Transcutaneous electrical nerve stimulation effects on pain-intensity and endogenous opioids levels among chronic low-back pain patients: A randomised controlled trial

Abstract

BACKGROUND:

Transcutaneous electrical nerve stimulation (TENS) is a promising non-pharmacological modality for the management of chronic low back pain (CLBP), but its efficacy and mode of action have not been clearly established.

OBJECTIVE:

To evaluate the responses of plasma beta-endorphin ( $\beta$ E), met-enkephalin (ME), and pain intensity (PI) among patients with CLBP exposed to TENS or sham-TENS.

METHODS:

This double-blind trial involved 62 participants (aged 53.29 $\pm$ 5.07 years) randomised into TENS group (frequency 100 Hz, burst-rate 2 Hz, burst-width 150 $\mu$ s, intensity 40 mA, duration 30 min), and sham-TENS group. The PI and plasma concentrations of $\beta$ E and ME were measured at baseline, immediately (0 hr), 1 hr, 24 hrs, and 48 hrs post-intervention. Data were analysed using general linear model repeated measures, ordinal regression, one-way analysis of variance, Kruskal-Wallis test, independent and paired samples $t$ -tests, Mann-Whitney U test, Wilcoxon signed-rank test, and Kendall’s tau coefficient.

RESULTS:

There was a significant temporal difference in PI between groups, $F$ (1, 58) $=$ 18.83, $p<$ 0.001; the TENS group had better pain relief. The relative analgesic effect of TENS started immediately after the intervention (median difference [ $M D$ ] $=-$ 3, $p<$ 0.001), peaked at 1 hr ( $MD=-$ 4, $p<$ 0.001), and worn out by 24 hrs ( $MD=-$ 1, $p=$ 0.029). However, there was no significant difference in $\beta$ E and ME between the groups from 0 hr to 24 hrs post interventions, and no significant correlation between the PI, and $\beta$ E, or ME.

CONCLUSION:

TENS significantly reduced PI up to 24 hrs after treatment.

Keywords

Beta-endorphin met-enkephalin physiotherapy spondylosis TENS analgesic

1. Introduction

Chronic low back pain (CLBP) is defined as lower back pain that exceeds three months of onset [1, 2]. It affects about 15–45% of the global population [1]. CLBP is a major cause of years lived with disability [3], low socioeconomic status, and poor quality of life [4, 5, 6]. CLBP can be diagnosed through history taking, physical examination, diagnostic imaging, and diagnostic injection [2]. The most common cause of CLBP is lumbar spondylosis – an age-related degenerative change of the lumbar spine [2, 4]. There are four primary categories of CLBP interventions: physiotherapy (modalities, exercises, and behavioural techniques), pharmacotherapy, injection therapy, and surgical intervention [2, 6]. The efficacy of each intervention remains equivocal [2, 6]; hence, a multidisciplinary approach has been recommended [7].

Transcutaneous electrical nerve stimulation (TENS) is a non-pharmacological and non-invasive physiotherapy modality [6]. A meta-analysis of a few randomised control trials (RCTs) has shown that TENS could be useful in CLBP management [8], the post-stimulation analgesic effect of TENS can last from five minutes to 24 hrs [9]. TENS appears to have a safety advantage over other CLBP interventions [6]. Prolonged treatment with pharmaco- and injection therapies have been associated with adverse effects, and surgeries may result in complications such as nerve injury and paralysis [5, 10, 11]. Moreover, TENS can be self-administered, cost-effective, portable, and easy to maintain [9].

TENS mode of action remains controversial [9, 12], the theories of TENS analgesic mechanism can be classified under spinal inhibition, endogenous opioid production, and psychological effect [9]. The spinal inhibitory mechanism involves presynaptic inhibition of central nociceptive transmission (pain gate theory) [6, 13], decreased inflammation-induced dorsal horn neuron sensitisation [14], altered levels of neurotransmitters such as gamma-aminobutyric acid [15], and modulation of the activity of the glial cells in the spinal cord [16]. The endogenous opioid mechanism is based on a complex physiological system that modulates noxious stimuli by integrating multiple opioid receptors and endogenous opioid peptides such as enkephalins, endorphins, dynorphins, and endomorphins [17, 18, 19, 20]. Furthermore, the descending inhibitory activities relayed via the periaqueductal grey and the rostral ventral medulla in the brainstem are mediated via opioidergic pathways [13]. Also, TENS may produce psychological pain relief as demonstrated by the differences seen between sham-TENS and no treatment [6, 9].

The pain intensity (PI) in people with CLBP is usually accessed through subjective measures such as the visual analogue scale [21] and numeric pain rating scale (NPRS) [22]. The pain measure is meaningful when interpreted with the concept of minimum clinically important difference (MCID). The MCID is the smallest difference in pain score, which participants perceive as a significant improvement in their condition [23]; this gives credence to anchor clinical relevance to the patient’s experience [22]. Investigating the analgesic effect of TENS is clinically relevant and exploring the relationships between PI and concentrations of endogenous opioids after TENS may provide further insight into the pain modulation theory [24, 25].

The objectives of this RCT were to evaluate the (a) extent of analgesic effect of TENS on participants with CLBP, (b) changes in plasma concentration of beta-endorphin ( $\beta$ E) and met-enkephalin (ME), (c) temporal correlation among $\beta$ E, ME and PI at baseline, immediately (0 hr), 1 hr, 24 hrs, and 48 hrs after TENS for CLBP, and (d) sex and age-group differences in the levels of PI, $\beta$ E, and ME at baseline. The study hypotheses were that there will be no significant baseline-adjusted temporal (0 hr, 1 hr, 24 hrs, and 48 hrs) difference in the (a) PI between TENS and sham-TENS groups, (b) $\beta$ E levels between TENS and sham-TENS groups, and (c) ME levels between TENS and sham-TENS groups. (d) There will be no significant time-based correlation between PI and $\beta$ E, or PI and ME, (e) there will be no significant sex differences in the baseline levels of PI, $\beta$ E, and ME, and (f) there will be no significant age-group differences in the baseline levels of PI, $\beta$ E, and ME.

2. Materials and methods

2.1 Study design

This stratified-randomised, double-blind, sham- controlled trial involved 70 patients with CLBP. Figure 1 shows the CONSORT flowchart. The Health Research and Ethics Committee of the University of Nigeria Teaching Hospital (UNTH), Enugu, approved the study (NHREC/05/01/2008B-FWA00002458 – IRB00002323). The study was conducted following the guidelines of the revised Declaration of Helsinki 2013 and reported in adherence to CONSORT 2010. The participants signed a written informed consent before participating in the study. The trial was retrospectively registered with the Pan African Clinical Trial Registry on 22 April 2020 (PACTR202004888024754).

2.2 Participants, setting, and eligibility criteria

Participants were recruited from the UNTH and National Orthopaedic Hospital, Enugu (NOHE), Nigeria, between 6 October to 24 November 2016 ( $n=$ 40) and 3 April to 29 May 2018 ( $n=$ 30). Consultant orthopaedic doctors in each hospital identified potential participants (newly diagnosed with CLBP due to lumbar spondylosis) and referred them to research assistants (RA). Lumbar spondylosis was diagnosed by positive medical history, physical examination, and the presence of symptomatic degenerative changes in a magnetic resonance image of the lumbar spine. The RA explained the study protocol, objectives, potential benefits, and adverse effects to the participants and inquired if they wish to participate. Potential participants, who volunteered to participate, were invited to sign an informed consent form, and subsequently randomised into a study group. The study was conducted (outside both hospitals) at the Department of Medical Rehabilitation (DMR), University of Nigeria Enugu Campus (UNEC).

