Paving the way for a better management of pain in patients with spinal cord injury: An exploratory study on the use of Functional Electric Stimulation(FES)-cycling

Abstract

Context/objective:

Chronic pain is common in patients with spinal cord injury (SCI), for whom it negatively affects quality of life, and its treatment requires an integrated approach. To this end, lower limb functional electrical stimulation (FES) cycling holds promise.

Objective:

To investigate pain reduction in a sample of patients with SCI by means of lower limb rehabilitation using FES cycling.

Design, setting and participants:

Sixteen patients with incomplete and complete SCIs, attending the Neurorobotic Unit of our research institute and reporting pain at or below the level of their SCI were recruited to this exploratory study.

Interventions:

Patients undertook two daily sessions of FES cycling, six times weekly, for 6 weeks.

Outcome measures:

Pain outcomes were measured using the 0–10 numerical rating scale (NRS), the Multidimensional Pain Inventory for SCI (MPI-SCI), and the 36-Item Short Form Survey (SF-36). Finally, we assessed the features of dorsal laser-evoked potentials (LEPs) to objectively evaluate Aδ fiber pathways.

Results:

All participants tolerated the intervention well, and completed the training without side effects. Statistically significant changes were found in pain-NRS, MPI-SCI, and SF-36 scores, and LEP amplitudes. Following treatment, we found that three patients experienced high pain relief (an NRS decrease of at least 80%), six individuals achieved moderate pain relief (an NRS decrease of about 30–70%), and five participants had mild pain relief (an NRS decrease of less than 30%).

Conclusion:

Our preliminary results suggest that FES cycling training is capable of reducing the pain reported by patients with SCI, regardless of American Spinal Injury Association scoring, pain level, or the neurological level of injury. The neurophysiological mechanisms underlying such effects are likely to be both spinal and supraspinal.

Introduction

Spinal cord injury (SCI) causes diverse sensory-motor deficits, including spasticity, muscle weakness and atrophy, and pain. Chronic pain can negatively affect mood, functioning, and quality of life, and is reported by approximately 70% of patients with SCI.¹ Such pain can be nociceptive, visceral, or neuropathic. Nociceptive pain has a prevalence of about 60% in SICI, and typically arises from damage to musculoskeletal tissue caused by trauma or inflammation, disuse of muscles or joints (e.g. due to upper limb paresis or being a wheelchair user), or spasticity-related muscle spasms and contractures.^2,3 Visceral pain is thought to originate from abnormal smooth muscle motility, mainly within the bowel and bladder.

Neuropathic pain has a prevalence of 30–40% in SCI and is related to neural tissue damage; it can develop acutely or post-injury at the site of the SCI (both peripheral and central neuropathic pain) or inferior to it (a central neuropathic pain condition).^2–4 In particular, neuropathic pain originates from structural, (mal)adaptive neuroplastic processes within both the site of injury (i.e. in the intact portion of the spinal cord above the SCI) and the brain [(mal)adaptive responses to SCI].⁵ These processes may involve persistent, local neuronal hyperexcitability due to the neural damage itself, and a dysfunctional relationship between the fast-transmitting lateral spinothalamic tract and the slow medially located polysynaptic pathway. These local modifications can, in turn, yield spontaneous activity in disinhibited polysynaptic pathways, sensitization of the spinothalamic tract, and ultimately, (mal)adaptive neuroplasticity processes within the thalamus or cortex. Consequently, these distal can further strengthen the local, dysfunctional changes, leading to exaggerated responsiveness to stimuli at or below average activation thresholds, which contributes to neuropathic pain, muscle spasticity, and autonomic dysreflexia.^4,5 However, the features of a patient’s pain depend also on the characteristics of the injury, including the degree of sensory and motor loss as defined by American Spinal Injury Association (ASIA) scoring, the neurological level of injury (NLI)(defined as the lowest spinal segment with intact sensation and Medical Research Council grade of >3 for muscle strength against gravity), and the sparing of sacral functions.⁴

The pathophysiological complexity of chronic pain in patients with SCI requires an integrated approach to managing that pain, consisting of drugs, psychology, physiotherapy, and surgical interventions.⁶ Physiotherapy aims to prevent and treat muscle and joint disuse as a means to manage the pain itself, in addition to promoting self-efficacy, activity, and participation in daily living activities. Regular training can significantly reduce nociceptive and neuropathic pain in people with SCI.^6–8 However, the neurophysiological correlates of these benefits are poorly understood. It has been proposed that physical exercise changes the perception of pain and reduces pain-related stress and depression, and consequently improves quality of life.

