroke tended towards greater percentage decreases of VAS and reduced medication than patients with haemorrhagic stroke. Finally, patients with normal motor scores tended to decrease or withdraw their post-operative analgesic drugs more frequently than did patients with motor deficits. However, these tendencies did not reach the significance threshold (Tables 3 and 4). For all the other categories, the different statistical tests did not show any variables that approached a significant level (Table 3). 3.3. Discussion Good or excellent overall results with MCS were obtained in 52% of our patients (groups 1 and 2), in agreement with previously published rates of success

Table 2b Decrease/withdrawal in postoperative medical treatment and positive answer to the question about iterative surgery according to the rate of pain relief (%)

Group 1 (70–100%R) nZ3 Group 2 (40–69%R) nZ13 Group 3 (10–39%) nZ11 Group 4 (0–9%) nZ4 a

Decrease/withdrawal in postoperative medical treatment

‘Yes’ answer to the question on iterative surgery

Total number of patients with at least one positive variable

2 8 7 0

2 12 8 0

2/3 12/13 a 10/11 0/4

Note the contradictory responses in group 3.

C. Nuti et al. / Pain 118 (2005) 43–52

ranging from 45 to 77% of patients (Mertens et al., 1999; Nguyen et al., 1999; Tsubokawa et al., 1991; 1993). Our series did not include patients with trigeminal neuropathic pain, in whom MCS has been shown to give the highest rate of pain relief with satisfactory reported results in up to 100% of patients (Ebel et al., 1996; Meyerson et al., 1993; Nguyen et al., 1997; 1999; Rainov and Heidecke, 2003). This latter bias in patients’ selection, associated with the patients’ heterogeneous characteristics (the inclusion of nine patients with pain unrelated to stroke), could explain why our overall results are at the lower end of the success scale compared to other reports. Compared to the rate of success of pharmacological therapy (about 30% in central pain) (McQuay et al., 1996; Sindrup and Jensen, 1999), the complete failure of MCS in only four patients (12.9%) is also an encouraging result in favour of the efficacy of MCS in this context of refractory pain. Taking into account the severity of the pain in the patients who were included in the study, its resistance to well-conducted medication and the paucity of side effects from MCS in these patients, it is suggested that this therapy can be proposed as a safe and useful therapeutic alternative, provided that the refractoriness to pharmacological therapy is established by well-conducted drug trials. Methods for evaluating clinical benefit may also to some extent explain why the success rates with MCS differ from one investigator to another. The evaluation of

49

therapies using a single criterion, such as VAS ratings or pain relief percentage may underestimate their relevance (Monhemius and Simpson, 2003), as illustrated in the group 3 patients. In these patients, there are patent discrepancies among the five variables of interest that were considered here. Despite a moderate percentage of declared pain relief (10–39%), the decrease in postoperative drug intake (or drug withdrawal) and/or the positive answer to the question on iterative surgery in 10 out of 11 patients confirms that MCS may have a beneficial effect on the patients. Accordingly, more than 50% of patients in the whole series experienced sufficient pain relief to decrease their drug intake (complete drug withdrawal in 35.5% of cases) and almost 70% would accept surgery again if the same pain relief was guaranteed. These data highlight the need to use a series of objective variables in addition to subjective scores, to complete the evaluation of the technique. Despite objective and subjective data in favour of a prolonged analgesic effect of MCS, the question of a placebo effect remains theoretically possible. The following arguments, however, argue against that interpretation. Firstly, as illustrated in Fig. 1, the effect of MCS remained stable in the improved groups over a 4-year mean follow-up. Even though the placebo effect may be prolonged, such a long period of efficacy would rather suggest other processes. Secondly, the request to check that the stimulator was functioning because of a

Table 3 Statistical tests used to determine the potential predictive factors for the analgesic effect of MCS Explicative variables

Output variables Mean %R

Second postoperative %R (end of the first month)

Mean VAS

Postoperative % decrease of VAS

Postoperative drug intake: identical or modified (decreased or withdrawn)

ns

ns

ns

*

Unpaired t-test Allodynia:CversusK

ns

‘Yes/No’ response

Chi-2 ns

Hyperpathia:CversusK

ns

*

ns

ns

ns

ns

Motor status: N versus Im

ns

ns

ns

ns

*

ns

SEPs: N versus Im

ns

ns

ns

ns

ns

ns

ST (QST/LEPs): N versus Im

ns

ns

ns

ns

ns

ns

‘brain’ versus ‘spine or brachial plexus’

ns

ns

ns

ns

ns

ns

Isch versus Haem

ns

ns

ns

*

*

ns

Interval between pain and surgery

Regression test ns

ns

Unpaired t-test ns

ns

Second postoperative %R (end of the first month)

Paired t-test: ns regression test: (RZ0.744; P! 0.0001)

ns

ns

The first column variables are tested as potential predictive factors for the analgesic effect of MCS; the first line variables are output data. Abbreviations: %R, percentage of relief obtained; VAS, visual analog scale; C, presence; -, absence; SEP, somatosensory evoked potentials; ST, spino-thalamic (QST, quantitative sensory testing and LEPs, laser-evoked potentials); N, normal; Im, impaired; Isch, ischemic stroke; Haem, haemorrhagic stroke; ns, non significant. *: indicates results demonstrating a tendency toward significance (see Table 4).

