The Effect of Transcutaneous Electrotherapy on Lumbar Range of Motion and Paraspinal Muscle Characteristics in Chronic Low Back Pain Patients: A Systematic Review and Meta-Analysis

Chronic low back pain (CLBP) affects paraspinal muscle size, quality (e.g., fatty infiltration), range of motion (ROM), and strength. Although transcutaneous electrotherapies are used to treat CLBP, their effects on paraspinal-related outcomes are not fully known. The aim of this systematic review and meta-analysis was to assess the overall effect of transcutaneous electrotherapies on trunk/lumbar ROM, paraspinal muscle morphology, and trunk muscle function (including strength and endurance) in CLBP patients. A systematic search of four databases and two study registers was conducted between 1 February 2022 and 15 September 2022. Two reviewers were responsible for screening and data extraction. Of the 3939 independent records screened, 10 were included in the systematic review and 2 in the meta-analysis. The results suggest there is limited evidence that both EMS and EMS plus exercise are superior to passive and active controls, respectively, for improving trunk muscle endurance. There is limited evidence that neither TENS nor mixed TENS are superior to controls for improving trunk muscle endurance. There is limited evidence that NMES is superior to passive controls for improving trunk muscle strength. The effect of transcutaneous electrotherapy on the other investigated outcomes was inconclusive. Future transcutaneous electrotherapy studies should focus on paraspinal-based outcomes that are under-studied.


Introduction
Chronic low back pain (CLBP) affects approximately 20% of the global population [1] and is the leading cause of years lived with disability worldwide [2], in spite of large healthcare expenditures towards its treatment [3]. There is growing evidence of morphological and functional changes to the paraspinal muscles in the presence of CLBP, such as reduced muscle size [4] and activation [5] and increased fat infiltration [6]. Changes to paraspinal muscle strength and endurance may also occur in conjunction with CLBP [7]. However, guidelines for the development of systematic reviews for CLBP have not highlighted the importance of evaluating objective changes to paraspinal muscle composition and function nor have they traditionally recommended examining fear-avoidance beliefs and pain catastrophizing as part of patient-centered outcomes [8,9].
Conservative treatments for CLBP include manual therapy, exercise, pharmacological interventions, and transcutaneous electrotherapy, among others. Transcutaneous electric nerve stimulation (TENS) is an electrotherapy that aims to stimulate sensory nerve fibres and is thought to promote analgesia by activating endogenous inhibitory pathways in the central nervous system and by reducing peripheral nociceptive output [10]. TENS continues to be used in the treatment of CLBP in spite of a lack of clear endorsement by medical guidelines [11,12]. The Cost B13 Working Group on Guidelines for Chronic Low Back Pain does not recommend TENS for CLBP, citing 'strong evidence' that it is not more effective than placebo or sham TENS [11], while the Low Back Pain Working Group of the North American Spine Society's Evidence-Based Clinical Guideline Development Committee found conflicting evidence that TENS improves pain or function in the shortmedium term [12]. However, recent systematic reviews [13,14] evaluating the effect of TENS on CLBP, with a focus on intervention timing and outcome assessment, have found a modest but positive effect on pain and function under certain conditions. For example, Jauregui et al. (2016) found significant weighted mean differences in pain intensity from pre-to post-treatment in patients treated with TENS for less than, but not more than, 5 weeks [13]. Wu et al. (2018) found that TENS improves functional disability when follow-up is within 6 weeks of treatment, compared to controls [14]. One limitation of these reviews is that their inclusion criteria were broad: for example, Wu et al. (2018) [14] included studies investigating both CLBP and LBP more generally. Another limitation is that neither examined paraspinal muscle-related outcomes, which to our knowledge, have never been included in a systematic review of TENS for CLBP.
Other forms of transcutaneous electrotherapy used to treat CLBP include interferential current (IFC) and electromyostimulation (EMS). IFC is created when two medium frequency alternating currents are crossed to create a 'low frequency' resultant current. It is claimed that IFC produces lower skin impedance, allowing current to penetrate tissue more deeply than conventional TENS [15]. IFC is most used for pain relief but may be used for muscle therapy. Electromyostimulation (EMS) aims to stimulate motor fibres to produce involuntary muscle contraction and is used to improve activation and strength in weakened but innervated muscles [15]. Neuromuscular electrical stimulation (NMES), sometimes called electrical stimulation (ES), is a form of EMS generally delivered to muscle under static conditions to evoke visible muscle contractions. Functional electrical stimulation (FES) pairs muscle stimulation with a functional task such as grasping an object [16], while whole-body electromyostimulation (WB-EMS) involves performing resistance training in conjunction with muscle stimulation [17]. Until recently, no systematic review had evaluated the effect of EMS on CLBP outcomes. The Philadelphia Panel (2001) was unable to find eligible studies investigating the effect of EMS in subjects with CLBP [18]. Poitras and Brosseau (2008) attempted to assess the effect of EMS in CLBP as part of a larger review article but found no relevant studies [19]. Recently, Linzmeye et al. (2022) reviewed the effect of NMES on lumbopelvic muscle function and paraspinal muscle thickness in CLBP patients. They reported that NMES helps improve paraspinal muscle endurance, while the effects on paraspinal muscle thickness were mixed [20]. While this was the first systematic review to specifically investigate the effects of transcutaneous electrotherapy on paraspinal muscle morphology, (i) it was limited to NMES-like modalities, (ii) did not assess other measurable markers of paraspinal muscle function (such as ROM), and (iii) both stand-alone and mixed interventions were compared/included in the same analysis.
To date, no study has comprehensively examined the effect of both sensory and motor transcutaneous electrotherapy on a wide array of observer-measured CLBP outcomes. Therefore, the aim of this systematic review was to assess the overall effect of transcutaneous electrotherapies on the following outcomes: trunk/lumbar ROM, paraspinal muscle morphology, and trunk muscle function (including strength and endurance) in patients with CLBP. The effect of transcutaneous electrotherapies was compared with passive (sham TENS, placebo) and active (other modalities) controls. Separate analyses were conducted for isolated electrotherapy interventions and mixed interventions. Focus was paid to short (post-intervention), medium (1 month post-intervention), and long-term (>1 month post-intervention) outcomes.