Furthermore, physical exercise can improve chronic musculoskeletal pain, thus influencing neuropathic SCI pain indirectly.^6–8 Training has to be intensive, repetitive, assisted, and task-oriented for successful sensorimotor recovery, which, in turn, improves pain potential. Cycling exercises using electrical passive pedaling systems equipped for functional electrical stimulation (FES) is a promising use of technological devices for lower limb rehabilitation in patients with SCI.⁹

The main aim of passive movement of the legs is to fully preserve the range of motion of the joints and muscles that are immobilized. Overall, FES cycling maintains muscle tone and flexibility of the joints and muscles, and prevents joint contractures, thus positively affecting spasticity, spasm, range of motion of the lower limbs, and muscle atrophy in individuals with SCI. Moreover, passive cycling influences cortical excitability, spasticity, bone mineral density, and myofibrillar protein content to similar degrees as active cycling.^10–19

FES involves the application of electrical currents to nerves and muscles using surface or implanted electrodes. FES-induced action potentials can activate muscular contraction even in paralyzed muscles.¹⁰ Furthermore, FES entrains activity-dependent neural plasticity mechanisms approximately at the myelomeres that correspond to the stimulated muscles, with potentially positive consequences on pain.¹¹ In this regard, it has been proposed that FES activates the dorsal columns, inhibits C-fibers, and consequently leads to decreased pain perception. FES might, therefore, lead to spinal stimulation that, in turn, decreases the excitability of small nerve fibers.

Therefore, FES and cycling training may combine to reduce pain either directly or because of the above mentioned positive effects on osteoarticular and muscle structures. Multiple sessions of FES cycling have been shown to provide SCI patients with significant cardiovascular, musculoskeletal, and neurological benefits in terms of spasticity, joint range of motion, muscle trophism, mobility, general health, and quality of life.^20–26 However, whether and how this occurs for pain remains undetermined.^20–26

This study is the first, to our knowledge, that is aimed at investigating the efficacy and neurophysiological underpinnings of FES cycling for the improvement of pain in individuals with SCI.

Materials and methods

Study population and setting

We consecutively enrolled patients with SCI attending our institute into this pilot study. The inclusion criteria were: (1) traumatic or non-traumatic, non-progressive SCI between the T3 and T11 vertebra to avoid possible co-occurrence of lower motor neuron lesion (see below); (2) acute/subacute phase (i.e. up to 12–18 months post-injury); (3) a 0–10 numerical rating scale (NRS) of at least 5 concerning below-level or at-level SCI pain; and (4) aged less than 65 years. Pressure ulcers, severely limited range of motion of the hips and knee joints, severe cognitive impairment, severe symptoms of depression or anxiety, lower motor neuron lesion, previous robotic rehabilitation, and severe pulmonary or heart disease were used as exclusion criteria. The clinical-demographic characteristics of the patients are summarized in Table 1. The local Institutional Review Board approved the study, and each patient provided their written informed consent to participate and to publication of the data.

Table 1.

Clinical-demographic characteristics.

ASIA	SCI pain level	Pain subtypes*			NLI	Sex	Age (y)	Post-SCI (m)
B	at-level		NeuP		T9	F	28	9
A	at-level		NeuP	o	T7	F	42	12
C	below-level		NeuP		T4	M	42	11
D	below-level		NeuP		T6	F	34	6
B	at-level	NocP			T6	F	42	4
B	at-level	NocP	NeuP	o	T6	M	42	6
A	at-level		NeuP		T7	M	20	13
A	at-level		NeuP		T10	M	38	13
B	below-level	NocP	NeuP		T4	F	34	10
C	below-level	NocP		o	T4	F	37	5
A	at-level	NocP	NeuP		T5	M	37	8
D	below-level		NeuP		T7	M	25	10
A	at-level	NocP		o	T10	F	26	7
B	below-level	NocP	NeuP		T8	M	35	14
A	at-level	NocP			T4	F	46	8
C	below-level	NocP			T10	M	30	9
A(6) B(5) C(3) D(2)	at-level(9) below-level(7)	56%	69%	25%	T4(4) T5(1) T6(3) T7(3) T8(1) T9(1) T10(3)	M = 8 F = 8	35 ± 7	9 ± 3

Data are reported as numbers or mean ± sd. * according to the International Spinal Cord Injury Pain (ISCIP) classification; neuropathic pain NeuP; nociceptive pain NocP; o other pain.

