Unilateral repetitive tibial nerve stimulation improves neurogenic claudication and bilateral F-wave conduction in central lumbar spinal stenosis

Unilateral repetitive tibial nerve stimulation improves neurogenic claudication and bilateral F-wave conduction in central lumbar spinal stenosis

Journal of Orthopaedic Science xxx (2017) 1e7 Contents lists available at ScienceDirect Journal of Orthopaedic Science journal homepage: http://www...

843KB Sizes 1 Downloads 15 Views

Journal of Orthopaedic Science xxx (2017) 1e7

Contents lists available at ScienceDirect

Journal of Orthopaedic Science journal homepage: http://www.elsevier.com/locate/jos

Original Article

Unilateral repetitive tibial nerve stimulation improves neurogenic claudication and bilateral F-wave conduction in central lumbar spinal stenosis Noritsuna Nakajima a, *, Toshikazu Tani b, Katsuhito Kiyasu a, Masashi Kumon a, Shinichirou Taniguchi c, Ryuichi Takemasa a, Nobuaki Tadokoro a, Kazuya Nishida d, Masahiko Ikeuchi a a

Department of Orthopaedic Surgery, Kochi Medical School, Kochi University, Japan Department of Orthopaedic Surgery, Kubokawa Hospital, Japan c Department of Orthopaedic Surgery, Kansai Medical University, Takii Hospital, Japan d Department of Orthopaedic Surgery, Chikamori Hospital, Japan b

a r t i c l e i n f o

a b s t r a c t

Article history: Received 25 July 2017 Received in revised form 20 October 2017 Accepted 13 December 2017 Available online xxx

Background: Repetitive electrical nerve stimulation of the lower limb may improve neurogenic claudication in patients with lumbar spinal stenosis (LSS) as originally described by Tamaki et al. We tested if this neuromodulation technique affects the F-wave conduction on both sides to explore the underlying physiologic mechanisms. Methods: We studied a total of 26 LSS patients, assigning 16 to a study group receiving repetitive tibial nerve stimulation at the ankle (RTNS) on one leg, and 10 to a group without RTNS. RTNS conditioning consisted of a 0.3-ms duration square-wave pulse with an intensity 20% above the motor threshold, delivered at a rate of 5 Hz for 5 min. All patients underwent the walking test and the F-wave and M-wave studies for the tibial nerve on both sides twice; once as the baseline, and once after either the 5-min RTNS or 5-min rest. Results: Compared to the baselines, a 5-min RTNS increased claudication distance (176 ± 96 m vs 329 ± 133 m; p ¼ 0.0004) and slightly but significantly shortened F-wave minimal onset latency (i.e., increased F-wave conduction velocity) not only on the side receiving RTNS (50.7 ± 4.0 ms vs 49.2 ± 4.2 ms; p ¼ 0.00081) but also on the contralateral side (50.1 ± 4.6 ms vs 47.9 ± 4.2 ms; p ¼ 0.011). A 5-min rest in the group not receiving RTNS neither had a significant change on claudication distance nor on any F-wave measurements. The M response remained unchanged in both groups. Conclusions: The present study verified a beneficial effect of unilaterally applied RTNS of a mild intensity on neurogenic claudication and bilateral F-wave conduction. Our F-wave data suggest that this type of neuromodulation could be best explained by an RTNS-induced widespread sympathetic tone reduction with vasodilation, which partially counters a walking-induced further decline in nerve blood flow in LSS patients who already have ischemic cauda equina. © 2017 The Japanese Orthopaedic Association. Published by Elsevier B.V. All rights reserved.

1. Introduction Repetitive electrical nerve stimulation of the lower limb provides a beneficial effect on neurogenic claudication in patients with lumbar spinal stenosis (LSS). This type of neuromodulation has

* Corresponding author. Department of Orthopaedic Surgery, Kochi Medical School, Kochi University, 185-1 Oko-cho Kohasu, Nankoku 783-8505, Japan. Fax: þ81 88 880 2388. E-mail address: [email protected] (N. Nakajima).

received little attention since originally described by Tamaki et al. (1986) [1] and confirmed by Takahashi et al. (1988) [2], despite its potential as a useful treatment option for LSS. More recently (2014) [3], one study reconfirmed this stimulation-induced effect, showing in a series of LSS patients that repetitive tibial nerve stimulation at the ankle (RTNS) at the mild intensity of 5/s for 5 min doubled their maximal walking distances, at which point they could no longer continue walking due to an aggravation in leg symptoms (i.e., absolute claudication distance) [4]. The same study also revealed that conduction velocity of the tibial nerve F-wave significantly increased on the side receiving

https://doi.org/10.1016/j.jos.2017.12.006 0949-2658/© 2017 The Japanese Orthopaedic Association. Published by Elsevier B.V. All rights reserved.