Materials and Methods
The current review protocol was registered with PROSPERO (ID number: CRD42023383928). This systematic review followed the recommendations suggested by the Preferred Reporting Items for Systematic reviews and Meta-Analyses (PRISMA) 2020 explanation and elaboration guidebook [21]. We also adhered to the guidelines proposed by Cochrane Neck and Back [22] for systematic reviews. The conceptualization and design of this systematic review was led by DW and supported by MF. Data extraction and the risk-of-bias assessment were conducted jointly by DW and BR. Data analysis and writing was performed by DW, and all three authors participated in the review and editing process.

Types of Studies
We included randomized and quasi-randomized control trials assessing the effect of transcutaneous electrotherapy in CLBP patients in comparison with a passive or active control (see below). In line with the Philadelphia Panel's consensus opinion [18], only studies with ≥5 participants per treatment group were included. To adequately assess the effect of interventions over time, we only included studies with ≥8 treatments per group. Only English-or French-language articles were included.

Participants
Participants aged 18-70 with a diagnosis of CLBP, defined as persistent pain between the lower ribs and gluteal fold, with or without leg pain, of at least 12 weeks duration were included in this review. Participants diagnosed with a specific spinal pathology, defined as one of the following: infection, tumour, previous lumbar surgery, osteoporosis, fracture, structural deformity (ex. scoliosis), inflammatory disorder (ex. ankylosis spondylosis), or cauda equina syndrome, were excluded. However, participants with lumbar disc herniation were included-provided they did not present with radicular symptoms-in line with evidence that disc degeneration and annulus tears (visible on T2-weighted MRI) are not necessarily painful [23]. Studies that included participants with mixed lower and upper back pain were excluded. Additionally, studies that included a mix of acute (<4 weeks) and chronic LBP patients as well as sub-acute (4-12 weeks) and chronic LBP patients were excluded, in line with suggestions that these conditions be considered separately [24].