Procedures

Patients were provided with one session of FES-cycling in the morning and one in the afternoon, five times a week, using a motorized cycle ergometer equipped for FES. The FES device consistently delivers physiological activation of muscles during cycling, that is, hip and knee flexion–extension.^27,28 It is possible to achieve three main goals by coupling cycling with FES: (1) to perform pedaling in an assisted modality; (2) to apply an individually adapted resistance; and (3) to finely tune muscle contractions to help the patient in the pedaling task by using their muscle strength.²⁰

The first two sessions were aimed at familiarizing the patient with the equipment. Then, the duration of the sessions started at 15 min and was increased daily by 5 min, up to a maximum of 45 min. A full 45-min session consisted of a warm-up phase (5 min at 40 cycles/min), an FES preparation phase (5 min at 30% of maximum stimulation output, MSO), an active phase (30 min at 30 cycles/min and 100% MSO, with a resistance of 5 Nm), and a cool-down phase (5 min at 20 cycles/min). However, these parameters were subject to individual variations within and along sessions to avoid uncomfortable conditions and fatigue. The FES setup involved bipolar adhesive electrodes on the knee, hip and trunk extensors, and hip/trunk ab/adductors were used to deliver square, biphasic, alternated waves with a pulse -width of 500 µs, a stimulation frequency of 50 Hz, an MSO of 30–100%, a maximum resistance of 5 Nm, and an individually adapted current amplitude to achieve a visible muscle twitch (approximately 25–75 mA, 140 mA at most). The entire training program lasted six weeks. All subjects continued other regular rehabilitation activities (e.g. physiotherapy following the Bobath principles and occupational therapy) during the study, except robotic-assisted gait training.

Patients were evaluated at baseline (TPRE), immediately after (TPOST) and three months after the end of the training (TPOST3). During the follow-up period, patients continued to receive their common conventional treatments.

Outcome measures

As pain outcome measures, we adopted a 0–10 NRS as a unidimensional measure,²⁹ the Multidimensional Pain Inventory for SCI (MPI-SCI) as a multidimensional pain measure,³⁰ and the 36-Item Short Form Survey (SF-36) as a multidimensional pain inventory interference item.³¹ All such measures have established or adequate validity for assessing pain after SCI.³² Specifically, each patient received two evaluations, the second within two days of the first, from two physicians (XX and XX), neither of whom were involved in the patients’ care in compliance with International Spinal Cord Injury Pain Basic Data Set (ISCIPBDS) and the International Spinal Cord Injury Pain (ISCIP) classifications. These were developed to standardize the reporting of pain among the SCI population for evaluating and comparing the results from different clinics.^33–35 In particular, these approaches allow the three worst pain types to be investigated (with a clear distinction between nociceptive pain, neuropathic pain, other pain, and unknown pain) in terms of average intensity and interference, location, frequency, duration, and the impact of the pain on physical, social and emotional functions, and sleep.^33,34 The pain-related data included the character and the intensity (as per NRS) of each pain type. Using a detailed pain history (to include time since its occurrence, number of pain sites, pain-related disability, psychological status, and general health)³⁶ and a complete neurological examination that included the ASIA score and patient drawings showing all pain sites, each pain was matched separately by a neurologist with the characteristics classified according to the three tiers of the ISCIP classification, i.e. pain type (nociceptive, neuropathic, other, or unknown pain), pain subtype (musculoskeletal, visceral, other nociceptive pain), and pain level (at-level, below-level, other neuropathic pain).^33,34 All Cohen’s ĸ values, which were calculated to test the inter-rater agreement for the test−retest cases, were greater than 0.60 (defined as “good” agreement). Consistent with the ISCIPBDS approach, we referred to the average pain over the preceding week.