Please cite this article in press as: Nakajima N, et al., Unilateral repetitive tibial nerve stimulation improves neurogenic claudication and bilateral F-wave conduction in central lumbar spinal stenosis, Journal of Orthopaedic Science (2017), https://doi.org/10.1016/j.jos.2017.12.006

2

N. Nakajima et al. / Journal of Orthopaedic Science xxx (2017) 1e7

RTNS in LSS patients, but not in age- and sex-matched healthy subjects, suggesting an RTNS-induced improvement in motor nerve conduction in already impaired LSS patients. To further clarify this neuromodulation, in the present exploratory study, we investigated whether the effects of RTNS applied to one leg in LSS patients occurred specifically in the stimulated nerve or more extensively in other unstimulated nerves, by conducting the electrophysiological study on both sides. Additionally, we compared the data for the two groups of central LSS patients; RTNS(þ) group and RTNS() group. We also sought to determine whether RTNS affects the proximal motor segment, the distal motor segment, or both, by comparing the latency changes of F-wave vs M-wave induced by RTNS. 2. Material and methods 2.1. Patients For our study, we selected a total of 26 LSS patients out of those who had been consecutively referred to Kochi Medical School for decompression surgery between October 2016 and March 2017. Our inclusion criteria comprised the following: (1) the presence of neurogenic claudication; (2) a claudication distance of less than 500 m; (3) radiologically-confirmed central LSS affecting both legs without radicular symptoms caused by lateral recess compression of an individual nerve root; and (4) electrophysiologicallyconfirmed M responses elicitable from the abductor hallucis (AH) after tibial nerve stimulation at the ankle. We assigned the 26 patients into two groups in the order of their presentation at the hospital. The first group consisted of 16 patients (10 men, 6 women), were aged between 59 and 90 years (mean, 72.6 years) with a mean height of 157.9 cm, and underwent RTNS in the supine position (RTNS(þ) group) for assessing the primary endpoint of this exploratory study, i.e., whether unilateral RTNS affects the F-wave on both sides; while the following 10 patients (5 men, 5 women) were aged between 67 and 85 years (mean, 75.3 years) with a mean height of 154.0 cm, and simply lay supine during the same period of time (RTNS() group). MRI showed the most intense compression of the thecal sac at L3-4 in 2 patients and at L4-5 in 14 in the RTNS(þ) group, and at L3-4 in 1 patient and at L4-5 in 9 in the RTNS() group (Table 1). All agreed in writing to participate in the study after reading an informed consent form approved by the hospital ethics committee. 2.2. Sequence of study (Fig. 1) All patients first underwent a walking test. The patients then sat comfortably on an adjustable armchair for 30 min to obtain complete relief from their claudication symptoms before the baseline Fwave and M-wave studies. F-waves and M-waves were recorded from the AH on both sides after tibial nerve stimulation at the ankle and at the knee, respectively, with the patients lying supine on an

Table 1 Demographic data.

Sex (M/F) Age (mean ± SD [years]) Height (mean ± SD [cm]) Level most intensely compressed (L3-4/L4-5) a b

RTNS (þ) group (n ¼ 16)

RTNS () group (n ¼ 10)

P value

10/6 72.6 ± 7.8 157.9 ± 10.6 2/14

5/5 75.3 ± 6.6 154.0 ± 6.8 1/9

0.826a 0.379b 0.304b 1a

Calculated according to 2  2 Chi square test. Calculated according to unpaired t-test.