50

C. Nuti et al. / Pain 118 (2005) 43–52

Table 4 Predictive factors

Abbreviation: %R, percentage of pain relief; VAS, visual analog scale; postop, postoperative; Isch, ischemic stroke; Haem, haemorrhagic stroke; Trt, treatment; A, allodyniae; trt Z, unchanged postoperative medical treatment; trt Y, decrease or withdrawal of postoperative medical treatment; Motor N, normal motor status; Motor aN, abnormal motor status.

loss of the analgesic effect in eight patients ignorant of such a MCS dysfunction (no subjective sensation of MCS was induced in our patients), with a recovery of a similar analgesic effect after repairing the device, is in favour of a real analgesic effect and not a placebo. Thirdly, if we consider that the effect of MCS may be related to a placebo response, there is no objective reason for a selective effect on the central pain component with an absence of any effect on the peripheral one, as observed in the four patients with such an association of two abnormal pain processes. One of the main contributions of our series is the length of the follow-up, which makes our study, to our knowledge, the longest prospective report on patients with MCS. Despite the absence of correlation with clinical and paraclinical data, the stability of the results after a mean follow-up of 4 years suggests that much of the analgesic effect of the procedure might be predicted from the first month. Indeed, despite a large and systematic preoperative

exploration, including a clinical and neurophysiological assessment, we failed to identify any other parameters predictive of long-term efficacy. Neither the preoperative motor status, nor pain semiology, nor the type/localization of the lesion, nor the objective somato-sensory dysfunction predicted patient’s outcome after MCS. Even though the distribution of the patients in each group was homogenous, the small size of the different subgroups tested in this study might explain these negative results. Knowledge of the neurophysiological mechanisms at the basis of the analgesic effects of MCS would clearly help improve the selection of those patients who could benefit from surgery. Pharmacological tests have been performed with this aim and it has been proposed that only patients who respond to thiamylal and ketamine analgesia (but not to morphine) would benefit from MCS (Yamamoto et al., 1997), but these results have not been duplicated (Saitoh et al., 2000). Absence of a severe lesion on the corticospinal tracts was also proposed as a positive factor for prognosis

C. Nuti et al. / Pain 118 (2005) 43–52

(Katayama et al., 1998). Even though this finding was not statistically replicated in this study, the tendency towards the decrease of analgesic drugs in patients with normal motor status is in agreement with the finding by Katayama et al. (1998) that a limited motor deficit could be a favourable predictive factor for the analgesic effect of MCS. The preoperative results of repetitive transcranial magnetic stimulation (rTMS) have also been proposed as a predictive test for MCS (Lefaucheur et al., 2001; Migita et al., 1995) but double-blind studies are still needed to formally pronounce on the interest of a preoperative rTMS test. Finally, the absence of a severe lesion on non-nociceptive sensory modalities as assessed by QST have also been proposed as a positive predictive value for the success of MCS (Drouot et al., 2002). However, in definite subgroups of patients, QST was not found to predict the effect of MCS by rTMS (Lefaucheur et al., 2004), as verified in our present population. Our results are also, from this point of view, consistent with those from Katayama et al. (1998) who reported no correlation between sensory symptoms including allodynia and hyperpathia, SEPs and the MCS effect. Taken together, these results suggest that (spino-thalamic and lemniscal) sensory abnormalities may not be predictive of the efficacy of MCS. Differences in the clinical aspects of neuropathic pain have not been shown to influence clinical outcome, including allodynia, which did not significantly influence the results of MCS. There was however a tendency for allodynic patients to decrease or withdraw their postoperative drug intake. We cannot decide whether the decrease in postoperative treatment in allodynic patients was related to the pharmacoresistance of this symptom. This is because most allodynic patients with postoperative treatment decrease or withdrawal belonged to groups 1 or 2 (i.e. to the groups with the better global effects of MCS). Retrospectively, it is impossible to know the relative influence of global versus specific antiallodynic effects of MCS in these patients. Our iterative clinical analysis at every visit could not determine the respective effect of stimulation on both components, as most patients were unable to describe it in a clear and reliable form. More than ten years after the description of the technique by Tsubokawa et al. (1991), no clear evidence for a preoperative selection of responders can be advanced. Both larger samples of patients and better understanding of the MCS mechanisms could help in patient selection, with the objective of improving the clinical results.

Acknowledgements This work was supported in part by Projet Hospitalier de Recherche Clinique (PHRC, tenders 1996–1999).