Types of Interventions
We included studies that use transcutaneous electrotherapy as the primary intervention for CLBP. In cases of studies using transcutaneous electrotherapy plus another intervention, transcutaneous electrotherapy had to account for at least 40% of the treatment program. This cut-off value was previously used by Macedo et al. (2009) in their systematic review of motor control exercises for persistent LBP [24]. Studies that compared the effect of transcutaneous electrotherapy against a passive or active control were included. Passive controls were defined as the following: sham electrotherapy (defined as having the device modified so that no current passes to skin-surface electrodes), usual care, and/or no treatment. Active controls were defined as the following: any non-transcutaneous electrotherapeutic intervention for CLBP, such as ultrasound, hot/cold packs, exercise, mobilization/manipulation, massage/soft tissue therapy, acupuncture, and non-transcutaneous electrotherapy. Studies comparing two transcutaneous electrotherapeutic treatments were be excluded, as were studies comparing the effect of transcutaneous electrotherapy plus another intervention against a third intervention (ex. TENS + hot pack vs. ultrasound), in line with recommendations by the Cochrane Back and Neck Group [22]. Additionally, studies where the method of determining stimulation intensity is according to manufacturers' specifications, or where it is not described, were excluded.

Types of Outcome Measures
We included studies that assess at least one of the following outcomes: lumbar or trunk range of motion, paraspinal muscle morphology (muscle cross-sectional area, fat infiltration, thickness, stiffness), or paraspinal muscle function (muscle contraction, strength, or endurance). The effect of the interventions in the short term (post-intervention), medium (1 month post-intervention), and long term (>1 month post-intervention) were considered.

Search Strategy
The following bibliographic databases were searched for studies pertaining to CLBP and transcutaneous electrotherapy: PubMed, Scopus, Web of Science, and Embase. Additionally, the follow study registers were searched for protocols of the included studies:

Study Selection
Two reviewers (DW, BR) initially screened the search results for potential studies based on the study title and, where reasonable, the abstract. After excluding articles during this first round, the full text of the remaining articles was read by the same reviewers, and a global yes/no decision was made for each potential study based on the inclusion criteria identified in Section 2.1. In the case of disagreement over the inclusion of an article, a third reviewer (MF) was consulted, and a consensus decision between the three reviewers was taken. Study screening was managed using SR Accelerator (https://sr-accelerator.com, accessed on 15 September 2022).

Data Extraction and Risk of Bias
Two reviewers (DW, BR) independently extracted data from each included study using a modified version of the extraction template developed by Cochrane Back and Neck [22]. Participant characteristics, interventions, comparisons, outcomes, analysis approach, results, and study sponsorship were recorded.
Risk of bias was assessed using the revised Cochrane risk-of-bias tool (RoB 2) [25] for randomized trials. This tool examines five domains: randomization bias, bias due to deviations from intended interventions, missing outcome data bias, measurement bias, and bias in selection of the reported result. Bias is assessed on a per-outcome basis, and the tool uses signaling questions and an algorithm to guide reviewers to judgment. Each outcome is rated as follows: low risk, some concerns, and high risk. Both reviewers (DW, BR) strictly followed the instructions outlined in the full guidance document for the RoB 2 tool. In case of disagreement, a third reviewer (MF) was consulted.