examination table. Next, the RTNS(þ) group had RTNS conditioning on one leg, a 0.3-ms duration square-wave pulse with an intensity 20% above the motor threshold, applied to the tibial nerve at the ankle at 5 Hz for 5 min with the same stimulating electrodes as used in the F-wave study. The RTNS() group simply continued lying supine without RTNS during the same period of 5 min. Subsequently, both groups underwent the second F-wave and M-wave studies on both sides and finally, the second walking study. 2.3. Walking test LSS patients walked around a circular course of 40 m on a smooth flat surface in a marked corridor. They walked at a comfortable and reasonable speed as instructed and supervised by one of the authors (NN). We measured the claudication distance [4], at which point the patients could no longer continue walking due to worsening leg symptoms, using a distance measuring wheel with a digital display of 0.1 m resolution (M474WS-160BS, Shiro Industry Co., Osaka, Japan). The walking test was performed twice; once as the baseline measurement for both groups, and once either after the 5-min RTNS or 5-min rest without RTNS. 2.4. Electrophysiologic studies The subject lay supine on an examination table in a quiet, warm room maintained at a temperature between 21 and 23  C. The Fwave study consisted of stimulating the tibial nerve at the ankle with a bipolar stimulator (RY-441B, Nihon Kohden, Tokyo, Japan) and recording the late response from the AH with a pair of disc electrodes (NE-142B, Nihon Kohden, Tokyo, Japan) placed over the belly and the tendon of the AH after application of an abrasive gel to reduce skin impedance. Electrical stimulation of a square wave, 0.2 ms in duration and of supramaximal intensity adjusted to a value 20% stronger than that required for eliciting the maximal M response, was delivered 100 times at 1 Hz. We used a NihonKohden Neuropack (Neuropack MEB2200, Nihon Kohden, Tokyo, Japan) with the band-pass filter set at 20 Hz and 3 KHz for acquisition of F-waves, which were defined as a deflection of at least 50 mV measured from peak to peak [5]. F-wave measurements comprised: (1) persistence, i.e., percentage of stimuli that produced F-waves; (2) peak-to-peak amplitude averaged for only those trials with detectable F responses (i.e., response average) [6]; (3) minimal onset latency; (4) mean onset latency, i.e., the average of all F-wave onset latencies; (5) chronodispersion, i.e., the difference between the minimal and maximal F-wave latencies [7]; and (6) F-wave conduction velocity (FWCV) calculated as: FWCV (m/s) ¼ 2D (mm)/F (ms)  M (ms)  1.0 (ms) [8], where “D” represents the surface distance measured from the stimulus site to the T12 spinous process by way of the knee and greater trochanter of the femur, and “F” and “M” are the respective latencies of the F-wave and M-wave. The F-wave latencies were measured twice by 1 investigator with an interval of more than one month. The levels of intra-rater reliability of the measurements were assessed using the intraclass correlation coefficient (ICC). We also conducted a motor nerve conduction study below the knee, measuring the latency and the baseline to peak amplitude of the M response recorded from the AH after supramaximal stimulation of the tibial nerve at the knee. 2.5. Statistical analysis We decided on a sample size of 16 patients, based on a previous study of Kumon et al. [3], for assessing the primary endpoint of the current exploratory experiment, although this type of experiment

Please cite this article in press as: Nakajima N, et al., Unilateral repetitive tibial nerve stimulation improves neurogenic claudication and bilateral F-wave conduction in central lumbar spinal stenosis, Journal of Orthopaedic Science (2017), https://doi.org/10.1016/j.jos.2017.12.006

N. Nakajima et al. / Journal of Orthopaedic Science xxx (2017) 1e7

3

Baseline walking study 30-minute rest RTNS(+) group

RTNS(-) group

Baseline F-wave study on both sides

Baseline F-wave study on both sides

Baseline M-wave study on both sides

Baseline M-wave study on both sides

5-minute RTNS on one side

5-minute rest

Test F-wave study on both sides

Test F-wave study on both sides

Test M-wave study on both sides

Test M-wave study on both sides

Test walking study

Test walking study

Fig. 1. Simplified diagram illustrating a sequence of studies, divided into the two major categories of those with and without repetitive tibial nerve stimulation (RTNS). F-wave and M-wave studies were conducted by stimulating the tibial nerve at the ankle and at the knee, respectively.

generally presents a difficulty in estimating the exact sample size. We used SPSS software, version 18 (SPSS Inc., Chicago, Illinois) for all statistical analyses. Categorical data were compared using the 2  2 Chi square test. For continuous variables, we assessed the normality of distribution first using the ShapiroeWilk test. For normally distributed data, 2-group comparisons were performed with the unpaired t-test. For non-normal distributions, we used the Kruskal Wallis H-test for 4-group comparisons and Wilcoxon signed-rank test for 2-group comparisons. Values are given as mean ± SD and the significance level was set at p < 0.05. 3. Results The two groups showed no significant differences in age (p ¼ 0.379), sex (p ¼ 0.826), height (p ¼ 0.304) or level of the most intense thecal sac compression (p ¼ 1) (Table 1). 3.1. Walking test A comparison of the baseline claudication distances showed no significant difference between the RTNS(þ) and RTNS() groups (176 ± 96 m vs 139 ± 110 m; p ¼ 0.341). The claudication distances in the RTNS(þ) group significantly increased (p ¼ 0.0004) from the baseline value of 176 ± 96 m to 329 ± 133 m after 5-min RTNS, whereas those in the RTNS() group showed no significant changes (p ¼ 0.14) from the baseline value of 139 ± 110 m to 151 ± 112 m after a 5-min rest without RTNS (Table 2).