In addition, patients were evaluated on the Modified Ashworth Scale (MAS; from hip, knee and ankle flexor and extensor muscles) and the NRS on spasticity, the Penn Spasm Frequency Scale (PSFS), and the 12-Item Short Form Survey (SF-12) for quality of life (a scale ranging from 12 to 119, with highest value corresponding to the worst quality of life). These were used to determine the potential effect of spasticity on pain assessment.^37,38 Finally, we assessed the features of dorsal laser-evoked potentials (LEPs) to study the Aδ fiber pathways.^39,40

Laser evoked potentials

LEPs provide useful information on the lesions of the small, myelinated (Aδ) pathways within the spinothalamic tract that mediate signaling on pain and temperature.^39–45

LEPs were recorded from the dorsum skin area along the midline having mostly impaired sensation (assessed by clinical neurological examination) in patients with incomplete SCI. By contrast, in patients with a complete SCI lesion, LEPs were recorded in the dorsal zone of pain between the NLI and the distal area of abolished sensation.^39,40 Therefore, it can be assumed that the damaged myelomeres were always above the LEP dermatome. We stimulated the skin along the midline because we did not selectively study the spinothalamic tracts (requiring stimulation of the paravertebral skin at a distance of 3 cm from the midline, thus avoiding the contribution from ipsilateral spinothalamic tract).³⁹ The main advantages of recording dorsal LEPs consist of minimizing the LEP peripheral components and localizing the lesion site with precision.³⁹ LEP measurements might thus offer more useful information concerning SCI-related pain.

Laser stimuli were delivered by using an Nd:YAG laser device. We first measured the laser sensory and laser pain thresholds, rated as “painful–perceived”, “non-painful–perceived”, and “non-perceived”. Stimulation intensity was progressively decreased from painful–perceived until the stimulus was rated as non-perceived; then, the intensity was progressively increased until the stimulus was rated as painful–perceived. This procedure was repeated three times. The laser sensory threshold was defined as the average value between non-perceived and non-painful–perceived, whereas the pain threshold was set at the average value between non-painful–perceived and painful–perceived.

Cortical LEPs were recorded using laser stimuli delivered at twice the intensity of the pain threshold, every 7–15 s (25 stimuli in all). The laser beam was slightly displaced randomly to avoid sensitization and habituation. Patients were asked to lay relaxed, fixate on a sign on the ceiling, avoid eye blinking, and rate the pain intensity of each laser stimulus using a 0–10 NRS (with 0 to indicate “no pain” and 10 “the worst pain imaginable”). Cortical responses were recorded with Ag–AgCl electrodes at Cz referenced to linked earlobes (A1–A2), sampled at 500 Hz and filtered at 0.2–100 Hz. The sweep was set at 2.5 s. Skin impedance was kept below 5 kΩ. In this way, the main negative (N2) and positive (P2) components of the LEPs were recorded. All trials containing artifacts were rejected from the subsequent analysis after visual inspection. The remaining trials (15 ± 6) were time-averaged, thus providing estimates of N2 latency (measured in ms from stimulus onset to N2 peak) and N2P2 peak-to-peak amplitude (measured in μV).^40–47

Statistical analysis

A Friedman or a repeated-measure ANOVA was used, where appropriate (depending on the Kolmogorov–Smirnov test of normality, p > 0.05), to compare the differences in clinical (considering the minimal clinically important difference, MCID) and neurophysiological outcomes using the factor time (three levels: TPRE, TPOST, and TPOST3). Success rate with the 95% confidence interval (CI) of the proportion of patients achieving each outcome was calculated by using the Adjusted Wald method [W(95%CI)]. Success rates for the composite outcome (pain NRS and MAS) were also reported. Statistical significance was set at p < 0.05. Post-hoc t-tests were Bonferroni-corrected for multiple comparisons. The data collectors did not administer the treatment and were blinded to the patient's treatment.