Statistical Analysis
Studies were grouped according to intervention vs comparator (active or passive), at short-, medium-, and long-term follow-up (subgroup analysis). Meta-analyses were conducted, using a random-effects model, when comparisons within a group were sufficiently homogenous with respect to PICO variables (population, intervention, comparator, outcome). A minimum of three comparisons were needed for a comparison to be eligible for meta-analysis; in such cases, the overall treatment effect of the intervention, with 95% confidence intervals, was calculated for each outcome. For continuous variables measured using different scales, the standardized mean difference (SMD) was calculated. Statistical heterogeneity was conducted using the Q-test and reported as the I 2 statistic. We interpreted the statistic as follows: <40% suggests a low risk of heterogeneity, 40-75% a moderate heterogeneity, and >75% a high risk of heterogeneity [22]. We used the Review Manager statistical software (RevMan version 5.4.1) to conduct the meta-analyses.

Search Results
We performed an electronic search for eligible articles across the following four databases: PubMed, Embase, Scopus, and Web of Science. We also hand searched reference lists for articles that the electronic search might have missed. A total of 5839 records were found through the search. After removing duplicates, 3938 titles were screened and 89 were selected for a full-text review. Ten studies were included in the qualitative review and two studies (spanning three comparisons) were eligible for meta-analysis. The full results of our search are presented below ( Figure 1). ciently homogenous with respect to PICO variables (population, intervention, comparator, outcome). A minimum of three comparisons were needed for a comparison to be eligible for meta-analysis; in such cases, the overall treatment effect of the intervention, with 95% confidence intervals, was calculated for each outcome. For continuous variables measured using different scales, the standardized mean difference (SMD) was calculated. Statistical heterogeneity was conducted using the Q-test and reported as the I 2 statistic. We interpreted the statistic as follows: <40% suggests a low risk of heterogeneity, 40-75% a moderate heterogeneity, and >75% a high risk of heterogeneity [22]. We used the Review Manager statistical software (RevMan version 5.4.1) to conduct the meta-analyses.

Search Results
We performed an electronic search for eligible articles across the following four databases: PubMed, Embase, Scopus, and Web of Science. We also hand searched reference lists for articles that the electronic search might have missed. A total of 5839 records were found through the search. After removing duplicates, 3938 titles were screened and 89 were selected for a full-text review. Ten studies were included in the qualitative review and two studies (spanning three comparisons) were eligible for meta-analysis. The full results of our search are presented below ( Figure 1).

Risk of Bias
The risk of bias was evaluated on a per-outcome basis and ranged from 'some concerns' to 'high'. Specifically, bias in selected of the reported results was judged to be of at least 'some concerns' for all outcomes because no protocols for statistical analysis could be found for any of the included studies; therefore, a comparison between the published report and the protocol could not be performed, which automatically elevates the risk of bias for this domain. In total, there were 'some concerns' of bias for 16 comparisons [17,[26][27][28][29][30][31] and a high risk of bias for 4 comparisons [32][33][34]. The risk-of-bias assessment is presented in Supplementary Table S1.

Mixed TENS versus Active Controls at Post-Intervention
Two studies [28,34]

Mixed EMS versus Passive Controls at Post-Intervention
In their comparison of WB-EMS with passive control, Weissenfels et al. (2018) [17] reported that WB-EMS was significantly more effective at improving isometric trunk extension strength (MD = 8.29, 95% CI = 0.9, 16.4), but not isometric trunk flexion strength, at post-intervention.  [28] additionally looked at trunk muscle endurance 1 and 2 months post-intervention. There were no significant differences between TENS and sham TENS for any of the trunk endurance outcomes at either time point.

Mixed TENS versus Active Controls at Post-Intervention
Three comparisons across two studies [28,33] compared mixed TENS intervention with an active control for static trunk muscle endurance. For trunk flexion endurance, the meta-analysis revealed that mixed TENS was comparable to the active control (pooled SMD = −0.30, 95% CI = −0.81, 0.20). The comparisons made in this analysis had a moderate degree of heterogeneity (I 2 = 42%). For trunk extension endurance, the meta-analysis revealed that mixed TENS was comparable to the active control (pooled SMD = −1.86, 95% CI = −3.77, 0.06). The comparisons made in this analysis had a high degree of heterogeneity (I 2 = 94%). The results are presented below in Figures 2 and 3.