2.2.1 Inclusion criteria

Being an ambulatory male or female participant aged 20 to 70 years, who had been clinically diagnosed with CLBP resulting from lumbar spondylosis, with PI $\geqslant$ 5 on NPRS. Being TENS naïve, and available for the study and follow-up assessments.

2.2.2 Exclusion criteria

The following subjects were excluded from the study: individuals who had lumbar spine surgery or severe spinal conditions such as Pott’s disease, disc protrusion, lumbar (spinal and foramina) stenosis, radiculopathy, and spondylolisthesis $>$ 1 cm. People with skin lesions within L1 to L5 landmark, positive (pain on) vertical oscillatory thrust of any spinal vertebrae beyond L1 to L5, pain from hips, and sacroiliac joints, pregnancy, implanted electronic device (pacemaker), participants using steroids, central nervous system stimulants, and a therapeutic electrical stimulation modality were absolutely excluded from the study. Participants with a history of cancer, and any other systemic disease that can refer pain to the lower back, and participants in psychological distress.

2.2.3 Participants’ preparation

The participants were reminded of the study date, time, and venue, twice (four and two days before the study) through phone calls. They were instructed to refrain from consuming any drug, caffeine, alcohol, exercises, and other self-care-ameliorating factors for at least 48 hours before the intervention.

2.3 Randomisation, concealment, and group assignment

The study was conducted in four batches; to avoid keeping participants waiting until the required sample size is achieved. The first three batches involved 20 eligible participants, while the last batch involved ten eligible participants (Total $=$ 70). For each batch, the authors adopted a stratified randomisation method to control for potential sex- and age-related randomisation bias [26]. The participants were grouped by sex and stratified into age ranges (class width $=$ 10 years). Members of each stratum were then randomised to either sham-TENS (control) or TENS (trial) groups using concealed (computer-based) randomly generated numbers. The assignment of the participants into groups was facilitated by an independent lecturer from the DMR UNEC.

The participants, pain assessors (who assessed PI through NPRS), and laboratory analysts (who collected blood samples and analysed the $\beta$ E and ME levels) were all blinded to the assigned groups. However, two physiotherapists (who administered the therapy) were aware of the assigned groups but were blinded to the study hypotheses.

2.4 Interventions

Licenced physiotherapists administered the interventions starting with an explanation of the study protocol to the participants. Participants were told that they might, or might not feel the electrical sensation during treatment, depending on their pain threshold. The treatment group received a bi-channel burst modulation TENS (MediHightec Medical-MH6200 Combo; made in Taiwan) through two (50 mm by 50 mm) pairs of self-adhesive electrodes on side-lying position. Each pair of electrodes were placed 10 cm apart (paraspinal), across the surface landmarks of the L1 and L5 spinous processes [27, 28]. The TENS was delivered for 30 min, 100 Hz frequency, 2 Hz burst rate, 150 $\mu$ s burst width at 10 $\mu$ s per step, and 40 mA output intensity. The TENS dosage was similar to other burst TENS studies [17, 19, 29]. Maximum tolerable intensity (MTI) was recommended for optimum analgesic effect [24]; therefore, we determined the average MTI for the first five participants ( $=$ 40 mA), and it was tolerable for all participants.

Participants in the control group received a sham TENS – a pair of the active TENS machine with all parameters set at zero, except the timer. The sham TENS was set to beep at the end of the 30 min to mimic the active TENS. All participants were treated on side-lying, under aseptic conditions (skin toileting and personal electrodes).

2.5 Outcome measurements

The outcome measures were completed at baseline, and remeasured immediately (0 hr), 1 hr, 24 hrs, and 48 hrs after the interventions. All outcomes were conducted by study personnel who were blinded to group allocation. The primary outcomes for this study were PI, $\beta$ E, and ME. PI was measured using NPRS, scored from 0 (no pain) to 10 (most severe pain). A systematic review revealed that most studies defined the MCID as a one-unit change on the pain scale from the baseline score [22]. The NPRS is a reliable, valid, and responsive tool for assessing CLBP intensity [30].

The plasma concentrations of $\beta$ E and ME were measured by collecting the participants’ (5 ml) blood samples from the antecubital vein, through standard venepuncture technique [26], into ethylenediamine tetraacetic acid (EDTA) tubes (Caremax Co., Ltd; made in China). The samples were centrifuged at 3000 rpm for 5 min, the plasma was harvested and stored frozen at $-$ 20 ${}^{\circ}$ C until required for analysis. Afterwards, the plasma concentrations and rate of clearance of $\beta$ E and ME were analysed using a commercial human enzyme-linked immunosorbent assay (ELISA) kit (BIOTANG; made in the USA). The absorbance was read at 450 nm using a microwell plate reader (Diagnostic Automation-DAR 800; made in the USA).

Participants’ demographic information and duration of low back pain (PD) were recorded on a bio-data form. The secondary outcomes: resting heart rate (RHR), systolic (SBP) and diastolic blood pressure (DBP) were monitored using an automated electronic device (Omron digital BP monitor, Model 11 EM 403c; Tokyo Japan) [26]. The body temperature (Temp) was captured with a non-contact telephoto thermometer (Rossmax HC700; made in Switzerland).

2.6 Sample size and statistical analysis

The authors calculated the sample size based on published systematic reviews on TENS for CLBP [6, 8]. We assumed a moderate effect size of 0.16, with an alpha error probability of 0.05 and a power of 0.90 (G*Power 3.1.9.4 software). The output showed that a sample of 62 participants would have ample power for repeated measures analysis of variance within- and between interactions. However, we recruited 70 participants in anticipation of 10% to 15% attrition.

The data were analysed with SPSS 26 software (IBM Corp., Armonk, NY, USA). Normally distributed baseline data: age, PD, $\beta$ E, ME, SBP, DPB, RHR, and temperature were analysed using mean $\pm$ standard deviation, and independent samples $t$ -test for groups comparison. Sex distribution (binomial variable) was analysed using frequency and Chi-square ( $\chi^{2}$ ), while median $\pm$ semi-interquartile range and Mann-Whitney U test were used to compare the PI between the groups. The sex differences in the baseline $\beta$ E and ME were analysed with independent samples $t$ -test while PI was analysed with the Mann-Whitney U test (reported as Z-statistic). The age-group differences in the baseline $\beta$ E and ME were analysed with one-way ANOVA, while PI was analysed with the Kruskal-Wallis test.

Changes in the PI, $\beta$ E, and ME were computed by subtracting the baseline value from the value at the time: 0 hr, 1 hr, 24 hrs, and 48 hrs. We first applied a two-way mixed-design analysis of covariance (ANCOVA) controlling for age and pain duration but discovered that both variables were not statistically significant. Therefore, they were dropped from further analysis. General linear model (GLM) repeated measures were used to complete (a) within- and between-groups analyses, and (b) the interactions among time, group, and sex with regards to concentrations of $\beta$ E and ME. Then, ordinal regression was used to complete the analysis of PI, using PI at baseline as the dependent variable.