Results

TPRE

Sixteen patients were enrolled into this pilot study according to the inclusion/exclusion criteria. Their clinical–-demographic characteristics are summarized in Table 1. All patients reported an NRS score of at least 5; eight reported scores above8 and mild lower limb spasticity and spasms, with related consequences on the cognitive–behavioral conceptualization of chronic pain and quality of life (Tables 2 and 3). Neuropathic pain was the most frequently encountered (69%), regardless of the NLI and the AIS grade (both p > 0.1), followed by nociceptive (56%) and unknown pain (25%). Musculoskeletal pain was the most frequently encountered nociceptive pain (44%), followed by neuropathic pain (72% at-level and 20% below-level). Thirty-eight percent of patients had only one type of pain, with a prevalence of neuropathic at-level SCI pain at higher AIS grades (Table 1). The other patients reported two or three different types of pain, with a prevalence of musculoskeletal pain at lower AIS grades (Table 1). There were no significant differences between socio-demographic (age, sex) or SCI characteristics and pain subtypes. The ĸ value of the type of pain between the two examiners was 0.694. All patients were already being treated with conventional medications such as anticonvulsants and antidepressants. LEPs were abnormal in 13 of the 16 patients (missing in four, abnormal in nine, due to increased latency and reduced amplitude for both below-level and at-level SCI pain) and normal in three patients (both below-level and at-level SCI pain) (Table 4 and Fig. 1). The laser sensory threshold was normal in all 16 patients, whereas the laser pain threshold was abnormal in all 16 (four lacked at-level SCI pain, 12 had increased thresholds for both below-level and at-level SCI pain).

Table 2.

The pain scores of individual patients at baseline (TPRE) and at different times of follow-up (TPOST and TPOST3), together with summary group (median and interquartile range).

Parameter	Patient n.	TPRE	TPOST	TPOST3
NRS-pain	1	9	7	6
	2*	5	4	4
	3*	5	3	4
	4	9	7	8
	5*	10	7	7
	6	10	7	7
	7*	5	2	4
	8*	10	0	2
	9*	5	3	3
	10*	5	2	2
	11	8	6	8
	12*	9	0	3
	13*	5	3	3
	14*	5	0	1
	15*	8	5	5
	16	5	3	5
		7(5–9)	3 (2–7)	4(3–7)
MPI-SCI	1	72	58	72
	2	48	43	44
	3*	60	51	52
	4	72	63	70
	5*	72	54	57
	6	48	42	44
	7	72	64	69
	8*	48	38	40
	9	48	40	40
	10*	48	39	41
	11	48	42	45
	12*	60	45	52
	13*	60	48	50
	14	72	52	68
	15*	72	47	52
	16	48	43	47
		60(48–72)	46(42–52)	51(44–60)
SF-36	1	86	68	79
	2*	94	73	77
	3*	84	56	73
	4	58	41	58
	5*	61	45	49
	6	60	48	56
	7*	79	57	67
	8	78	55	77
	9*	75	46	64
	10*	67	52	54
	11*	52	38	47
	12*	84	57	75
	13	93	59	90
	14	75	52	73
	15*	81	50	59
	16	81	52	68
		78(64–84)	52(48–57)	67(57–75)

*denotes the patients who achieved the minimal clinically important difference.

Table 3.

The spasticity and quality of life scores of individual patients at baseline (TPRE) and at different times of follow-up (TPOST and TPOST3), together with summary group (median and interquartile range).

Parameter	Patient n.	TPRE	TPOST	TPOST3
M	1	0	0	0
	2	18	14	17
	3	12	7	8
	4	6	4	6
	5	0	0	0
	6	6	4	6
	7	18	14	18
	8	6	4	4
	9	0	0	0
	10	12	8	9
	11	12	8	10
	12	18	14	16
	13	12	8	12
	14	6	5	5
	15	18	14	16
	16	0	0	0
		9(4–13)	6(3–9)	7(3–13)
NRS-spasticity	1	2	1	2
	2	0	0	0
	3	0	0	0
	4	0	0	0
	5	0	0	0
	6	1	1	1
	7	0	0	0
	8	0	0	0
	9	1	1	1
	10	1	1	1
	11	0	0	0
	12	2	1	1
	13	2	1	2
	14	1	1	1
	15	2	1	2
	16	2	1	2
		1(0–2)	1(0–1.2)	1(0–1.2)
PSFS	1	2	2	3
	2	0	0	0
	3	2	1	3
	4	2	2	2
	5	1	1	1
	6	1	1	0
	7	1	1	0
	8	2	2	1
	9	1	1	1
	10	0	0	1
	11	0	0	1
	12	0	0	1
	13	0	0	1
	14	1	1	1
	15	1	1	0
	16	0	0	1
		1(0–1.2)	1(0.1.1)	1(0.7–1.1)
SF-12	1	105	43	40
	2	99	63	55
	3	88	65	45
	4	72	67	69
	5	104	52	80
	6	74	64	59
	7	71	59	57
	8	81	44	76
	9	104	42	72
	10	86	48	73
	11	94	50	76
	12	119	41	55
	13	100	79	74
	14	70	77	45
	15	118	57	74
	16	73	79	43
		91(74–104)	58(47–65)	64(52–74)

Table 4.