EMS versus Passive Controls at Post-Intervention
Two studies [30,32] [27] re-examined trunk extensor endurance at 6 months post-intervention. There were no significant differences between the mixed EMS and core exercise groups. However, the mixed EMS group continued to demonstrate grater static trunk endurance than the core exercise group (83.60 (41.92) vs. 37.10 (13.70), p < 0.05).

EMS versus Passive Controls at Post-Intervention
Two studies [30,32] assessed the effect of EMS on resting multifidus thickness. Batistella et al. (2019) [32] did not find significant between-group differences, while Pelegrini et al. (2019) [30] reported that the Aussie Current group had significantly greater thickness than the control group  [30] reported that the Aussie Current group had significantly greater resting multifidus thickness than the control group (4.23 (0.60) vs. 3.71 (0.50), p = 0.0161). Multifidus thickness was not reported on by Batistella et al. (2019) [32] for this time point.

Discussion
This is the first systematic review to evaluate the effect of transcutaneous electrotherapies on multiple aspects of paraspinal muscle function in CLBP patients. Previously, Linzmeyer et al. (2022) [20] assessed the effect of EMS on paraspinal muscle strength and size in this population and concluded that NMES improved paraspinal muscle strength, with mixed effects on muscle size. Our review included the four studies that appeared in Linzmeyer et al. (2022) [20] plus an additional two studies that used WB-EMS, a mixed intervention. We also included three sensory electrotherapy studies for which spinal/trunk ROM was an outcome of interest. Furthermore, we separated stand-alone electrotherapy interventions from mixed interventions and made separate comparisons for active and passive controls. These distinctions are important, as they help researchers assess the specific effect of an intervention and help guide clinical practice.
Three studies looked at the effect of sensory electrotherapy on spinal ROM. At postintervention, two of the studies found that participants who received sensory electrotherapy had better improvements to ROM than those who did not, while one study found that TENS-whether stand-alone or in conjunction with exercise-was less effective than exercise alone in improving ROM. This latter study (Kofotolis et al., 2008) [28] also reported that neither TENS nor mixed TENS were more effective than placebo TENS at improving ROM. In contrast to the results reported by Elserty et al. (2016) [33] and Lara-Palomo et al. (2013) [29], where sensory electrotherapy had a beneficial effect, the application of TENS by Kofotolis et al. (2008) [28] appears to have decreased treatment efficacy. This difference in outcome may in part be explained by differences in sensory application parameters between the studies. Elserty et al. (2016) [34] used conventional TENS (120 Hz) and Lara-Palomo et al. (2013) [29] used IFC, both of which are thought to provide immediate pain relief. In contrast, Kofotolis et al. (2008) [28] used low-frequency TENS (4 Hz), which has an onset of pain relief of at least 20 min post-treatment [15]. Given that participants in Kofotolis et al.'s mixed TENS group performed exercise 5 min after receiving TENS, it is possible that at the time of exercise, these participants had some lingering discomfort from the TENS treatment and were not yet able to benefit from it. The only study that re-examined outcomes at 1 and 2 months post-intervention-Kofotolis et al. [28]-found mixed TENS to be better than sham TENS at improving trunk extension ROM, although exercise alone was superior to all the other interventions at improving trunk ROM in general.
EMS paired with strength training-whether superimposed or sequential-appears to have a beneficial effect on isometric trunk flexion and extension strength. The three studies that used this intervention and measured paraspinal muscle strength reported significant improvements in strength from baseline [17,26,31]. However, while WB-EMS was significantly more effective than no training [17], neither WB-EMS nor NMES plus stabilization exercises were more effective than conventional training and stabilization exercises alone, respectively, for improving trunk muscle strength [26,31]. These results are broadly in line with research that suggests that NMES improves maximal voluntary contractions (MVC) in healthy subjects [35].
Two studies [28,33] examined the effect of TENS on trunk muscle endurance. Based on the results, exercise is superior to TENS at improving endurance in flexion and extension, with benefits lasting at least 2 months post-intervention. The relative benefit of TENS compared to placebo TENS for improving trunk endurance appears negligible. With respect to the three mixed TENS interventions [28,33]-which were all LF TENS (≤10 Hz) [36] either proceeding or following exercise-the meta-analyses revealed no significant difference from active controls. There is no obvious mechanism by which TENS directly improves muscle function; although a by-product of LF TENS is the generation of a slight muscle twitch, this treatment aims to stimulate small-diameter afferent nerve fibers [15] and does not induce tetany [15]. However, mixed TENS was superior to placebo TENS for improving trunk endurance, likely due to the exercise component.
EMS appears to generate short-term improvements in trunk endurance, and two of the three studies that tested EMS as a stand-alone intervention reported significant improvements in trunk endurance outcomes at post-intervention [30,32]. As with the strength outcomes, EMS was superior to passive but not active controls. When combined with exercise, EMS was superior to exercise alone at post-intervention. Importantly, in the study by Dimer da Luz et al. (2019) [27], both the EMS and mixed-EMS groups maintained improvements from baseline at long-term (6 months) follow-up, suggesting that EMS may have long-term effects on muscle physiology and/or neural activation in line with resistance training. Similar results were reported in a study of young men who underwent 24 weeks of resistance training followed by 24 weeks of de-training; after the de-training period, the participants reported significantly more strength than at baseline (p < 0.05) [37]. In contrast, the EMS groups in Batistella et al. (2020) [32] and Pelegrini et al. (2019) [30] reported a decrease in trunk endurance from post-intervention to 1 month post-intervention. There is no clear explanation for these differences, as all three studies were similar in terms of PICO variables: all contained a sample exclusively of women and all used medium-frequency electrotherapy for the same total number of sessions (12).  [30] alone. Additionally, the EMS groups in both studies had a decrease in multifidus thickness from post-intervention to 1 month post-intervention, suggesting a de-training effect that impacted the multifidus directly. A similar result was reported in a study of older adults who performed either trunk strengthening exercises or walking-balance exercises. Although the trunk strengthening group had significant increases in multifidus thickness at L4-L5 and L5-S1 after 12 weeks of training, these gains were lost after 6 weeks of de-training [38].