The following post hoc tests were completed to explain the pair-wise differences. Paired- and independent samples $t$ -tests were used for within- and between-group analyses of $\beta$ E and ME, respectively. Within- and between-groups differences in PI were analysed using Wilcoxon matched-pair signed-rank and the Mann-Whitney U tests, respectively. Finally, Kendall’s tau ( $\tau$ ) coefficient was used for the temporal correlation between PI levels and $\beta$ E, and ME concentrations. Kendall’s tau coefficient is more suitable for correlational analysis involving an ordinal variable such as PI [31].

Table 1
Participant characteristics

	TENS group Mean $\pm$ S.D.		Sham-TENS group Mean $\pm$ S.D.		All Mean $\pm$ S.D.		$t$ -statistic	$p$ -value
N	30		32		62
Sex:
Male	26		24		50		1.350 ${}^{\text{a}}$	0.245
Female	4		8		12
Age (years)	53.23	$\pm$ 5.14	53.34	$\pm$ 5.08	53.29	$\pm$ 5.07	$-$ 0.085	0.933
PD (months)	13.13	$\pm$ 10.68	13.06	$\pm$ 7.37	13.10	$\pm$ 9.04	0.031	0.976
PI (NPRS)	9.0	$\pm$ 0.50	9.0	$\pm$ 0.38	9.0	$\pm$ 0.50	$-$ 0.816 ${}^{\text{b}}$	0.415
$\beta$ E (ng/mL)	0.51	$\pm$ 0.04	0.49	$\pm$ 0.03	0.50	$\pm$ 0.03	1.893	0.063
ME (pg/L)	76.85	$\pm$ 5.17	73.97	$\pm$ 4.69	75.36	$\pm$ 5.10	2.303	0.025 ${}^{*}$
SBP (mmHg)	114.00	$\pm$ 7.02	117.78	$\pm$ 7.36	115.95	$\pm$ 7.39	$-$ 2.066	0.043 ${}^{*}$
DPB (mmHg)	74.57	$\pm$ 5.40	74.25	$\pm$ 7.00	74.40	$\pm$ 6.22	0.199	0.843
RHR (bpm)	82.93	$\pm$ 12.26	77.00	$\pm$ 11.61	79.87	$\pm$ 12.20	1.958	0.055
Temp. ( ${}^{\circ}$ C)	35.91	$\pm$ 0.47	36.04	$\pm$ 0.39	35.97	$\pm$ 0.43	$-$ 1.206	0.233

PD: Pain duration; PI: Pain intensity; $\beta$ E: beta-endorphin; ME: met-enkephalin; SBP: Systolic blood pressure; DBP: Diastolic blood pressure; RHR: Resting heart rate; Temp: Body temperature. Values given for PI are Median $\pm$ Semi-interquartile range. ${}^{\text{a}}=$ Chi-Square statistics. ${}^{\text{b}}=$ Mann-Whitney U test reported as Z-statistic. ${}^{\ast}=$ Statistic was significant at $p<$ 0.05 level (2-tailed).

Figure 1.

CONSORT flowchart.

3. Results

3.1 Recruitment and demographic data

Seventy individuals with CLBP provided written consent to participate in the study. They were stratified according to sex and age group and randomised ( $n=$ 35 each) into TENS and sham TENS groups. After randomisation, eight participants (TENS group $=$ 5, sham TENS group $=$ 3) withdrew from the study for personal reasons (11.4% attrition). Therefore, 62 participants that completed the study (TENS $=$ 30, sham TENS $=$ 32) were included in the statistical analyses. Figure 1 shows the CONSORT flowchart.

The participants were 50 males and 12 females, within the age ranges of 20 to 29 ( $n=$ 0), 30 to 39 ( $n=$ 0), 40 to 49 ( $n=$ 13), 50 to 59 ( $n=$ 40), and 60 to 70 years ( $n=$ 9). The mean age $\pm$ SD was 53.29 $\pm$ 5.07 years (range 42 to 63 years). Participants reported severe low back pain of 9.00 $\pm$ 0.50 (median $\pm$ semi-inter-quartile range) over an average duration of 13.10 $\pm$ 9.04 months. The baseline plasma levels of $\beta$ E and ME were 0.50 $\pm$ 0.03 ng/mL, and 75.36 $\pm$ 5.10 pg/L, respectively. There was no significant sex difference in PI ( $Z=-$ 0.82, $p=$ 0.415), $\beta$ E ( $t=$ 0.85, $p=$ 0.396), and ME ( $t=$ 1.03, $p=$ 0.306) at baseline. Similarly, there was no significant baseline difference in PI ( $H$ [2, 59] $=$ 4.16, $p=$ 0.125), $\beta$ E ( $F$ [2, 59] $=$ 5.13, $p=$ 0.601), and ME ( $F$ [2, 59] $=$ 0.74, $p=$ 0.482) across the age groups. The participants’ vital signs: temperature (35.97 $\pm$ 0.43 ${}^{\circ}$ C), RHR (79.87 $\pm$ 12.20 bpm), systolic (115.95 $\pm$ 7.39 mmHg), and diastolic (74.40 $\pm$ 6.22 mmHg) blood pressure were within normal ranges at baseline (Table 1).

Table 2
Analyses of between- and within-groups effects of TENS on $\beta$ E, ME and PI

Test source	Mean square	Partial Eta squared	$F$ -statistic (df)	$p$ -value
Between-participants effects
PI ${}^{\text{x}}$ Groups	29.40	0.25	18.83 (1, 58)	0.001 ${}^{*}$
PI ${}^{\text{x}}$ Sex	2.78	0.03	1.78 (1, 58)	0.188
PI ${}^{\text{x}}$ Sex ${}^{\text{x}}$ Groups	4.44	0.05	2.85 (1, 58)	0.097
$\beta$ E ${}^{\text{x}}$ Groups	0.12	0.36	33.13 (1, 58)	0.001 ${}^{*}$
$\beta$ E ${}^{\text{x}}$ Sex	0.03	0.12	7.68 (1, 58)	0.007 ${}^{*}$
$\beta$ E ${}^{\text{x}}$ Sex ${}^{\text{x}}$ Groups	0.03	0.13	8.85 (1, 58)	0.004 ${}^{*}$
ME ${}^{\text{x}}$ Groups	2718.14	0.36	33.12 (1, 58)	0.001 ${}^{*}$
ME ${}^{\text{x}}$ Sex	604.91	0.11	7.37 (1, 58)	0.009 ${}^{*}$
ME ${}^{\text{x}}$ Sex ${}^{\text{x}}$ Groups	697.90	0.13	8.50 (1, 58)	0.005 ${}^{*}$
Within-participants effects ${}^{\text{a}}$
PI ${}^{\text{x}}$ Time	174.26	0.91	338.97 (4, 232)	0.001 ${}^{*}$
PI ${}^{\text{x}}$ Time ${}^{\text{x}}$ Groups	37.18	0.69	72.37 (4, 232)	0.001 ${}^{*}$
PI ${}^{\text{x}}$ Time ${}^{\text{x}}$ Sex	0.15	0.01	0.51 (4, 232)	0.731
PI ${}^{\text{x}}$ Time ${}^{\text{x}}$ Groups ${}^{\text{x}}$ Sex	0.09	0.01	0.31 (4, 232)	0.875
$\beta$ E ${}^{\text{x}}$ Time	0.00 ${}^{\text{b}}$	0.07	4.42 (4, 232)	0.002 ${}^{*}$
$\beta$ E ${}^{\text{x}}$ Time ${}^{\text{x}}$ Groups	0.00 ${}^{\text{b}}$	0.04	2.20 (4, 232)	0.069
$\beta$ E ${}^{\text{x}}$ Time ${}^{\text{x}}$ Sex	0.00 ${}^{\text{b}}$	0.05	3.25 (4, 232)	0.013 ${}^{*}$
$\beta$ E ${}^{\text{x}}$ Time ${}^{\text{x}}$ Groups ${}^{\text{x}}$ Sex	0.00 ${}^{\text{b}}$	0.01	0.70 (4, 232)	0.593
ME ${}^{\text{x}}$ Time	86.66	0.07	4.32 (4, 232)	0.002 ${}^{*}$
ME ${}^{\text{x}}$ Time ${}^{\text{x}}$ Groups	42.35	0.04	2.11 (4, 232)	0.080
ME ${}^{\text{x}}$ Time ${}^{\text{x}}$ Sex	71.21	0.06	3.55 (4, 232)	0.008 ${}^{*}$
ME ${}^{\text{x}}$ Time ${}^{\text{x}}$ Groups ${}^{\text{x}}$ Sex	14.25	0.01	0.71 (4, 232)	0.585

