https://orcid.org/0000-0002-8400-6463
Abstract
Specific neck exercises (SNE) targeting deep cervical flexors and extensors are commonly used for the treatment of nonspecific neck pain (NSNP). However, whether SNE are more effective than alternative exercise interventions (AEI) remains unclear. Furthermore, it has been postulated that SNE may be most effective when they are tailored and targeted to patients with evidence of motor control dysfunction, yet this notion also remains unproven. The objectives of this study were to compare the effectiveness of SNE with that of AEI for reducing pain and disability in people with NSNP and to assess whether the effectiveness of SNE is increased when the exercises are tailored and provided to patients with evidence of motor control dysfunction.
For this systematic review with meta-analysis, Medline, Web of Science, Scopus, and Physiotherapy Evidence Database were searched. Eligibility criteria included randomized controlled trials evaluating the effectiveness of SNE against that of AEI in people with NSNP. Meta-analysis included subgroup analyses to determine the effect of exercise tailoring and participant selection criteria on the effectiveness of SNE.
Twelve studies were included. Meta-analysis revealed greater effectiveness of SNE in the short to medium term for reducing pain (pooled standardized mean difference [SMD] = −0.41; 95% CI = −0.76 to −0.06; P = .02) and disability (pooled SMD = −0.41; 95% CI = −0.78 to −0.04; P = .03) but no differences in the long term for pain (pooled SMD = −1.30; 95% CI = −3.35 to 0.75; P = .21) and disability (pooled SMD = −1.81; 95% CI = −4.29 to 0.67; P = .15), although evidence was limited for the latter. The effectiveness of SNE was not superior in studies that included only participants with motor control dysfunction or when exercises were tailored to each participant. Overall, the studies were of low quality. Grading of Recommendations Assessment, Development and Evaluation revealed low certainty, serious risk of bias, and inconsistency of findings for short- to medium-term effects and very low certainty, serious risk of bias, and very serious inconsistency for long-term effects.
The preferential use of SNE may be recommended to achieve better short- to medium-term outcomes, although the low quality of evidence affects the certainty of these findings. Currently used strategies for selecting patients and tailoring SNE are not supported by the evidence and therefore cannot be recommended for clinical practice.
SNE are more effective than AEI for reducing pain and disability in patients with NSNP in the short to medium term, but overall evidence is of low quality, affecting the certainty of the findings. Tests of muscle dysfunction (mostly the craniocervical flexion test) currently used in studies to select patients and tailor SNE do not result in greater effectiveness of these exercises.
Evidence suggests SNE are more effective than other forms of exercise, although evidence is overall of low quality. Use of the craniocervical flexion test in isolation to select participants and/or tailor SNE cannot be recommended.
Introduction
Neck pain is highly prevalent1 and is the sixth largest cause of disability in the world,2 causing considerable economic impact.3 Between 50% and 75% of people will not recover from an acute episode and will experience recurrent neck pain within the next 1 to 5 years,4,5 Approximately 68% of people will develop persistent and chronic neck pain.6 Chronic neck pain can arise from a traumatic episode (eg, whiplash) or have no traumatic origin. In the latter, in the absence of any identifiable pathoanatomical cause,7 most patients are classified as having chronic nonspecific neck pain (NSNP).
Clinical practice guidelines for the management of NSNP advocate the use of exercise8–14; however, there is no consensus as to what type of exercise is most beneficial. General exercise8–10,12,13 and specific neck exercises (SNE)8,12 have been recommended.
Studies assessing the effects of strengthening exercises,15–29 postural reeducation,30–34 scapular stabilization and/or scapular muscle strengthening exercises,35–38 sensorimotor training,39–41 and stretching22,32,42–44 have reported favorable outcomes.