Limitations
This systematic review had a number of limitations. First, while we were able to include ten studies in the qualitative synthesis, only two were eligible for meta-analysis due to the heterogeneity of the interventions and outcomes. As a result, we were only able to run a meta-analysis for two outcomes, limiting the scope of our findings. Additionally, there was considerable variety in the studies' CLBP definition and inclusion criteria. Seven studies defined CLBP as low back pain of at least 3 months duration [26,27,29,30,[32][33][34], two defined it as low back pain experienced on ≥50% of days in the previous 3 months [17,31], while one defined it as 'chronic'-without specifying a baseline for symptom duration-whilst noting that participants were recruited from a pool of individuals who had LBP for at least 6 months [28]. In terms of inclusion criteria, three studies required participants to have at least a moderate amount of pain or disability at baseline [26,27,29] and the remaining seven did not [17,28,[30][31][32][33][34]. Requiring study participants to have at least a moderate degree of pain or disability allows for minimum important changes (MIC) to be detected. This value has been reported on for a number of common self-reported LBP questionnaires: the Visual Analogue Scale (MIC = 15), Numerical Pain Rating Scale (MIC = 2), the Oswestry Disability Index (MIC = 20%), and the Roland Morris Disability Questionnaire (MIC = 5) [39]. Although self-reported questionnaires were not the focus of this review, participants in three studies [28,30,32] had baseline ODI scores of <20%, which may have limited the potential for their muscle strength and endurance to improve.
One strength of the studies included in this review is that they provided electrotherapy protocols in sufficient detail to allow for replication. However, four of the seven [17,[26][27][28]31,33,34] studies that included an exercise component were vague in at least one aspect of the exercise protocol. Two studies did not clearly specify training volume [17,31], one did not clearly specify the parameters for exercise progression [28], and one neglected to provide the exercise program altogether [34]. Researchers should clearly define all aspects of an intervention protocol to allow for reproducibility.
Another limitation of this review is that there were relatively few studies included per outcome. Although many RCTs investigating the effect of transcutaneous electrotherapy on CLBP have been published, most have focused on self-reported outcomes such as pain and disability. Unfortunately, we were forced to exclude some studies investigating trunk-muscle-related outcomes for not meeting the inclusion criteria, including a lack of control group [40], insufficient treatment duration [41], and participant age [42]. If research is to paint a comprehensive picture of the effect of transcutaneous electrotherapy on trunkmuscle-related outcomes in CLBP patients, care should be taken to account for variables that can impede or confound the effects of an intervention. For example, 5-6 weeks is considered the minimum amount of time for NMES interventions to produce muscle hypertrophy [35,43], as strength gains made from shorter interventions will be the result of neural adaptations [43]. NMES researchers interested in paraspinal morphology outcomes can produce superior research by avoiding interventions of a shorter duration. With respect to age, smaller multifidus and erector spinae CSAs and higher levels of fat infiltration have been reported in older compared to younger individuals [44,45], with 60 years being identified as the age after which notable morphological changes occur [43]. Half of the studies in this review included participants > 60 years old [17,[29][30][31]33], which may have blunted the intervention effect. Future studies may want to only include participants aged 18-60 years to limit the confounding effect of age.
Finally, all the studies included in this systemic review had at least a moderate risk of bias, though this was due in large part to bias in the selection of the reported result, since none of the included studies published a protocol. Future RCTs can minimize the risk of bias by including the study protocol in their published report.