${}^{\ast}F$ -statistic was significant at $p<$ 0.05 level (2-tailed). ${}^{\text{a}}=$ sphericity assumed. ${}^{\text{b}}=<$ 0.005. Time $=$ baseline, immediately after, one hour, twenty-four hours, and forty-eight hours after treatment. Groups $=$ TENS and sham TENS. Sex $=$ Male and Female. PI: pain intensity. $\beta$ E: beta-endorphin (pg/L). ME: met-enkephalin (pg/L).

3.2 Between- and within-group analysis of PI

A GLM repeated measures was completed to determine the main effects of group and time on PI and the interaction between time, group, sex, and PI (Table 2). There was a significant difference in the PI between the groups, $F$ (1, 58) $=$ 18.83, $p<$ 0.001. Similarly, there was a significant difference in the reported PI across time series, $F$ (4, 232) $=$ 338.97, $p<$ 0.001. However, there was no significant interaction between PI and sex. Age was earlier dropped from the model for failing the ANCOVA test.

Table 3
Baseline versus time differences of $\beta$ E and ME plasma levels, and PI within the sham-TENS group ( $n=$ 32)

Parameters	Post-test Mean $\pm$ S.D.		Baseline Mean $\pm$ S.D.		Mean difference (95% CI)		$t$ -statistic (df $=$ 31)	$p$ -value
Beta-endorphin ( $\beta$ E) ng/mL
$\beta$ E ${}_{0}$ – $\beta$ E ${}_{\text{b}}$	0.51	$\pm$ 0.04	0.49	$\pm$ 0.03	0.02	(0.00, 0.03)	1.54	0.133
$\beta$ E ${}_{1}$ – $\beta$ E ${}_{\text{b}}$	0.49	$\pm$ 0.04	0.49	$\pm$ 0.03	0.00	( $-$ 0.02, 0.01)	$-$ 0.37	0.718
$\beta$ E ${}_{24}$ – $\beta$ E ${}_{\text{b}}$	0.49	$\pm$ 0.02	0.49	$\pm$ 0.03	0.00	( $-$ 0.02, 0.01)	$-$ 1.05	0.303
$\beta$ E ${}_{48}$ – $\beta$ E ${}_{\text{b}}$	0.48	$\pm$ 0.03	0.49	$\pm$ 0.03	$-$ 0.01	( $-$ 0.03, $-$ 0.01)	$-$ 3.01	0.005 ${}^{\ast}$
Met-enkephalin (ME) pg/L
ME ${}_{0}$ –ME ${}_{\text{b}}$	75.84	$\pm$ 5.32	73.97	$\pm$ 4.67	1.87	( $-$ 0.60, 4.35)	1.54	0.133
ME ${}_{1}$ –ME ${}_{\text{b}}$	73.59	$\pm$ 6.15	73.97	$\pm$ 4.67	$-$ 0.38	( $-$ 2.47, 1.72)	$-$ 0.37	0.718
ME ${}_{24}$ –ME ${}_{\text{b}}$	72.94	$\pm$ 3.49	73.97	$\pm$ 4.67	$-$ 1.03	( $-$ 3.04, 0.98)	$-$ 1.05	0.303
ME ${}_{48}$ –ME ${}_{\text{b}}$	71.44	$\pm$ 3.91	73.97	$\pm$ 4.67	$-$ 2.53	( $-$ 4.24, $-$ 0.82)	$-$ 3.01	0.005 ${}^{\ast}$
Pain intensity (PI)
PI ${}_{0}$ –PI ${}_{\text{b}}$	7	$\pm$ 0.50	9	$\pm$ 0.38	$-$ 2	$\pm$ 0.50	$-$ 5.00	$<$ 0.001 ${}^{*\text{a}}$
PI ${}_{1}$ –PI ${}_{\text{b}}$	8	$\pm$ 0.50	9	$\pm$ 0.38	$-$ 1	$\pm$ 0.50	$-$ 4.90	$<$ 0.001 ${}^{*\text{a}}$
PI ${}_{24}$ –PI ${}_{\text{b}}$	9	$\pm$ 0.50	9	$\pm$ 0.38	0	$\pm$ 0.50	$-$ 3.74	$<$ 0.001 ${}^{*}$
PI ${}_{48}$ –PI ${}_{\text{b}}$	9	$\pm$ 0.50	9	$\pm$ 0.38	0	$\pm$ 0.50	$-$ 2.36	0.018 ${}^{*}$

$\beta$ E: beta-endorphin. ME: met-enkephalin. PI: pain intensity. Subscripts: b, 0, 1, 24, and 48 $=$ baseline, immediately after, one hour, twenty-four hours, and forty-eight hours after treatment, respectively. PI $=$ median $\pm$ semi-interquartile range; differences were calculated with Wilcoxon signed-rank test. $t$ -statistic $=$ paired samples $t$ -test. ${}^{\ast}$ Statistic was significant at $p<$ 0.05 level (2-tailed). ${}^{\text{a}}$ Minimal clinically significant difference $=$ median PI difference less than at least one unit from baseline [22].

Table 4

Baseline versus time differences of $\beta$ E and ME plasma levels, and PI within TENS group ( $n=$ 30)