The LEP features in each patient at baseline (TPRE) and follow-up (TPOST and TPOST3), together with summary group (mean ± sd).

Parameter	Patient n.	TPRE	TPOST	TPOST3
N2 latency (ms)	1	no LEP
	2*	401	382	368
	3*	211	204	200
	4	332	225	220
	5*	332	332	337
	6	314	304	301
	7*	447	443	430
	8*	no LEP
	9*	253	245	242
	10*	208	208	205
	11	313	303	294
	12*	438	446	435
	13*	no LEP
	14*	374	377	375
	15*	231	226	224
	16	no LEP
		321 ± 84	308 ± 89	303 ± 86
N2P2 amplitude (μV)	1	no LEP
	2*	7	8	7
	3*	7	7	7
	4	10	11	10
	5*	9	11	9
	6	6	7	7
	7*	8	9	8
	8*	no LEP
	9*	7	8	7
	10*	9	11	10
	11	6	7	6
	12*	5	6	5
	13*	no LEP
	14*	7	9	8
	15*	10	13	12
	16	no LEP
		7.6 ± 1.6	8.9 ± 2.1	8 ± 2

*denotes the patients who achieved the minimal clinically important difference in pain outcome measures.

Figure 1.

Average N2P2 LEP components for each patient with SCI elicited by stimulating below the level of the lesion. LEP were detectable in 12/16 of patients, of which three were normal (NLI reported in red characters).

TPOST and TPOST3

All 16 participants completed the trial without reporting any side effects. Patients significantly improved their MPI-SCI scores (χ² = 31, p < 0.0001; TPOST t₍₁₅₎= 7.6, p < 0.001; TPOST3 t₍₁₅₎= 4.2, P= 0.001) (Table 2); seven patients achieved the MCID [Δ= 20%; W= 0.44(0.23–0.66)] and were among those showing the highest NRS pain score changes.

NRS pain scores were also significantly improved (χ²= 26, p < 0.0001; TPOST t₍₁₅₎= 4.1, P= 0.001; TPOST3 t₍₁₅₎= 3.8, P= 0.002) (Table 2); 11 patients achieved the MCID [Δ= 18%; W= 0.83(0.62–0.97)], five of whom achieved an improvement of at least 50%. The remaining individuals showed no global changes or had a mild response to the treatment at TPOST. Concerning the distribution of pain scores at TPOST, the NRS score was approximately 6 (interquartile range, IQR= 3–8). At TPOST3, two patients again reported an NRS score above 8, whereas the other individuals continued to report an NRS score of about 5 (IQR= 4–7).

Lastly, SF-36 scores significantly changed over time (χ²= 31, p < 0.0001; TPOST t₍₁₅₎= 13, p < 0.001; TPOST3 t₍₁₅₎= 5.7, P= 0.002) (Table 2), and nine patients achieved the MCID [Δ= 15%; W= 0.56(0.33 to −0.76)].

A significant improvement was also found for MAS scores (χ²= 21, p < 0.0001; TPOST t₍₁₅₎= 6.7, p < 0.0001; TPOST3 t₍₁₅₎= 3.9, P= 0.002; MCID not yet standardized) and NRS spasticity scores (χ²= 10, P= 0.007; TPOST t₍₁₅₎= 4.5, P= 0.0009; TPOST3 P= 0.2; MCID not yet standardized), whereas changes in PSFS scores were not significant (P= 0.4) (Table 3). SF-12 scores also significantly improved (χ²= 18, p < 0.0001; TPOST t₍₁₅₎= 6.2, P= 0.0001; TPOST3 t₍₁₅₎= 7.7, p < 0.0001; MCID not yet standardized) (Table 3). The correlation between the successful relief of pain and a simultaneous improvement in MAS was not significant (r= 0.184, P= 0.5).