3.2. F-wave study The 4 sets of baseline F-waves recorded from both groups bilaterally showed no significant differences in persistence, amplitude, mean onset latency, minimal onset latency, chronodispersion and conduction velocity (Table 3). In the RTNS(þ) group, the F-wave values recorded on the side receiving RTNS significantly changed from the baselines after a 5min RTNS; both mean onset latency (54.1 ± 4.1 ms to 53.9 ± 4.3 ms; p ¼ 0.026) and minimal onset latency (50.7 ± 4.0 ms to 49.2 ± 4.2 ms; p ¼ 0.00081) shortened, and F-wave conduction velocity increased (49.7 ± 4.2 m/s to 51.6 ± 4.4 m/s; p ¼ 0.00081). No other measures such as persistence, amplitude, and chronodispersion changed significantly with RTNS (Fig. 2 and Table 4). F-waves recorded on the contralateral side also changed significantly from the baselines to the values after a 5-min RTNS; F-wave minimal onset latency shortened (p ¼ 0.011) from 50.1 ± 4.1 ms to 47.9 ± 4.2 ms, and F-wave conduction velocity increased (p ¼ 0.0019) from 50.1 ± 4.6 m/s to 53.0 ± 5.1 m/s. No other measures such as persistence, amplitude, mean latency and chronodispersion changed significantly with RTNS (Fig. 3 and Table 4). In contrast, F-waves recorded in the RTNS() group showed no significant changes from the baselines to the values after a 5-min rest without RTNS for any F-wave measures on both sides (Table 5). ICC values in measuring F-wave latencies showed very high levels of reliability; 0.990 for the mean onset latency and 0.975 for the minimal onset latency.

Please cite this article in press as: Nakajima N, et al., Unilateral repetitive tibial nerve stimulation improves neurogenic claudication and bilateral F-wave conduction in central lumbar spinal stenosis, Journal of Orthopaedic Science (2017), https://doi.org/10.1016/j.jos.2017.12.006

4

N. Nakajima et al. / Journal of Orthopaedic Science xxx (2017) 1e7

Table 2 Claudication distance in both RTNS(þ) and RTNS() groups. RTNS (þ) group

Claudication distance (m) a

RTNS () group

Pre-RTNS (mean ± SD)

Post-RTNS (mean ± SD)

P valuea

Baseline (mean ± SD)

Test walking study (mean ± SD)

P valuea

176.3 ± 96.3

329.4 ± 132.9

0.0004

138.5 ± 109.6

151.0 ± 111.6

0.14

Calculated according to Wilcoxon signed-rank test.

Table 3 Baseline F-wave data in both RTNS(þ) and RTNS() groups. RTNS (þ) group

Persistence (%) Peak-to-peak amplitude (mV) Mean latency (msec) Minimal onset latency (msec) Chronodispersion (msec) Conduction velocity (m/sec) a

P valuea

RTNS () group

Side of RTNS (mean ± SD)

Contralateral side (mean ± SD)

Rt (mean ± SD)

(mean ± SD)

79.8 ± 15.9 317.0 ± 116.6 54.4 ± 4.1 50.7 ± 4.0 10.3 ± 3.3 49.7 ± 4.2

69.6 ± 21.7 274.5 ± 134.6 54.7 ± 4.4 50.1 ± 4.1 10.6 ± 2.7 50.1 ± 4.6

83.6 ± 17.8 298.3 ± 126.5 52.9 ± 3.0 48.1 ± 3.1 12.5 ± 3.3 49.0 ± 4.0

85.2 ± 11.6 303.3 ± 78.0 53.9 ± 3.4 49.3 ± 3.1 10.9 ± 3.4 47.9 ± 4.1

Lt

0.20 0.45 0.67 0.35 0.40 0.69

Calculated according to Kruskal Wallis H-test.

Fig. 2. Sixteen consecutive M responses and F-waves recorded on the side receiving RTNS before (left) and after (right) RTNS in a 69-year-old patient. Analyses of 100 traces showed a change in F-wave minimal latency from 54.5 ms to 52.4 ms (arrowheads). Table 4 Comparison between pre- and post-RTNS F-wave data in RTNS(þ) group. Side of RTNS

Persistence (%) Peak-to-peak amplitude (mV) Mean latency (msec) Minimal onset latency (msec) Chronodispersion (msec) Conduction velocity (m/sec) a

Contralateral side a

Pre-RTNS (mean ± SD)

Post-RTNS (mean ± SD)

P value

Pre-RTNS (mean ± SD)

Post-RTNS (mean ± SD)

P valuea

79.8 ± 15.9 317.0 ± 116.6 54.4 ± 4.1 50.7 ± 4.0 10.3 ± 3.3 49.7 ± 4.2

78.6 ± 17.3 304.6 ± 160.7 53.9 ± 4.3 49.2 ± 4.2 11.1 ± 3.6 51.6 ± 4.4

0.87 0.66 0.026 0.00081 0.35 0.00081

69.6 ± 21.7 274.5 ± 134.6 54.7 ± 4.4 50.1 ± 4.1 10.6 ± 2.7 50.1 ± 4.6

72.3 ± 18.6 274.4 ± 181.6 54.3 ± 4.2 47.9 ± 4.2 12.6 ± 5.4 53.0 ± 5.1

0.51 0.89 0.27 0.011 0.22 0.0019

Calculated according to Wilcoxon signed-rank test.