Parameters	Post-test Mean $\pm$ S.D.	Baseline Mean $\pm$ S.D.	Mean difference (95% CI)	$t$ -statistic (df $=$ 29)	$p$ -value
Beta-endorphin ( $\beta$ E) ng/mL
$\beta$ E ${}_{0}$ – $\beta$ E ${}_{\text{b}}$	0.53 $\pm$ 0.07	0.51 $\pm$ 0.04	0.02 (0.00, 0.04)	1.69	0.102
$\beta$ E ${}_{1}$ – $\beta$ E ${}_{\text{b}}$	0.52 $\pm$ 0.05	0.51 $\pm$ 0.04	0.01 (0.00, 0.03)	1.90	0.067
$\beta$ E ${}_{24}$ – $\beta$ E ${}_{\text{b}}$	0.52 $\pm$ 0.04	0.51 $\pm$ 0.04	0.01 (0.00, 0.03)	1.70	0.100
$\beta$ E ${}_{48}$ – $\beta$ E ${}_{\text{b}}$	0.53 $\pm$ 0.04	0.51 $\pm$ 0.04	0.02 (0.01, 0.04)	3.22	0.003 ${}^{\ast}$
Met-enkephalin (ME) pg/L
ME ${}_{0}$ –ME ${}_{\text{b}}$	78.90 $\pm$ 9.85	76.85 $\pm$ 5.17	2.05 ( $-$ 0.84, 4.93)	1.45	0.158
ME ${}_{1}$ –ME ${}_{\text{b}}$	78.30 $\pm$ 6.94	76.85 $\pm$ 5.17	1.45 ( $-$ 0.61, 3.51)	1.44	0.162
ME ${}_{24}$ –ME ${}_{\text{b}}$	78.40 $\pm$ 5.86	76.85 $\pm$ 5.17	1.55 ( $-$ 0.62, 3.71)	1.46	0.155
ME ${}_{48}$ –ME ${}_{\text{b}}$	80.20 $\pm$ 5.81	76.85 $\pm$ 5.17	3.35 (1.12, 5.58)	3.07	0.005 ${}^{\ast}$
Pain intensity (PI)
PI ${}_{0}$ –PI ${}_{\text{b}}$	4 $\pm$ 0.50	9 $\pm$ 0.50	$-$ 5 $\pm$ 0.50	$-$ 4.88	$<$ 0.001 ${}^{*\text{a}}$
PI ${}_{1}$ –PI ${}_{\text{b}}$	4 $\pm$ 0.50	9 $\pm$ 0.50	$-$ 5 $\pm$ 0.50	$-$ 4.86	$<$ 0.001 ${}^{*\text{a}}$
PI ${}_{24}$ –PI ${}_{\text{b}}$	8 $\pm$ 0.50	9 $\pm$ 0.50	$-$ 1 $\pm$ 0.50	$-$ 4.18	$<$ 0.001 ${}^{*\text{a}}$
PI ${}_{48}$ –PI ${}_{\text{b}}$	9 $\pm$ 1.00	9 $\pm$ 0.50	0 $\pm$ 0.00	$-$ 2.45	0.014 ${}^{*}$

Table 5

Baseline-adjusted between groups comparison of PI, $\beta$ E, and ME plasma levels

Parameters	TENS group Mean $\pm$ S.D. ( $n=$ 30)	Sham group Mean $\pm$ S.D. ( $n=$ 32)	Mean difference (95% CI)	$t$ -statistic (df $=$ 60)	$p$ -value
Beta-endorphin ( $\beta$ E) ng/mL
$\beta$ E ${}_{0}$ – $\beta$ E ${}_{\text{b}}$	0.02 $\pm$ 0.05	0.02 $\pm$ 0.05	0.00 ( $-$ 0.02, 0.03)	0.33	0.744
$\beta$ E ${}_{1}$ – $\beta$ E ${}_{\text{b}}$	0.01 $\pm$ 0.04	0.00 $\pm$ 0.04	0.01 (0.00, 0.03)	1.58	0.118
$\beta$ E ${}_{24}$ – $\beta$ E ${}_{\text{b}}$	0.01 $\pm$ 0.04	0.00 $\pm$ 0.04	0.01 (0.00, 0.04)	1.99	0.052
$\beta$ E ${}_{48}$ – $\beta$ E ${}_{\text{b}}$	0.02 $\pm$ 0.04	$-$ 0.01 $\pm$ 0.03	0.03 (0.02, 0.06)	4.34	$<$ 0.001 ${}^{\ast}$
Met-enkephalin (ME) pg/L
ME ${}_{0}$ –ME ${}_{\text{b}}$	2.05 $\pm$ 7.73	1.87 $\pm$ 6.87	0.18 ( $-$ 3.53, 3.88)	0.09	0.927
ME ${}_{1}$ –ME ${}_{\text{b}}$	1.45 $\pm$ 5.52	$-$ 0.38 $\pm$ 5.82	1.83 ( $-$ 1.06, 4.71)	1.26	0.211
ME ${}_{24}$ –ME ${}_{\text{b}}$	1.55 $\pm$ 5.80	$-$ 1.03 $\pm$ 5.57	2.58 ( $-$ 0.31, 5.47)	1.79	0.079
ME ${}_{48}$ –ME ${}_{\text{b}}$	3.35 $\pm$ 5.98	$-$ 2.53 $\pm$ 4.75	5.88 (3.14, 8.61)	4.30	$<$ 0.001 ${}^{\ast}$
Pain intensity (PI)
PI ${}_{0}$ –PI ${}_{\text{b}}$	$-$ 5 $\pm$ 0.50	$-$ 2 $\pm$ 0.50	$-$ 3 $\pm$ 1.50	$-$ 6.70	$<$ 0.001 ${}^{*\text{a}}$
PI ${}_{1}$ –PI ${}_{\text{b}}$	$-$ 5 $\pm$ 0.50	$-$ 1 $\pm$ 0.50	$-$ 4 $\pm$ 2.00	$-$ 6.89	$<$ 0.001 ${}^{*\text{a}}$
PI ${}_{24}$ –PI ${}_{\text{b}}$	$-$ 1 $\pm$ 0.50	0 $\pm$ 0.50	$-$ 1 $\pm$ 0.50	$-$ 2.18	0.029 ${}^{*\text{a}}$
PI ${}_{48}$ –PI ${}_{\text{b}}$	0 $\pm$ 0.00	0 $\pm$ 0.50	0 $\pm$ 0.50	$-$ 1.08	0.279

$\beta$ E: beta-endorphin. ME: met-enkephalin. PI: pain intensity. Subscripts: b, 0, 1, 24, and 48 $=$ baseline, immediately after, one hour, twenty-four hours, and forty-eight hours after treatment, respectively. PI $=$ median $\pm$ semi-interquartile range; differences were calculated using Mann-Whitney U test, reported as Z-statistic. $t$ -statistic $=$ independent samples $t$ -test. ${}^{\ast}$ Statistic was significant at $p<$ 0.05 level (2-tailed). ${}^{\text{a}}$ Minimal clinically significant difference $=$ median PI difference less than at least one unit from baseline [22].

The within-groups analyses (Tables 3 and 4) showed a significant difference between the PI reported at baseline and at 0 hr to 48 hrs after sham-TENS and actual TENS, respectively. This implies that sham-TENS had a positive psychological effect. However, the between-group post hoc analysis (Table 5) showed that the TENS group had a significantly superior analgesic effect relative to sham-TENS. The baseline-adjusted differences in PI were significantly lower for TENS group than sham-TENS group at 0 hr (median difference [ $M D$ ] $=-$ 3, $Z=-$ 6.70, $p<$ 0.001), 1 hr ( $MD=-$ 4, $Z=-$ 6.89, $p<$ 0.001), and 24 hrs ( $MD=-$ 1, $Z=-$ 1.08, $p=$ 0.029). However, there was no significant difference at baseline and 48 hrs after interventions (Fig. 2).

Figure 2.

Temporal differences in pain intensities (NPRS) between the groups. Legend: a $=$ significant difference between TENS and Sham-TENS groups at the given time, b $=$ no significant difference between groups (Mann-Whitney U test, the alpha level was set at 0.05).