Extensive research has demonstrated neuromuscular disturbances in people with NSNP. Changes in motor output, such as reduced movement speed and less fluent movements, have been observed.45–47 Altered coordination between deep and superficial neck muscles, reduced specificity of neck muscle activity, and increased neck flexor and extensor coactivation have all been observed.48–50 Changes in muscle timing, such as delayed activation of the deep neck flexors, sternocleidomastoid, and splenius capitis51,52 and delayed relaxation of the sternocleidomastoid following activation,53 may also occur. Reduced neck muscle endurance is a further common feature, affecting the superficial and deep neck flexors and neck extensors.54,55
SNE targeting the deep neck flexors and extensors are effective for reverting some of these neuromuscular disturbances,56–59 resulting in improved neuromuscular function. However, whether SNE are more effective than other types of exercises for reducing pain and disability in NSNP needs to be established. Three previous reviews60–62 found no differences between the effectiveness of SNE and that of other specific forms of exercise (eg, strengthening), but each review included only a small number of studies in the meta-analysis (3,62 5,60 and 761). A recent systematic review 63 found SNE to be more effective; however, it was also limited by the number of relevant studies, because it included only studies that used a pressure biofeedback unit during the intervention, even though this approach is not considered mandatory for performing SNE.59,64 Furthermore, the review included nonrandomized clinical trials, which have been shown to bias results.65 These existing systematic reviews60–63 did not perform an assessment of treatment adherence, a key factor when assessing the effectiveness of exercise interventions.66–68
Clinical practice guidelines8,9,13,69 recommend evaluating motor control and strength impairments and subclassifying patients accordingly; however, the usefulness of these assessment strategies in obtaining better treatment outcomes has not been evaluated. It has been suggested that research on the effectiveness of SNE should use adequate selection of participants and appropriate exercise tailoring.60,70–76 It has been postulated that treatment needs may vary according to the specific presentation of the patient, with differences in motor control impairments, movement patterns, or causative factors between individuals. Hence, SNE that are designed to enhance motor control may have a larger effect in people who have NSNP and are selected according to predefined criteria (ie, displaying evidence of motor control dysfunction). It has also been proposed that the use of tailored SNE treatment targeting detected motor control disturbances and adapted and/or advanced according to each specific individual may be more effective than the application of a standardized SNE program with the same dosage and type of exercise given to all participants. The importance of precise exercise tailoring and appropriate dosing has also been highlighted by NSNP clinical guidelines.10,14,77 However, no systematic review has assessed whether the effectiveness of SNE is superior when clinical tests of muscle dysfunction are used to determine inclusion in the study or when exercises are tailored to the performance level of each participant.
The aim of this systematic review was twofold: first, to compare the effectiveness of SNE with that of other forms of exercise for reducing pain and disability in people with NSNP and, second, to assess whether the effectiveness of SNE (compared with other forms of exercise) is greater when the exercises are provided to patients who fulfill specific criteria of motor control impairments or when the exercises are tailored to the individual needs of each patient.
Methods
Search Strategy and Information Sources
The search was conducted from inception to April 5, 2021. Medline (search strategy is shown in Supplementary Table 1, Web of Science, Scopus, and Physiotherapy Evidence Database (PEDro) were searched using the keywords shown in Supplementary Table 2. The asterisk truncation operator was used to account for variations in spelling, and keywords in each row were combined using the Boolean operator “OR.” The results of searches produced by keywords in rows 1 and 2 were combined with the operator “AND,” and these were combined with the results of searches produced by keywords in row 3 using the operator “NOT.”
No date limit was set, and only randomized controlled trials were included. Search results were exported to Zotero software, version 5.0.88 (Corporation for Digital Scholarship, Vienna, VA, USA) for further processing. Reference lists of selected studies were also checked for further relevant studies.
Selection and Data Collection Processes
Inclusion and exclusion criteria for studies are shown in Supplementary Table 3. Titles and abstracts of all studies were screened for relevance, and the full text of potentially relevant articles was evaluated by 2 reviewers (I.V. and I.L.) to identify studies that met the inclusion/exclusion criteria. A data table (Suppl. Tab. 4) was used to extract information about authors, participant characteristics, exercise interventions, use of any tests of muscle dysfunction for participant selection or exercise tailoring, outcome measures, and results. In studies in which multiple comparisons were made, only those relevant to the aims of the systematic review were extracted. If necessary, authors were contacted for further information or clarification.
Data Items and Synthesis Methods
Meta-analyses were performed if 3 or more studies investigated the same or comparable outcome measure. Statistical analyses were performed in Review Manager 5.3 (The Cochrane Collaboration, Copenhagen, Denmark). Pain and disability outcomes were expressed as standardized mean differences (SMD) between groups (SNE vs alternative exercise interventions [AEI]) with a 95% CI. A random-effects model was used to assess the effect size, and the magnitude of the standardized effect was interpreted as small (0.2–0.5), moderate (0.5–0.8), or large (>0.8).78 A sensitivity analysis was performed to assess to what extent the estimated effect size was affected by the risk of bias of included studies. A primary meta-analysis included all studies, whereas a secondary meta-analysis excluded studies with a high risk of bias.