3.3. M-wave study In the RTNS(þ) group, the M-waves evoked in the AH by tibial nerve stimulation at the knee neither showed RTNS-induced changes in amplitude nor in latency on the side receiving RTNS (p ¼ 1 and P ¼ 0.69) or on the contralateral side (p ¼ 0.95 and p ¼ 0.90). The M-wave values in the RTNS() group also remained

unchanged from the baselines in amplitude and latency on both sides after a 5-min rest (Table 6). 4. Discussion Neurogenic claudication typically occurs in central LSS, which is classified radiographically as the stenosis located in the central

Please cite this article in press as: Nakajima N, et al., Unilateral repetitive tibial nerve stimulation improves neurogenic claudication and bilateral F-wave conduction in central lumbar spinal stenosis, Journal of Orthopaedic Science (2017), https://doi.org/10.1016/j.jos.2017.12.006

N. Nakajima et al. / Journal of Orthopaedic Science xxx (2017) 1e7

5

Fig. 3. Sixteen consecutive responses recorded on the contralateral side to the RTNS before (left) and after (right) RTNS in an 80-year-old patient. Analyses of 100 traces showed a change in F-wave minimal latency from 51.6 ms to 48.4 ms (arrowheads).

Table 5 Comparison between baseline and repeat F-wave study data in RTNS () group. Rt

Persistence (%) Peak-to-peak amplitude (mV) Mean latency (msec) Minimal onset latency (msec) Chronodispersion (msec) Conduction velocity (m/sec) a

Lt a

Baseline (mean ± D)

After 5-min rest (mean ± SD)

P value

Baseline (mean ± SD)

After 5-min rest (mean ± SD)

P valuea

83.6 ± 17.8 298.3 ± 126.5 52.9 ± 3.0 48.1 ± 3.1 12.5 ± 3.3 49.0 ± 4.0

82.9 ± 14.3 349.6 ± 207.8 53.7 ± 3.6 48.5 ± 2.8 13.0 ± 4.2 48.6 ± 3.7

0.60 0.17 0.053 0.12 1 0.32

85.2 ± 11.6 303.3 ± 78.0 53.9 ± 3.4 49.3 ± 3.1 10.9 ± 3.4 47.9 ± 4.1

84.2 ± 17.2 288.5 ± 84.6 54.5 ± 3.2 49.8 ± 3.0 11.4 ± 3.1 47.5 ± 4.1

1 0.81 0.051 0.48 0.77 0.42

Calculated according to Wilcoxon signed-rank test.

Table 6 M-wave data in both RTNS (þ) and RTNS() groups. RTNS(þ) group

RTNS() group

Side of RTNS

Contralateral side

Rt

Lt

Pre-RTNS Post-RTNS P Baseline After 5-min rest P Baseline After 5-min rest P Pre-RTNS Post-RTNS P valuea (mean ± SD) (mean ± SD) valuea (mean ± SD) (mean ± SD) valuea (mean ± SD) (mean ± SD) valuea (mean ± SD) (mean ± SD) Amplitude (baseline 4.5 ± 2.7 to peak) (mV) Latency (msec) 13.5 ± 2.2 a

4.7 ± 3.7

1

3.3 ± 2.1

4.1 ± 2.8

0.95

5.6 ± 3.4

6.5 ± 4.0

0.39

6.1 ± 3.0

6.5 ± 2.7

0.17

13.3 ± 1.4

0.69

13.6 ± 2.4

13.3 ± 1.4

0.90

12.5 ± 0.8

12.5 ± 0.8

0.60

12.7 ± 0.6

12.8 ± 0.7

0.11

Calculated according to Wilcoxon signed-rank test.

canal with thecal sac compression tending to affect both legs [9]. Claudication symptoms may result from compression-induced focal nerve ischemia, as generally believed, because of its rapidly reversible nature. Various previous studies support this view. Cerebrospinal fluid pressure [10] or epidural pressure [11] measured in LSS patients increased while standing upright and walking. Electrophysiologic exploration with experimental focal compression of the peripheral nerve in humans showed evidence of compression-induced nerve ischemia directly under the localized compression [12,13]. Animal experiments on nerve compression [14,15] and neurogenic claudication [16,17] also revealed reduced blood flow and/or venous congestion of the nerve in association with an increase in endoneurial fluid pressure at the site of compression. Considering all these findings, it suggests that in central LSS, an activity-dependent and/or posture-related increase

in compression of the cauda equina, which is already ischemic and vulnerable, may induce a further reduction in its blood flow, giving rise to neurogenic claudication. In fact, a prostaglandin E1 analog, commonly used for LSS as a vasodilator that increases blood flow and inhibits platelet aggregation, proved effective in relieving claudication symptoms [18]. Claudication symptoms can significantly impact the patients' daily activities, so that claudication distance serves as a useful measure of quantifying the severity of functional impairment in LSS patients. Our previous work in central LSS patients, confirming the pioneering work by Tamaki et al., [1] showed that RTNS applied to one leg significantly increased their claudication distances. However, this finding may have simply resulted from the rest during RTNS and not from the RTNS per se, because claudication distance generally varies from one trial to the next and the second period of