51

References Brown JA, Barbaro NM. Motor cortex stimulation for central and neuropathic pain: current status. Pain 2003;104(3):431–5. Drouot X, Nguyen JP, Peschanski M, Lefaucheur JP. The antalgic efficacy of chronic motor cortex stimulation is related to sensory changes in the painful zone. Brain 2002;125:1660–4. Ebel H, Rust D, Tronnier V, Bo¨ker D, Kunze S. Chronic precentral stimulation in trigeminal neuropathic pain. Acta Neurochir 1996;138: 1300–6. Garcia-Larrea L, Peyron R, Mertens P, Gre´goire MC, Lavenne F, Le Bars D, et al. Electrical stimulation of motor cortex for pain control: a combined PET- scan and electrophysiological study. Pain 1999;83: 259–73. Garcia-Larrea L, Convers P, Magnin M, Andre´-Obadia N, Peyron R, Laurent B, et al. Laser-evoked potential abnormalities in central pain patients: the influence of spontaneous and provoked pain. Brain 2002; 125:2766–81. Hanajima R, Ashby P, Lang AE, Lozano AM. Effects of acute stimulation through contacts placed on the motor cortex for chronic stimulation. Clin Neurophysiol 2002;113(5):635–41. Katayama Y, Kukaya C, Yamamoto T. Poststroke pain control by chronic motor cortex stimulation: neurological characteristics predicting a favorable response. J Neurosurg 1998;89:585–91. Lefaucheur JP, Drouot X, Nguyen JP. Interventional neurophysiology for pain control: duration of pain relief following repetitive transcranial magnetic stimulation of the motor cortex. Neurophysiol Clin 2001; 31(4):247–52. Lefaucheur JP, Drouot X, Menard-Lefaucheur I, Zerah F, Bendib B, Cesaro P, et al. Neurogenic pain relief by repetitive transcranial magnetic stimulation depends on the origin and the site of pain. J Neurol Neurosurg Psychiatry 2004;75:612–6. McQuay HJ, Trame`r M, Nye BA, Caroll D, Wiffen PJ, Moore RA. A systematic review of antidepressants in neuropathic pain. Pain 1996;68: 217–27. Mertens P, Nuti C, Sindou M, Guenot M, Peyron R, Garcia-Larrea L, et al. Precentral cortex stimulation for the treatment of central neuropathic pain. Stereotact Funct Neurosurg 1999;73:122–5. Meyerson BA, Lindblom U, Linderoth B, Lind G, Herregodts P. Motor cortex stimulation as treatment of trigeminal neuropathic pain. Acta Neurochir 1993;(Suppl. 58):150–3. Migita K, Uozumi T, Arita K, Monden S. Transcranial magnetic coil stimulation of motor cortex in patients with central pain technique and application. Neurosurgery 1995;36:1037–40. Monhemius R, Simpson BA. Efficacy of spinal cord stimulation for neuropathic pain: assesment by abstinence. Eur J Pain 2003;7:513–9. Nguyen JP, Keravel Y, Feve A, Uchiyama T, Cesaro P, Le Guerinel C, et al. Treatment of deafferentation pain by chronic stimulation of the motor cortex: report of a series of 20 cases. Acta Neurochirurgica 1997; (Suppl. 68):54–60. Nguyen JP, Lefaucheur JP, Decq P, Uchiyama T, Carpentier AC, Fontaine D, et al. Chronic motor cortex stimulation in the treatment of central and neuropathic pain. Correlations between clinical, electrophysiological and anatomical data. Pain 1999;82:245–51. Peyron R, Garcia-Larrea L, Deiber MP, Cinotti L, Convers P, Sindou M, et al. Electrical stimulation of precentral cortical area in the treatment of central pain: electrophysiological and PET study. Pain 1995;62: 275–86. Rainov NG, Heidecke V. Motor cortex stimulation for neuropathic facial pain. Neurol Res 2003;25:157–61. Saitoh Y, Shibata M, Hirano S-I, Hirata M, Mashimo T, Yoshimine T. Motor cortex stimulation for central and peripheral deafferentation pain. Report of eight cases. J Neurosurg 2000;92: 150–5.

52

C. Nuti et al. / Pain 118 (2005) 43–52

Sindrup SH, Jensen TS. Efficacy of pharmacological treatments of neuropathic pain: update and effect related to mechanism of drug action. Pain 1999;83:389–400. Tsubokawa T, Katayama Y, Yamamoto T, Hirayama T, Koyama S. Chronic motor cortex stimulation for the treatment of central pain. Acta Neurochir 1991;(Suppl. 52):137–9. Tsubokawa T, Katayama Y, Yamamoto T, Hirayama T, Koyama S. Chronic motor cortex stimulation in patients with thalamic pain. J Neurosurg 1993;78:393–401.

Wood CC, Spencer DD, Allison T, McCarthy G, Williamson PD, Goff WR. Localization of human sensorimotor cortex during surgery by cortical surface recording of somatosensory evoked potentials. J Neurosurg 1988;68:99–111. Yamamoto T, Katayama Y, Hirayama T, Tsubokawa T. Pharmacological classification of central post-stroke pain: comparison with the results of chronic motor cortex stimulation therapy. Pain 1997;72: 5–12.