3.3 Between- and within-group analysis of

\beta

E and ME

There were significant main effects of group, time, and sex on the concentrations of $\beta$ E and ME (Table 2). Within-group analyses for sham-TENS participants (Table 3) showed significant changes in the concentrations of $\beta$ E ( $MD=-$ 0.01, $t=-$ 3.01, $p=$ 0.005) and ME ( $MD=-$ 2.53, $t=-$ 3.01, $p=$ 0.005) at 48 hrs relative to the baseline. Likewise, at 48 hrs there were significant differences in $\beta$ E ( $MD=$ 0.02, $t=$ 3.22, $p=$ 0.003), and ME ( $MD=$ 3.35, $t=$ 3.07, $p=$ 0.005) relative to baseline among the participants in the TENS group (Table 4). However, there was no significant difference between baseline and 0 hr to 24 hrs levels of $\beta$ E and ME within both groups.

When both groups’ baseline-adjusted data were compared (Table 5), the mean $\beta$ E level among TENS group participants was significantly higher than the sham-TENS group ( $MD=$ 0.03 ng/mL, $t=$ 4.34, $p<$ 0.001), at 48 hrs, only. A similar result was observed for changes in ME levels, TENS group was significantly higher than the sham-TENS group at 48 hrs only ( $MD=$ 5.88 pg/L, $t=$ 4.30, $p<$ 0.001).

There were significant main effects of sex in the concentrations of $\beta$ E and ME (Table 2). Irrespective of the groups, the mean concentration of $\beta$ E ( $MD=-$ 0.04 ng/mL, 95% CI: $-$ 0.07 ng/mL, $-$ 0.004 ng/mL, $t=-$ 2.22, $p=$ 0.03) and ME ( $MD=-$ 5.49 pg/L, 95% CI: $-$ 10.44 pg/L, $-$ 0.54 pg/L, $t=-$ 2.22, $p=$ 0.03), were higher in females than males immediately (0 hr) after interventions. There were no sex differences at 1 hr, 24 hrs, and 48 hrs after interventions. The time-matched correlation among PI, $\beta$ E, and ME (Table 6) showed that only ME at baseline correlated significantly with PI ( $\tau=-$ 0.25, $p=$ 0.017).

Table 6
Correlation between PI, $\beta$ E, and ME at matched times ( $n=$ 62)

Parameter	Kendall’s tau ( $\tau)$	$p$ -value
PI ${}_{\text{b}}$ vs. $\beta$ E ${}_{\text{b}}$	$-$ 0.14	0.180
PI ${}_{0}$ vs. $\beta$ E ${}_{0}$	$-$ 0.10	0.300
PI ${}_{1}$ vs. $\beta$ E ${}_{1}$	$-$ 0.09	0.380
PI ${}_{24}$ vs. $\beta$ E ${}_{24}$	$-$ 0.16	0.119
PI ${}_{48}$ vs. $\beta$ E ${}_{48}$	0.20	0.051
PI ${}_{\text{b}}$ vs. ME ${}_{\text{b}}$	$-$ 0.25	0.017 ${}^{*}$
PI ${}_{0}$ vs. ME ${}_{0}$	$-$ 0.10	0.300
PI ${}_{1}$ vs. ME ${}_{1}$	$-$ 0.09	0.380
PI ${}_{24}$ vs. ME ${}_{24}$	$-$ 0.16	0.119
PI ${}_{48}$ vs. ME ${}_{48}$	0.20	0.051

${}^{*}$ Indicates $\tau$ is significant at $p<$ 0.05 level (2-tailed). PI: pain intensity. $\beta$ E: beta-endorphin (pg/L).ME: met-enkephalin (pg/L). Subscripts: b, 0, 1, 24, and 48 $=$ baseline, immediately after, one hour, twenty-four hours, and forty-eight hours after treatment, respectively.

3.4 Harm

Participants were observed for treatment-related adverse reactions for at least 10 min after each sample collection and then followed up with a once-daily phone call for three days. A few patients reported mild pain at the cubital fossa, where the blood samples were collected. The preponderance of the patients on the TENS group complained of tingling electrical sensation as the TENS intensity was gradually increased to 40 mA, but they adapted within three minutes of the treatment. In the TENS group, two out of six patients with Fitzpatrick skin type III had mild skin erythema ( $\leqslant$ 2 hrs) on the sites at which the electrodes were applied. All patients tolerated the administered dosage of TENS; thus, the standard parameters were not reduced. No adverse reaction was reported among the patients that received sham TENS.

4. Discussion

CLBP is a major public health problem [32], accurate diagnosis and effective management of CLBP will significantly reduce its global disability burden [33]. A multidisciplinary approach involving medication, physiotherapy, and surgery is being used to manage CLBP [2, 6]. Non-surgical and non-pharmacological approaches tend to be safer and cost-effective over the long term [9, 10]. TENS is a non-pharmacological and non-invasive modality commonly used to manage CLBP, but its efficacy and mode of action remain equivocal [9, 12]. The present study was designed to determine whether TENS is efficacious among people with CLBP, the time limit of the analgesic effect, and the correlation between the effect and plasma-borne endogenous opioids.

CLBP patients treated with TENS experienced significant pain relief up to 24 hr after the treatment, compared to their counterparts that received sham TENS. This finding agreed with another RCT [21]. Three groups ( $n=$ 30 each) of patients with low back pain were exposed to either: low-frequency TENS (4 Hz, 200 micros), high-frequency TENS (110 Hz, 200 micros), or placebo TENS for 45 min. The authors concluded that the observed effect justified the clinical prescription of TENS among the population [21]. Similarly, a meta-analysis of 13 TENS studies concluded that treatment of CLBP with TENS demonstrated significant pain reduction and that TENS should be incorporated into CLBP management [8].

However, the result is contrasting with the outcome of Khadilkar and colleagues’ systematic review of the effect of TENS versus placebo for CLBP [34]. They concluded that the evidence from the small number of placebo-controlled trials does not support the use of TENS in the routine management of CLBP but encouraged further research. Similarly, a recent overview of eight systematic reviews comprising 51 RCTs found no evidence for the effectiveness of TENS compared to sham TENS in patients with CLBP [6]. The authors could not complete a meta-analysis due to the methodological heterogeneity of available trials; but recommended further studies with higher methodological quality, reduced risk of bias, and better precision. All the recommendations were considered in the present study.

In comparison with sham TENS, the TENS group had their PI lowered by three to four points on the NPRS between zero to one hour after the treatment. When tested at the twenty-fourth hour, there was still one point difference in favour of the TENS group. The outcome was statistically significant, but of minimal clinical importance [22, 23]. We inferred that the analgesic effect of TENS can last for one hour in people with CLBP. We could not find any meta-analysis on the duration of TENS analgesia in CLBP. However, a review article [9] found studies that reported analgesic durations ranging from five minutes to 24 hrs. Bearing in mind the heterogeneity of TENS for CLBP studies, a carefully designed systematic review is warranted.

Further analyses were conducted to determine the effect of TENS on the plasma levels of endogenous opioids. There was a marginal serial increase in the mean $\beta$ E level within the groups. Between-groups analysis showed that the TENS group had a significantly higher $\beta$ E level at 48 hrs. This finding partially aligned with the findings of an opioidergic study by Hughes and colleagues [17] on the response of plasma $\beta$ E to TENS in healthy participants ( $n=$ 36). They reported that mean plasma $\beta$ E concentration increased in participants treated with high frequency (101 Hz to 108 Hz) and low intensity (26 mA to 44 mA) TENS, or low frequency (4 Hz to 7 Hz) and high-intensity TENS (45 mA to 65 mA) and decreased in the placebo group.