Please cite this article in press as: Nakajima N, et al., Unilateral repetitive tibial nerve stimulation improves neurogenic claudication and bilateral F-wave conduction in central lumbar spinal stenosis, Journal of Orthopaedic Science (2017), https://doi.org/10.1016/j.jos.2017.12.006

6

N. Nakajima et al. / Journal of Orthopaedic Science xxx (2017) 1e7

walking can be longer than the first after a short rest. The present study, therefore, compared the data for the two groups of LSS patients: one group underwent RTNS in the supine position; and the other simply lay supine during the same period of time. As a result, the group receiving RTNS showed a significant increase in claudication distance, while the other group without RTNS did not, indicating that an improvement in neurogenic claudication in the present study appears to be directly attributable to RTNS itself. In most previous electrophysiologic studies on neurogenic claudication, the authors devoted their efforts to more objectively documenting the dynamic neurophysiologic changes related to posture and exercise. They showed some evidence of transient conduction abnormalities by using various techniques such as spinal cord evoked potentials [19], F-wave [20e22], H-reflex [23], cortical somatosensory evoked potentials and motor-evoked potentials [24]. The present F-wave study followed in the same line of these previous electrophysiologic studies on neurogenic claudication. However, this time we intended to document the electrophysiologic changes associated with an RTNS-induced improvement in neurogenic claudication. The F-wave results from backfiring of antidromicallyactivated anterior horn cells; in other words, the F-wave impulse first travels away from the recording electrodes toward the spinal cord before it returns to activate the distal muscles [8]. Therefore, unlike the ordinary nerve conduction studies for evaluating a relatively short distal portion of the peripheral nerves, F-wave measurement helps in assessing motor conduction along the entire length of the peripheral motor axons, including the most proximal segment. In one study on neurogenic claudication, serial F-wave recordings before and after ambulation revealed dynamic alterations in F-wave persistence and latency, probably reflecting dynamic conduction abnormalities across the cauda equina [21]. One previous study of Kumon et al. [3] has shown that RTNS applied to one leg significantly reduced the F-wave minimal latency, and increased the velocity on the side receiving RTNS. The present study duplicated the results by adopting the same experimental settings. More importantly, we have also documented that unilateral RTNS induced a significant F-wave latency shift on the contralateral side. This particular finding implies that RTNS applied to one nerve may not specifically affect only the stimulated nerve, but rather more extensively involve other unstimulated nerves. The contralateral effects may be reminiscent of the phenomenon termed “mirror-image pain”, i.e., bilateral pain induced by unilateral nerve insult. The physiologic mechanisms causing “mirrorimage pain” remain elusive, although spinal and/or supraspinal sensitizations of pain processing and peripheral mechanisms have been postulated [25]. We can only speculate on physiological mechanisms underlying this remote effect of RTNS on F-wave conduction. As shown by a previous study [26], which focused on patients with chronic pain in one limb, transcutaneous electrical stimulation of mild intensity applied to painful areas for 20e45 min significantly increased the skin temperatures on both the stimulated and contralateral limbs only in those who experienced pain relief during stimulation. This result suggested a generalized, stimulation-induced diminution of sympathetic tone associated with an elimination of pain-induced vasoconstriction. Consistent with this view, one animal study on stimulation-induced changes in the nerve blood flow measured with the thermal diffusion method showed that unilateral sciatic nerve stimulation increased blood flow in the ipsilateral 7th lumbar nerve root, and this effect was eliminated by intraperitoneal injection of a sympathetic nerve blocking agent [27]. A body of evidence also supports vascular effects induced by repetitive peripheral nerve stimulation, causing improved tissue oxygenation [28e30]. Based on these findings, we postulated that RTNS applied to one leg may give rise to widespread sympathetic blockade with vasodilation, which partially counters