Concerning LEP features (Table 4), ANOVA analysis returned only a mild time effect for N2 latency (F_(2,30)= 4.9, P= 0.02) and a significant time effect for N2P2 amplitude (F_(2,30)= 16, P= 0.0001), with neither significant effects nor interactions appreciable with respect to ASIA, SCI pain level, and NLI. In particular, the N2 latency slightly decreased over time (TPOST P= 0.08; TPOST3 t₍₁₁₎= 2.5, P= 0.03), whereas the N2P2 amplitude significantly increased over time (TPOST t₍₁₁₎= 6.3, P= 0.0001; TPOST3 t₍₁₁₎= 2.6, P= 0.03).

Discussion

FES-cycling-based rehabilitation programs show promise for improving outcomes in individuals with SCI, but further studies are required.^20–26 Our data revealed that FES cycling training (provided using a relatively inexpensive electrometrical device) effectively reduced total pain symptoms and each specific quality of those symptoms, according to MPI-SCI and NRS pain scoring. A significant improvement was found in all clinical outcomes, which was clinically meaningful in a consistent proportion of the patients up to the three-month follow-up. Notably, the mean duration of the treatment effect was somewhat homogeneous, and the variability between participants was low. These effects represent a clinically impressive improvement compared to previous placebo-controlled NRS data on the reduction of neuropathic pain of different causes.⁴⁸ FES cycling also had clinically significant, positive consequences on quality of life, and was not associated with severe adverse events. These positive data might be related in part to the full adherence to the study protocol (no patient withdrew) and the short follow-up (no patient was lost at the three-month follow-up).

Furthermore, we found no significant differences in the response rates between patients who were or were not provided previously with pharmacological interventions such as anticonvulsants and antidepressants. This indicates that FES cycling could be suitable in patients for whom conventional medication is ineffective. Therefore, FES cycling might be indicated for pain management in patients with SCI. However, this conclusion has to be taken with caution because the sample was small and non-homogeneous with respect to ASIA score.

A particularly contentious issue is the association between neuropathic pain and the spasticity-related musculoskeletal pain.^38,49 The clinical differentiation between neuropathic and non-neuropathic pain (e.g. musculoskeletal pain, due to spasms, contractures, and overuse) is challenging. Neuropathic and spasticity-related pain are likely correlated through their shared neuroplasticity mechanisms, which are also the basis of functional recovery.^50–53 We found no correlation between pain improvement and spasticity reduction, although we can only speculate on this issue given that the enrolled sample was small and non-homogeneous with respect to ASIA score, and that spasticity changes were only partially appreciable. Furthermore, it must be considered that different subtypes of pain can correlate with spasticity. Therefore, it will be necessary to perform a multidimensional assessment of pain and spasticity in keeping with the multidimensional nature of such entities.^54–57

Given that the FES cycling group was not compared with a control group, one could argue that the positive pain outcomes were a placebo effect or simply due to a potentiation of ongoing pharmacological treatment. However, not only was pain relief observed at the end of the FES-based rehabilitation program, it lasted up to a further three months in a proportion of enrolled patients. Furthermore, the clinical changes were mirrored by increases in LEP amplitude and reduction of pain thresholds, which correlated with a reduction of severity and duration of the pain.^58,59 Finally, the improvements in pain symptoms experienced by our patients seemed somewhat greater to those reported from previous studies that compared tricyclic antidepressants and traditional anticonvulsants with placebo.^60–63

In our opinion, all these observations suggest a positive, neurophysiological effect on pain mechanisms by FES cycling. However, it remains to be clarified where in the body and how these neurophysiological effects occur, considering the possibility of a synergistic effect between ongoing drug treatment and FES cycling. In this regard, the neurophysiological mechanisms that mediate the effects on pain we describe are most likely as complex and as multileveled as pain in SCI is. The changes in LEP amplitude but not in latency suggest that pain improvement was, at least partly, due to changes in plasticity processes within the brain. The long-term effect on pain suggests the involvement of supraspinal mechanisms upon repeated exposure to FES cycling, which itself entrains neuroplastic adaptive mechanisms.¹¹ However, a spinal contribution cannot be ruled out. In particular, the motor practice likely favored the entrainment of residual connectivity across the injury site.^7,8,64 FES might also specifically contribute to this mechanism owing to the intense amount of proprioceptive information in parallel to cycling.^65–69 Accordingly, we may hypothesize that FES cycling exerts direct effects on pain, rather than operating on other, non-neuropathic pain, as suggested by the effect on LEP and the lack of correlation between pain and spasticity reduction (spasticity being a significant factor in patients with SCI).^70–73