We applied 100 Hz TENS, which resulted in no significant difference in the mean ME levels (compared with sham-TENS) at 0 hr to 24 hrs after treatment. Our finding was consistent with that of Han and colleagues [19], who conducted a study ( $n=$ 37) on the effect of low- (2 Hz) and high-frequency (100 Hz) TENS on Met-enkephalin-Arg-Phe (MEAP) and dynorphin A (Dyn A) immunoreactivity in human lumbar cerebrospinal fluid (CSF). Han and colleagues [19] found no significant difference in the mean MEAP (the analogue equivalent of ME) in CSF using 100 Hz TENS. The CSF samples were analysed for immunoreactive (ir) opioid peptides, MEAP from preproenkephalin, and Dyn A from preprodynorphin, respectively. Application of 2 Hz TENS resulted in a marked increase (67%, $p<$ 0.05) of ir-MEAP but not ir-Dyn A, whereas 100 Hz TENS produced a 49% increase in ir-Dyn A ( $p<$ 0.01) but not ir-MEAP. Though we analysed ME levels using blood plasma, we found no significant ME difference between the groups. The findings imply that high-frequency TENS may not stimulate endogenous endorphins such as ME [9]. We suggest that the analgesic effect of high-frequency TENS is through the pain gate theory rather than enkephalin or endorphins stimulation.

Furthermore, the present study found no significant temporal correlation between the reported PI and levels of $\beta$ E or ME. This outcome differed from the finding of Hughes and colleagues [17], who opined that TENS stimulated plasma $\beta$ E reduced evoked potential response – a measure of pain threshold in healthy subjects. There were several methodological differences between the present study and that of Hughes and colleagues [17], the previous study was conducted on a small sample of young healthy subjects ( $n=$ 36, mean age $=$ 25 years). The sampling technique, inclusion criteria, and sample size calculation were ambiguous [17]. Scholars have proposed three mechanisms for the analgesic effect of TENS: spinal inhibition, production of endogenous opioids, and psychological effect [6, 9]. The analgesic effect of TENS among our study participants might be attributed to the pain gate mechanism. The psychological effect of sham-TENS was equally noted. We suggest that future studies should focus on other opioidergic pathways for a holistic view of the system. Other studies may also look into the psychological effect of TENS.

4.1 Clinical implications of the findings

The pain relief duration shown in this study is of clinical significance to patients and caregivers. TENS analgesia rose sharply after treatment and reached its peak in 1 hr. It implies that TENS can be used as an adjunct therapy to reduce medication dosages and associated adverse effects in people with CLBP. Moreover, physiotherapists can utilise the analgesia period to administer exercises and other back rehabilitation with long term benefits.

4.2 Strengths of the study

This study adhered to the recommendations for methodological innovation in further studies on the efficacy, duration, and mode of action of TENS for CLBP [6, 8]. This novel design involved expert orthopaedic physicians in the screening of participants. The eligibility criteria were comprehensive, and stratified randomisation was adopted to avoid grouping bias. The endogenous opioids were measured with a high-tech molecular assay. To the best of our knowledge, this is the only TENS study on people with CLBP that completed a four-course (48 hrs) post-intervention temporal analysis of within- and between-groups changes in PI, and the corresponding plasma endogenous opioid levels. Age, sex, PD, group, and time interactions were also evaluated.

4.3 Limitations of the study

We could not design a sham-TENS device that mimics the electrical sensation felt with operational TENS; however, this is a common limitation of virtually all TENS studies [6]. We analysed plasma $\beta$ E and ME, which may not be the only endogenous opioids available in the blood after TENS application. Moreover, there are pieces of evidence that CSF-based $\beta$ E and ME may have an influence on pain modulation [19]. Therefore, the present study could not explore all the potential opioidergic pathways essential for a generalisable conclusion.

5. Conclusion

Thirty-minute burst TENS significantly reduced CLBP intensity immediately ( $-$ 3), at 1 hr ( $-$ 4), and 24 hrs ( $-$ 1) after the intervention. The peak analgesia was at 1 hr. However, there was no significant baseline-adjusted between-group temporal difference in plasma concentration of the $\beta$ E, and ME, within the analgesia phase. Correspondingly, there was no significant correlation between the concentrations of $\beta$ E and ME and the reported PI levels at any given time after the intervention. We deduced that burst TENS analgesia was beyond the elevation of plasma $\beta$ E and ME. We align with other scholars that suggested the pain gate mechanism as responsible for the analgesic effect of high-frequency TENS.

Author contributions

CIE, GCO, and OKO contributed to the conception of this study. CIE, EON, GCO, OKO, ACO, OAA, and MEK made substantial contributions to the design, acquisition of data and performed the statistical analysis. OKO, BMK, and CIE were responsible for drafting the article. OAA, ACO, EON, and GCO contributed to its critical revision. All authors approved the final manuscript for publication.

Availability of data and materials

The datasets analysed during the current study are available in the Zenodo repository, https://doi.org/10.52 81/zenodo.3739775.

Funding

None to report.

Footnotes

Acknowledgments

The authors thank the staff and management of the Department of Medical Rehabilitation, University of Nigeria, Enugu Campus (UNEC), and Safety Molecular Pathology Laboratory, Rangers Avenue, Enugu, Nigeria (formerly in UNEC), where this study was conducted. They also thank the staff and management of the University of Nigeria Teaching Hospital, Enugu, and National Orthopaedic Hospital, Enugu, Nigeria, especially the research assistants.

Conflict of interest

None to report.

References

Meucci

Fassa

Faria

NMX

. Prevalence of chronic low back pain: systematic review. Rev. Saúde Pública. 2015; 49: 73.

Middleton

Fish

. Lumbar spondylosis: Clinical presentation and treatment approaches. Curr Rev Musculoskelet Med. 2009; 2(2): 94-104.

GBD 2016 Disease and Injury Incidence and Prevalence Collaborators. Global, regional, and national incidence, prevalence, and years lived with disability for 328 diseases and injuries for 195 countries, 1990–2016: A systematic analysis for the Global Burden of Disease Study 2016. Lancet. 2017; 390(10100): 1211-1259.

van Oostrom

Monique Verschuren

de Vet

Picavet

. Ten year course of low back pain in an adult population-based cohort – the Doetinchem cohort study. Eur J Pain. 2011; 15(9): 993-998.

Andrew

Derry

Taylor

Straube

Phillips

. The costs and consequences of adequately managed chronic non-cancer pain and chronic neuropathic pain. Pain Pract. 2014; 14(1): 79-94.

Gibson

Wand

Meads

Catley

O’Connell

. Transcutaneous electrical nerve stimulation (TENS) for chronic pain – an overview of Cochrane Reviews. Cochrane Database Syst Rev. 2019; 4(4): CD011890.

Suman

Schaafsma

Buchbinder

van Tulder

Anema

. Implementation of a multidisciplinary guideline for low back pain: Process-evaluation among health care professionals. J Occup Rehabil. 2017; 27(3): 422-433.

Jauregui

Cherian

Gwam

Chughtai

Mistry

Elmallah

Harwin

Bhave

Mont

. A meta-analysis of transcutaneous electrical nerve stimulation for chronic low back pain. Surg Technol Int. 2016; 28: 296-302.

Wright

. Exploring the evidence for using TENS to relieve pain. Nursing Times. 2012; 11: 20-23. https://www.nursingtimes.net/clinical-archive/pain-management/exploring-the-evidence-for-using-tens-to-relieve-pain-09-03-2012/.