A subgroup analysis was performed to assess whether the effectiveness of SNE (compared with AEI) was superior in studies that used the presence of neck muscle dysfunction as an inclusion criterion or when exercises were tailored to the individual participant. A study was considered to use neck muscle dysfunction as a criterion if tests of muscle function or the assessment of aberrant movement patterns had been used to include/exclude participants in the study. A study was considered to tailor exercises to the individual participant if prescription and progression of the exercises had been adapted to each participant according to the level of muscle function or exercise capacity. The I2 statistic was used to assess heterogeneity between studies and was interpreted as low (<50%), moderate (50%–75%), and high (>75%).79
Assessment of Risk of Bias
Assessment of the risk of bias was performed using a modified version80 of the Cochrane Back Review Group tool,81 which incorporates a further criterion on the presence of a power calculation to justify the sample size. The tool consists of 13 criteria scored as “yes,” “no,” or “unclear,” indicating whether the criteria are met or information is not reported. Criterion 12 makes reference to the compliance of participants with the prescribed exercise program. There is currently no consensus on what is considered to be an acceptable rate of exercise compliance,67,82 and we considered that to meet this criterion, participants should have performed at least 80% of the prescribed program. Studies were evaluated independently by I.V. and I.L. Discrepancies were solved by consensus, and an independent third evaluator (D.F.) was consulted if needed.
As proposed by Furlan et al81 and Lascurain-Aguirrebeña et al8,0 studies were rated as having a low risk of bias when a minimum of 6 criteria were met, including criteria 1, 3, 5, and 12, because of their greater potential impact on treatment effect size. If studies met at least 6 criteria but did not meet 1 of those 4, then they were categorized as having a moderate risk of bias. All others were considered to have a high risk of bias.
The quality of evidence was assessed using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach.
Results
Study Selection
The selection of studies is summarized in the Preferred Reporting Items for Systematic Reviews and Meta-Analyses flow diagram shown in the Supplementary Figure. The search yielded 5809 studies; after the removal of duplicates and the exclusion of studies on the basis of the title or abstract, the full text of 103 studies was examined. Twelve of these 103 studies met the inclusion criteria. Several authors83–90 were contacted for further information or clarification about their study. No response was obtained from 385,87,90; hence, no further detail about AEI could be included for 2 studies,85,90 and no short-term data could be obtained for 1 study.87
Study Characteristics
Twelve randomized controlled trials comparing the effectiveness of SNE against AEI on pain and disability in NSNP were included (Suppl. Tab. 4). Studies used comparable outcomes. For measuring pain intensity, 8 studies used a visual analog scale,83,85,87–89,91–93 and 4 studies used a numerical pain rating scale.84,86,90,94 All studies measured disability with the Neck Disability Index.83–94 The majority of studies measured outcomes in the short to medium term (4–12 weeks postintervention), and only 3 studies measured outcomes in the long term (18,86 24,87 and 2689 weeks).
Three different SNE protocols were identified. Seven studies84–86,88,89,91,94 treated participants using the protocol proposed by Jull et al,94 4 studies83,90,92,93 based their treatment on the protocol described by Falla et al,95 and the remaining study87 used a mix of functional and nonfunctional exercises that aimed to strengthen deep flexor and extensor muscles and dissociate cervical and upper limb movements. With regard to AEI, 7 studies used strengthening and/or endurance exercises as a sole exercise intervention85,86,88,89,92–94 and 2 combined these with stretching exercises.84,90 Of these, 2 studies targeted the superficial neck flexors and extensors,84,92 3 studies targeted the superficial neck flexors only,91,93,94 1 study targeted the semipinalis cervicis muscle,86 2 studies targeted neck muscles in general (no further details could be obtained from the authors85,90), and 1 study involved strengthening of the craniocervical flexor muscles89; isotonic exercise was used in 4 studies,84,90,93,94 and isometric exercise was used in 6 studies.84–86,88–90 Postural exercises were used as a sole intervention87 in 1 study and were combined with isometric exercises in another.88 The remaining study83 used sensorimotor control exercises, such as eye following, gaze stability, and eye-head relocation.