compression-induced focal nerve ischemia in LSS patients, resulting in an improvement in both neurogenic claudication and F-wave conduction. Inconsistent with this view, however, the present study also showed that the M-response latency following tibial nerve stimulation at the knee remained unchanged on both sides after RTNS. The presence of F-wave latency shift in conjunction with unchanged motor conduction distally may indicate that RTNS specifically affects the proximal nerve segment rather than its widespread effect. This discrepancy may result in part from different sensitivity of the two electrophysiologic studies in assessing a diffuse or multisegmental process like widespread sympathetic tone reduction as postulated above. The ordinary nerve conduction study with Mresponse relies on the latency across a relatively short distal portion of the peripheral nerves. In contrast, F-wave latency reflects a longer path and accumulates all the segmental changes, which individually might be too small to show a clear deviation from the range of normal variability, detecting the diffuse or multi-segmental alteration better than a short segment study [8]. This point, however, requires further study. Regardless of the underlying mechanisms of its effect on neurogenic claudication, RTNS has practical value in providing a new therapeutic modality for LSS. There are some limitations to our study. First, repeated walking tests and F-wave studies were not conducted to investigate how long the effects of RTNS on neurogenic claudication and F-waves last. Second, we employed the stimulus intensity, duration, frequency and number of stimuli for RTNS similar to those originally described by Tamaki et al. [1], while the stimulation parameters most effective on neurogenic claudication and F-wave conduction remain to be elucidated. Nevertheless, this study will at least provide some basis for further refinement in its practical application to LSS treatment. We conclude that RTNS applied to one leg with a mild intensity of 5/s for 5 min significantly increases the claudication distances in central LSS patients. This clinical improvement is associated with a slight but significant shortening of tibial nerve F-wave latencies (i.e., an increase in F-wave conduction velocity) on both sides. These findings could be best explained by the RTNS-induced widespread sympathetic tone reduction with vasodilation, which partially counters a walking-induced further decline in nerve blood flow in LSS patients who already suffer from ischemic cauda equina. Conflict of interest The authors declare that they have no conflict of interest. References [1] Tamaki T, Tsuji H, Takano H, Takakuwa K, Noguchi T. Electrophysiological analyses on cauda equine intermittent claudication. Rinsho Seikei Geka 1986 Apr;21(4):513e7 [in Japanese]. [2] Takahashi K, Tomita K, Umeda S, Baba H, Kawahara N, Nagata S, Kikuchi Y, Nomura S. Transcutaneous electrical nerve stimulation for patients with lumbar spinal canal stenosis: preliminary report. J Jap Bioelectr Res Soc 1988 Nov;2(1):59e62 [in Japanese]. [3] Kumon M, Tani T, Ikeuchi M, Kida K, Takemasa R, Nakajima N, Kiyasu K, Tadokoro N, Taniguchi S. Influence of tibial transcutaneous repetitive electrical stimulation on neurogenic claudication and F-wave in lumbar stenosis. J Rehabil Med 2014 Nov;46(10):1046e9. [4] Kruidenier LM, Nicolai SPA, Willigendael EM, de Bie RA, Prins MH, Teijink JAW. Functional claudication distance: a reliable and valid measurement to assess functional limitation in patients with intermittent claudication. BMC Cardiovasc Disord 2009 Mar 2;9:9. https://doi.org/ 10.1186/1471-2261-9-9. [5] Lin JZ, Floeter MK. Do F-wave measurements detect changes in motor neuron excitability? Muscle Nerve 2004 Sep;30(3):289e94. [6] Taniguchi S, Kimura J, Yamada T, Ichikawa H, Hara M, Fujisawa R, Shimizu H, Tani T. Effect of motion imagery to counter rest-induced suppression of Fwave as a measure of anterior horn cell excitability. Clin Neurophysiol 2008 Jun;119(6):1346e52.

Please cite this article in press as: Nakajima N, et al., Unilateral repetitive tibial nerve stimulation improves neurogenic claudication and bilateral F-wave conduction in central lumbar spinal stenosis, Journal of Orthopaedic Science (2017), https://doi.org/10.1016/j.jos.2017.12.006