Limitations

Although the internal validity of our data is sufficiently proven by the clinical and neurophysiological demonstration of the effect FES cycling has on pain in patients with SCI, the external validity (generalizability) of our data is limited. We must reiterate that this is the first study explicitly addressing pain reduction in patients with SCI with the use of FES cycling, and acknowledge the small sample size and non-homogeneity with respect to ASIA score, the lack of a control group, the short follow-up period, and the lack of detail in the neurophysiological assessment.

In particular, it is questionable whether physiotherapy received in addition to FES cycling was responsible for the decreased pain. Furthermore, it is worth acknowledging that in general, studies on pain treatments have very high placebo rates for improvement. It is therefore necessary to ascertain whether the improvements we observed were much more significant than those usually described for other non-blinded daily treatments. As such, a randomized controlled trial will be necessary to confirm our preliminary findings.

One might be concerned about the reporting of neuropathic pain intensity if a patient had more than one pain. Analogously, one could also query the differential diagnosis between neuropathic and nociceptive pain, especially if the NLI is thoracic. In this regard, we adopted the ISCIPBDS and ISCIP classifications described above. Specifically, we first categorized the pain in terms of its location relative to the level of the SCI; then, we differentiated pain subtypes according to the patient's history and by examination and laboratory and radiological data. Such repeated assessment over time should allow the differentiation of pain type.^74,75 Lastly, each patient received further evaluations within two days from two independent physicians. All Cohen’s ĸ values were greater than 0.60, thus suggesting good agreement on the differential diagnosis of pain. Furthermore, the high percentage of patients having more than one pain and the prevalence of the neuropathic pain type is consistent with the results of other studies.^3,76

In our sample, no individual with complete SCI reported below-level pain. It is known that individuals with SCI with complete and incomplete injuries can have below-level SCI (neuropathic) pain and that a substantial proportion of patients with below-level pain have complete spinal lesions.⁷⁷ Conversely, it has been reported that no significant systematic correlation exists for overall neuropathic pain after SCI, including sex, injury level, or extent of injury.⁷⁸ Furthermore, none of these parameters predispose for at-level or below-level pain, except for complete SCI, which better correlates with below-level neuropathic pain.⁷⁸ However, the sample size was small and patient inclusion was sequential. Therefore, further research is necessary to ascertain whether a different response would be expected if the participants reported pain in areas where they were insensate.

Although FES cycling seems therefore applicable to all patients with SCI to reduce their pain (except for specific contraindications), future controlled studies should evaluate the cost-effectiveness and clinical efficacy of FES cycling in comparison to other interventions. Furthermore, symptoms of anxiety and depression will have to be explicitly included as an additional endpoint, given that they can significantly correlate with symptomatic neuropathy and treatment effects. Therefore, more extensive and supportive studies are required to confirm our promising data.

Conclusions

Chronic pain is a debilitating secondary condition affecting more than two thirds of individuals who sustain a SCI. Moreover, chronic post-SCI is associated with depression, unemployment, and reduced quality of life, as well as significant costs associated with healthcare and lost productivity. Pharmacological treatments for chronic pain are only modestly effective and have significant side effects. There is a need for a comprehensive, community-based program that can offer individuals with SCI and chronic pain opportunities to integrate active pain management strategies into their daily lives along with the benefits of ongoing participation in activities and interventions that are effective in pain relief and management. In this regard, FES cycling could be considered a practical and valuable means to provide patients with SCI with beneficial outcomes in terms of pain and quality of life, as suggested by both clinical assessment and LEP data, since it can be incorporated into their daily health and wellness-promoting activities. Whereas the use of FES cycling has been shown to improve cardiovascular, musculoskeletal, and neurological outcomes, our work offers preliminary evidence beyond anecdotal reports on how FES can contribute to reducing pain.

Funding Statement

No funding to be reported.

Disclaimer statements

Conflict of interest Authors have no conflict of interests to declare.

Ethical approval All procedures performed in studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Helsinki declaration and its later amendments or comparable ethical standards. The local Institutional Review Board approved the study.

Informed consent Patients provided their written informed consent to study participation and data publication.