Four studies83,90,91,94 included neck motor control impairment as an inclusion criterion, whereas 8 articles did not.84–89,92,93 The most commonly90,91,94 used criterion to assess neck motor control impairment was the inability to adequately perform at least the third stage of the craniocervical flexion test (CCFT). The remaining study83 used poor performance during active cervical movement control tests96 as an inclusion criterion. Exercise protocols were tailored to each specific participant in 8 studies,83,84,86,88,89,91,93,94 whereas nontailored exercises were used in 4 studies.85,87,90,92 For SNE, all studies83,84,86,88,89,91,93,94 tailored exercises according to the participant´s performance during the CCFT; exercises commenced at the level achieved in the first assessment and gradually progressed depending on each participant´s response. For AEI, 6 studies86,88,89,91,93,94 used a maximum voluntary contraction or maximum repetitions to tailor the exercise protocol to the participant´s capacity, whereas the remaining 2 studies gradually progressed the difficulty of the exercise84 or the speed and/or range of motion.83
Where the duration of face-to-face sessions with participants was reported, these ranged from 30 to 45 minutes.83,84,87,88,91,93,94 Their frequency varied; they were held 1,83,89,94 2,84,86 3,87,88,90,91,93 or 792 times per week. Two studies started with 1 session per week and progressed to 1 session every 2 weeks.83,89 In 6 studies, sessions were complemented with a home exercise program83,86,87,89,93,94 that required the participant to perform exercises 2 or 3 times per day with a duration of 10 to 20 minutes.83,94
Risk of Bias in Studies
Supplementary Table 5 summarizes the findings of the assessment of risk of bias. Five studies85,87,88,90,92 were classified as having a high risk of bias, and 7 studies83,84,86,89,91,93,94 were classified as having a moderate risk of bias. None of the studies was classified as having a low risk of bias.
Eleven studies83,85–94 did not ensure participant masking; thus, participants were masked in 1 study only.84 The assessor was not masked in 6 studies84,85,87,88,90,92 but was masked in the other 6 studies.83,86,89,91,93,94 Treatment compliance was considered acceptable in 2 studies83,92 and remained unclear for the remaining 10 studies.84–91,93,94 Because of the nature of the intervention, all studies could not mask the therapist. All studies used a correct randomization method; 5 studies83,84,86,89,91 analyzed participants in the group to which they were randomly allocated, whereas 7 studies did not.84,85,87,88,90,93,94 An adequate sample size calculation was not reported in 6 studies.84,85,88,90–92 The GRADE assessment (Suppl. Tab. 6) for short- to medium-term effects revealed low certainty, serious risk of bias, and inconsistency of findings. For long-term results, very low certainty, serious risk of bias, and very serious inconsistency and imprecision were found.
Results of Syntheses
Effectiveness of SNE vs AEI
All studies reported a reduction in pain intensity and disability in the short to medium term in both exercise groups. With regard to their relative effectiveness, the findings of the meta-analysis for pain intensity and disability are shown in Figures 1 and 2, respectively; these included 448 participants from a total of 11 studies. SNE were found to be more effective than AEI at reducing pain intensity over the short to medium term (P = .02), although the effect size was small (pooled SMD = −0.41; 95% CI = −0.76 to −0.06). Moreover, compared with AEI, SNE also showed a greater reduction in disability (P = .03), although the effect size was again small (pooled SMD = −0.41; 95% CI = −0.78 to −0.04). Heterogeneity in the results of the studies was moderate for pain (I2 = 70%) and disability (I2 = 72%). The sensitivity analysis revealed that findings were affected by the inclusion of 4 studies with a high risk of bias.85,88,90,92 In the secondary meta-analysis, the effect size for pain remained relatively unchanged (pooled SMD = −0.45; 95% CI = −0.84 to −0.05; P = .03), but the effect size for disability decreased and was no longer statistically significant (pooled SMD = −0.29; 95% CI = −0.68 to −0.10).
Effects of specific neck exercises versus alternative exercise interventions on pain intensity in the short to medium term (4–12 weeks) when a subgroup analysis was performed to compare studies that used motor control dysfunction as an inclusion criterion with those that did not.
Effects of specific neck exercises versus alternative exercise interventions on disability in the short to medium term (4–12 weeks) when a subgroup analysis was performed to compare studies that used motor control dysfunction as an inclusion criterion with those that did not.
With regard to long-term effects, 2 studies86,89 reported improvements in both exercise groups but no differences between groups, whereas the remaining study87 reported considerable differences in favor of SNE. The findings of the meta-analysis for pain intensity and disability in 106 participants are shown in Figures 3 and 4, respectively. Meta-analysis revealed no significant differences between interventions. Heterogeneity for pain intensity and disability was high, at 95% and 97%, respectively.