N. Nakajima et al. / Journal of Orthopaedic Science xxx (2017) 1e7 [7] Panayiotopoulos CP, Chroni E, Daskalopoulos C. The significance of F-chro syndrome and other nodispersion in the electrodiagnosis of Guillain-Barre motor neuropathies. Arch Neurol 1992 Mar;49(3):217e8. [8] Kimura J. Electrodiagnosis in diseases of nerve and muscle: principles and practice. 4th ed. New York: Oxford University Press; 2013. p. 149e79. [9] Truumees E. Spinal stenosis: pathophysiology, clinical and radiologic classification. AAOS Instr Course Lect 2005;54(29):287e302. [10] Magnaes B. Clinical recording of pressure on the spinal cord and cauda equina. part 2: position changes in pressure on the cauda equina in central lumbar spinal stenosis. J Neurosurg 1982 Jul;57(1):57e63. [11] Takahashi K, Kagechika K, Takino T, Matsui T, Miyazaki T, Shima I. Changes in epidural pressure during walking in patients with lumbar spinal stenosis. Spine 1995 Dec 15;20(24):2746e9. [12] Tani T, Ushida T, Kimura J. Sequential changes of orthodromic sensory nerve action potentials induced by experimental compression of the median nerve at the wrist. Clin Neurophysiol 2001 Jan;112(1):136e44. [13] Ikemoto T, Tani T, Taniguchi S, Ikeuchi M, Kimura J. Effects of experimental focal compression on excitability of human median motor axons. Clin Neurophysiol 2009 Feb;120(2):342e7. [14] Rydevik B, Lundborg G, Bagge U. Effects of graded compression on intraneural blood flow. J Hand Surg 1981 Jan;6(1):3e12. [15] Lundborg G, Myers R, Powell H. Nerve compression injury and increased endoneurial fluid pressure: a “miniature compartment syndrome”. J Neurol Neurosurg Psychiatry 1983 Dec;46(12):1119e24. [16] Takenobu Y, Katsube N, Marsala M, Kondo K. Model of neuropathic intermittent claudication in the rat: methodology and application. J Neurosci Meth 2001 Jan 15;104(2):191e8. [17] Ikawa M, Atsuta Y, Tsunekawa H. Ectopic firing due to artificial stasis in rat lumbar spinal canal stenosis model. Spine 2005 Nov 1;30(21):2393e7. [18] Takahashi J, Kobayashi H, Wakabayashi S, Deguchi M, Ito H, Mogami Y, Tanikawa H, Nakagawa H, Moriya H, Ashizawa R, Takahara K, Kinoshita H, Tateiwa Y, Misawa H, Tsutsumimoto T, Nakakohji T, Yuzawa Y, Sawaumi A, Hidai Y, Matsuda S, Nakamura I, Toba S, Kamimura M, Nakane T, Hirabayashi H, Hashidate H, Ogihara N, Mukaiyama K, Kato H, Ohtsuka K. The effect of a prostaglandin E1 derivative on the symptoms and quality of

[19]

[20]

[21] [22]

[23] [24]

[25]

[26] [27]

[28]

[29] [30]

7

life of patients with lumbar spinal stenosis. J Orthop Sci 2013 Mar;18(2): 208e15. Tsuji H, Tamaki T, Itoh T, Yamada H, Motoe T, Tatezaki S, Noguchi T, Takano H. Redundant nerve roots in patients with degenerative lumbar spinal stenosis. Spine 1985 Jan-Feb;10(1):72e82. Tang LM, Schwartz MS, Swash M. Postural effects on F wave parameters in lumbosacral root compression and canal stenosis. Brain 1988 Feb;111(Pt 1): 207e13. London SF, England JD. Dynamic F waves in neurogenic claudication. Muscle Nerve 1991 May;14(5):457e61. Bal S, Celiker R, Palaoglu S, Cila A. F wave studies of neurogenic intermittent claudication in lumbar spinal stenosis. Am J Phys Med Rehab 2006 Feb;85(2): 135e40. Pastor P, Valls-Sole J. Recruitment curve of the soleus H reflex in patients with neurogenic claudication. Muscle Nerve 1998 Aug;21(8):985e90. Baramki HG, Steffen T, Schondorf R, Aebi M. Motor conduction alterations in patients with lumbar spinal stenosis following the onset of neurogenic claudication. Eur Spine J 1999;8(5):411e6. Senba E, Okamoto K, Imbe H. Central sensitization and descending facilitation in chronic pain state. In: Wilke WS, editor. New insights into fibromyalgia. Rijeka Croatia: INTECH; 2012. p. 19e40. Abram SE, Asiddao CB, Reynolds AC. Increased skin temperature during transcutaneous electrical stimulation. Anesth Analg 1980 Jan;59(1):22e5. Takahashi K, Nomura S, Tomita K, Matsumoto T. Effects of peripheral nerve stimulation on blood flow of the spinal cord and the nerve root. Spine 1988 Nov;13(11):1278e83. Cramp FL, McCullough GR, Lowe AS, Walsh DM. Transcutaneous electric nerve stimulation: the effect of intensity on local and distal cutaneous blood flow and skin temperature in healthy subjects. Arch Phys Med Rehabil 2002 Jan;83(1):5e9. Sheffler LR, Chae J. Neuromuscular electrical stimulation in neurorehabilitation. Muscle Nerve 2007 May;35(5):562e90. n K, Hrafnkelsdo  ttir T, Jern S, Biber B, Mannheimer C, DuttaRoy S. TransHalle cutaneous electrical stimulation induces vasodilation in healthy controls but not in refractory angina patients. J Pain Symptom Manag 2010 Jul;40(1):95e101.

Please cite this article in press as: Nakajima N, et al., Unilateral repetitive tibial nerve stimulation improves neurogenic claudication and bilateral F-wave conduction in central lumbar spinal stenosis, Journal of Orthopaedic Science (2017), https://doi.org/10.1016/j.jos.2017.12.006