Conclusions
In conclusion, there is limited evidence that EMS is superior to passive, but not active, controls for improving paraspinal muscle endurance in CLBP patients at post-intervention. There is limited evidence that EMS is comparable to active controls for improving paraspinal muscle endurance at ≥1 month post-intervention and conflicting evidence for its benefit compared to passive controls. There is also limited evidence that EMS plus exercise is superior to exercise alone for improving paraspinal muscle endurance in CLBP patients at post-intervention but not at ≥1 month post-intervention. There is limited evidence that neither TENS nor mixed TENS are superior to controls (active and passive) for improving paraspinal muscle endurance in CLBP patients at post-intervention and ≥1 month postintervention. There is limited evidence that mixed NMES is superior to passive, but not active, controls for improving trunk muscle strength at post-intervention. There is conflicting evidence regarding the overall effect of EMS on paraspinal muscle thickness in CLBP patients. Finally, there is conflicting evidence about the overall effect of sensory electrotherapy on trunk/spinal ROM in CLBP patients, although higher pulse frequencies (80-120 Hz) [29,34] appear to be more effective than lower pulse frequencies (4 Hz) [28].
Our findings suggest that EMS is a viable alternative for improving paraspinal muscle endurance and may be particularly beneficial for patients with fear of movement or for whom exercise is not feasible. Although clinicians should prioritize the use of volitional exercise, when possible, EMS can form part of a multidisciplinary treatment strategy for CLBP alongside therapies such as exercise, mindfulness-based stress reduction [46], cognitive behavioral therapy [46], and oral supplementation with collagen peptides [47]. On the other hand, the results of this systematic review suggest there is no clinical basis for using TENS to improve paraspinal muscle endurance. To obtain a comprehensive understanding of the potential impact of transcutaneous electrotherapy for CLBP, future research can investigate outcomes that were not included in this review (ex: stand-alone EMS for paraspinal muscle strength, for which no eligible studies were found) or expand on the conflicting research into paraspinal muscle size and lumbar ROM. Where possible, researchers might try to adhere to the PICO variables of the studies included in this review in an intervention-dependent fashion. For example, the publication of more studies that used medium-frequency EMS and examined multifidus thickness would directly add to the existing body of research in this area and allow for future systematic reviews on this topic to be more robust.