Effects of specific neck exercises versus alternative exercise interventions on pain intensity in the long term (18–26 weeks).
Effects of specific neck exercises versus alternative exercise interventions on disability in the long term (18–26 weeks).
Effect of Participant Selection and Exercise Tailoring on the Effectiveness of SNE vs AEI
Subgroup analysis was only possible for short- to medium-term effects because of the limited number of studies with long-term follow-up assessments. The subgroup analysis comparing studies that used motor control impairment as an inclusion criterion versus those that did not is shown in Figures 5 and 6. The effectiveness of SNE versus AEI was no greater in studies in which participants were selected according to a predefined criterion of motor control impairment (P = .78 for pain intensity; P = .46 for disability). In studies that used neck motor dysfunction as an inclusion criterion, the heterogeneity was low (I2 = 24% for pain; I2 = 0% for disability), whereas in the other subgroup it was high (I2 = 79% for pain; I2 = 82% for disability).
Effects of specific neck exercises versus alternative exercise interventions on pain intensity in the short to medium term (4–12 weeks) when a subgroup analysis was performed to compare studies that used exercise tailoring with those that did not.
Effects of specific neck exercises versus alternative exercise interventions on disability in the short to medium term (4–12 weeks) when a subgroup analysis was performed to compare studies that used exercise tailoring with those that did not.
The subgroup analysis comparing studies that used exercise tailoring versus studies that did not tailor the exercises to the individual participant is shown in Figures 5 and 6. SNE was not more effective than AEI in studies that used some form of exercise tailoring (P = .80 for pain intensity; P = .30 for disability). For studies that used exercise tailoring, heterogeneity was moderate for pain (I2 = 62%) and for disability (I2 = 59%); for studies that did not use exercise tailoring, heterogeneity was high for pain intensity (I2 = 86%) and for disability (I2 = 87%).
Discussion
This systematic review assessed the effectiveness of neck exercise for the treatment of NSNP. Studies reported improvements in pain and disability in both SNE and AEI, providing evidence of the usefulness of exercise to treat NSNP. Results of the meta-analysis indicate that in the short to medium term, SNE are more effective than AIE and therefore may be used preferentially. However, the pooled effect size was only small, suggesting that SNE may not be considerably superior to other forms of exercise for relieving pain or reducing disability.
The observed small effect size may be due to a lack of adequate participant selection. The majority of studies84–89,92,93 did not apply any form of criteria for muscle dysfunction to select participants. It would be logical to expect that SNE are more effective for people presenting with motor control deficits. However, subgroup meta-analysis showed that studies that used specific criteria did not obtain a better outcome with SNE than those that did not, although only 4 studies83,90,91,94 could be included in the former subgroup. It is important to note that despite many tests of muscle dysfunction being described in the literature,59,96 these 4 studies83,90,91,94 used mostly the CCFT in isolation. Current evidence suggests that the isolated use of the CCFT to select participants for SNE does not improve treatment outcome. A more comprehensive series of tests may be required to assess muscle dysfunction for the appropriate selection of participants for SNE, given the complexity and diversity of motor control impairments that have been documented in people with neck pain.45–49,51,53–55,97 The inclusion of tests that assess muscle dysfunction during functional movements that reproduce patient’s symptoms and in different movement planes may improve patient selection for greater effectiveness of SNE.
Although the majority of studies83,84,86,88,89,91,93,94 attempted to tailor exercises to participants, the subgroup meta-analysis showed that the strategy used was not effective at increasing the effectiveness of SNE. Again, all studies that used exercise tailoring83,84,86,88,89,91,93,94 employed performance during the CCFT to tailor exercises to the individual participant and progress them accordingly. Evidence suggests that this approach may not have added value if used in isolation. Clinical guidelines should acknowledge that current evidence does not support the use of these tests in isolation for the subclassification of patients for most effective treatment with SNE. A combination of the CCFT with other forms of assessment to aid exercise tailoring and progression may be required to enhance the effectiveness of SNE.
For exercises to be effective, they have to be performed consistently. Despite the importance of adherence in studies that assess the effectiveness of exercise interventions, this criterion was met by only 2 studies.83,92 Previously,67 a dose-effect relationship of exercise has been established. It is difficult to conclude on the effectiveness of either SNE or AEI interventions when exercise adherence is not clearly monitored, as was the case in the majority of studies. Furthermore, most studies reported only the results of short- to medium-term follow-up assessments. Only 3 studies86,87,89 included longer-term follow-up assessments; 2 studies86,89 reported beneficial effects in both SNE and AEI groups but no differences between interventions, whereas 1 study87 reported considerable greater effectiveness of SNE. This discrepancy in the findings was reflected in the high heterogeneity (I2 ≥ 95%) noted in the meta-analysis, which found no long-term differences in effectiveness between interventions. Nevertheless, the total number of participants was small, and no clear conclusion regarding long-term effects can be drawn.