10.

McDougall

. Peripheral analgesia: hitting pain where it hurts. BBA Mol Basis Dis 2011; 1812: 459-467.

11.

Hayden

Ellis

Asbridge

Ogilvie

Merdad

Grant

DAG

Stewart

Campbell

. Prolonged opioid use among opioid-naive individuals after prescription for nonspecific low back pain in the emergency department. Pain. 2021; 162(3): 740-748.

12.

Sluka

Bjordal

Marchand

Rakel

. What makes transcutaneous electrical nerve stimulation work? Making sense of the mixed results in the clinical literature. Phys Ther. 2013; 93(10): 1397-1402.

13.

Ossipov

Dussor

Porreca

. Central modulation of pain. J Clin Invest. 2010; 120(11): 3779-87.

14.

Sabino

Santos

Francischi

de Resende

. Release of endogenous opioids following transcutaneous electric nerve stimulation in an experimental model of acute inflammatory pain. J Pain. 2008; 9(2): 157-163.

15.

Maeda

Lisi

Vance

Sluka

. Release of GABA and activation of GABA(A) in the spinal cord mediates the effects of TENS in rats. Brain Res. 2007; 1136(1): 43-50.

16.

Matsuo

Uchida

Nakajima

Guerrero

Watanabe

Takeura

Sugita

Shimada

Nakatsuka

Baba

. Early transcutaneous electrical nerve stimulation reduces hyperalgesia and decreases activation of spinal glial cells in mice with neuropathic pain. Pain. 2014; 155(9): 1888-1901.

17.

Hughes

Jr Lichstein

Whitlock

Harker

. Response of plasma beta-endorphins to transcutaneous electrical nerve stimulation in healthy subjects. Phys Ther. 1984; 64(7): 1062-1066.

18.

Gabis

Shklar

Geva

. Immediate influence of transcranial electrostimulation on pain and beta-endorphin blood levels: An active placebo-controlled study. Am J Phys Med Rehabil. 2003; 82(2): 81-85.

19.

Han

Chen

Sun

Yuan

Yan

Hao

Terenius

. Effect of low- and high-frequency TENS on Met-enkephalin-Arg-Phe and dynorphin A immunoreactivity in human lumbar CSF. Pain. 1991; 47(3): 295-298.

20.

Koneru

Satyanarayana

Rizwan

. Endogenous opioids: Their physiological role and receptors. Global Journal of Pharmacology. 2009; 3(3): 149-153.

21.

Warke

Al-Smadi

Baxter

Walsh

Lowe-Strong

. Efficacy of transcutaneous electrical nerve stimulation (TENS) for chronic low-back pain in a multiple sclerosis population: A randomized placebo-controlled clinical trial. Clin J Pain. 2006; 22(9): 812-819.

22.

Olsen

Bjerre

Hansen

Hilden

Landler

Tendal

Hróbjartsson

. Pain relief that matters to patients: Systematic review of empirical studies assessing the minimum clinically important difference in acute pain. BMC Med. 2017; 15(1): 35.

23.

Glassman

Copay

Berven

Polly

Subach

Carreon

. Defining substantial clinical benefit following lumbar spine arthrodesis. J Bone Joint Surg Am. 2008; 90: 1839-1847.

24.

Vance

Dailey

Rakel

Sluka

. Using TENS for pain control: the state of the evidence. Pain Manag. 2014; 4(3): 197-209.

25.

Damien

Colloca

Bellei-Rodriguez

CÉ

Marchand

. Pain modulation: From conditioned pain modulation to placebo and nocebo effects in experimental and clinical pain. Int Rev Neurobiol. 2018; 139: 255-296.

26.

Ezema

Omeh

Onyeso

OKK

Anyachukwu

Nwankwo

Amaeze

Ugwulebor

Nna

Ohotu

Ugwuanyi

. The effect of an aerobic exercise programme on blood glucose level, cardiovascular parameters, peripheral oxygen saturation, and body mass index among southern nigerians with type 2 diabetes mellitus, undergoing concurrent sulfonylurea and metformin treatment. Malays J Med Sci. 2019; 26(5): 88-97.

27.

Rojhani-Shirazi

Rezaeian

. The effects of Transcutaneous Electrical Nerve Stimulation on postural control in patients with chronic low back pain. J Med Life. 2015; 8(Spec Iss 2): 19-27.

28.

Neto

MLP

Maciel

LYS

Cruz

KML

Filho

VJS

Bonjardim

DeSantana

. Does electrode placement influence tens-induced antihyperalgesia in experimental inflammatory pain model? Braz J Phys Ther. 2017; 21(2): 92-99.

29.

Verruch

Fréz

Bertolini

. Comparative analysis between three forms of application of transcutaneous electrical nerve stimulation and its effect in college students with non-specific low back pain. Br J Pain. 2019; 2: 132-6.

30.

Sakulsriprasert

Vachalathiti

Kingcha

. Responsiveness of pain, functional capacity tests, and disability level in individuals with chronic nonspecific low back pain. Hong Kong Physiother J. 2020; 40(1): 11-17.

31.

Croux

Dehon

. Influence functions of the Spearman and Kendall correlation measures. Stat Methods Appl. 2010; 19: 497-515.

32.

Vassilaki

Hurwitz

. Insights in public health: perspectives on pain in the low back and neck: global burden, epidemiology, and management. Hawaii J Med Public Health. 2014; 73(4): 122-126.

33.

Buchbinder

Blyth

March

Brooks

Woolf

Hoy

. Placing the global burden of low back pain in context. Best Pract Res Clin Rheumatol. 2013; 27(5): 575-589.

34.

Khadilkar

Odebiyi

Brosseau

Wells

. Transcuta-neous electrical nerve stimulation (TENS) versus placebo for chronic low-back pain. Cochrane Database Syst Rev. 2008; 2008(4): CD003008.

Transcutaneous electrical nerve stimulation effects on pain-intensity and endogenous opioids levels among chronic low-back pain patients: A randomised controlled trial

Abstract

BACKGROUND:

OBJECTIVE:

METHODS:

RESULTS:

CONCLUSION:

Keywords

1. Introduction

2. Materials and methods

2.1 Study design

2.2 Participants, setting, and eligibility criteria

2.2.1 Inclusion criteria

2.2.2 Exclusion criteria

2.2.3 Participants’ preparation

2.3 Randomisation, concealment, and group assignment

2.4 Interventions

2.5 Outcome measurements

2.6 Sample size and statistical analysis

Table 1 Participant characteristics

3.1 Recruitment and demographic data

Table 2 Analyses of between- and within-groups effects of TENS on β E, ME and PI

Table 3 Baseline versus time differences of β E and ME plasma levels, and PI within the sham-TENS group ( n = 32)

Table 6 Correlation between PI, β E, and ME at matched times ( n = 62)

4. Discussion

4.1 Clinical implications of the findings

4.2 Strengths of the study

4.3 Limitations of the study

5. Conclusion

Author contributions

Availability of data and materials

Funding

Footnotes

Acknowledgments

Conflict of interest

References

Table 1
Participant characteristics

Table 2
Analyses of between- and within-groups effects of TENS on $\beta$ E, ME and PI

Table 3
Baseline versus time differences of $\beta$ E and ME plasma levels, and PI within the sham-TENS group ( $n=$ 32)

Table 6
Correlation between PI, $\beta$ E, and ME at matched times ( $n=$ 62)