The overall quality of the evidence was low, as demonstrated by the methodological quality assessment. Five of 12 studies85,87,88,90,92 were classified as having a high risk of bias because of the absence of concealed allocation,85–88,90,92,93 no patient and assessor masking,85,87,88,90,92 lack of adequate management and reporting of dropouts,84,85,87,88,90 and nonavoidance of cointerventions,87–90,92 among others. Their contribution to the results of the meta-analysis was considerable. Although the meta-analysis that included all studies showed significant effects in favor of SNE for both pain and disability in the short to medium term (albeit the effect size was small), when studies with a high risk of bias were excluded from the meta-analysis, differences were no longer significant for disability. It should be noted that the largest effect sizes in pain and disability in both the short to medium term and the long term were reported by 2 studies with a high risk of bias,85,87 both in favor of SNE. The absence of good quality evidence was also reflected in the findings from the GRADE assessment, which concludes that the certainty regarding the superior effectiveness of SNE over AEI in the short to medium term is low, and very low certainty exists regarding longer-term effects. In addition to the high risk of bias shown by some studies,85,87,88,90,92 GRADE assessment also revealed concerns regarding the consistency of the findings. There was heterogeneity in the results of the studies, which may have been caused by the varied quality of their research methodology or differences in the exercise dosage prescribed. The latter ranged from 1 session per week to 3 sessions per week. Furthermore, some studies also included home exercise programs, whereas others used only face-to-face sessions.
To our knowledge, this is the most comprehensive systematic review of the effectiveness of SNE compared with AIE in NSNP. Additionally, no previous systematic review has reported long-term effects or assessed the effect of participant selection criteria or exercise tailoring on the effectiveness of SNE; therefore, to our knowledge, our review is the first to report on these. Our findings of superior effectiveness of SNE in the short to medium term (albeit with a small effect size) differ from those of 3 previous systematic reviews60–62 that reported no difference between interventions. Although our meta-analysis included 12 studies, previous systematic reviews included only 3 studies,62 5 studies,60 and 7 studies61 in their meta-analysis. Tsiringakis et al63 also concluded that SNE are more effective than AIE but limited their review to only studies that used the pressure biofeedback unit, even though this device is not essential for performing SNE.59,64 Our findings and those of Tsiringakis et al63 differ, however, on the quality of the evidence available. Four studies83,89,91,94 that they classified as high quality were classified as having a moderate risk of bias in our review, and 4 studies85,88,90,92 that they reported as having fair quality were classified as having high risk in the present study. This was also the case for the review by Zoete et al61; for 3 studies89,93,94 that they classified as high quality (PEDro score = 8/11), our assessment indicated a moderate risk of bias, and 2 studies87,88 with moderate quality (PEDro score = 6/11) had a high risk of bias. Differences are most likely related to the methodological quality assessment tool employed. In previous reviews,61,63 the PEDro tool98 was used; this tool does not assess whether there was risk of selective reporting, cointerventions were avoided, or exercise compliance was acceptable. The PEDro tool has been shown to underestimate the risk of bias compared with the Cochrane tool.99
Clinical practice guidelines8,10,12,13 and systematic reviews71–74,100,101 suggest that multimodal treatments involving exercise, manual therapy, and education should be used for most effective management of NSNP. Evidence shows that both AIE and SNE are effective for reducing pain and disability in NSNP. Studies support the preferential use of SNE against other forms of exercise for the short to medium term; hence, those developing clinical guidelines may want to consider advocating its preferential use. However, the size of the effect is small, and the findings are based to some extent on studies with a high risk of bias. Despite the widespread use of SNE,56,59,102,103 studies that include long-term outcomes comparing SNE and AEI are scarce. More specific participant selection and exercising tailoring may be the key to achieving higher effectiveness of SNE and larger effect sizes compared with other exercise interventions, although currently used methods appear to offer little advantage. Evidence suggests that the use of the CCFT to select patients and tailor SNE interventions provides no benefit. Other tests or criteria of muscle dysfunction should be explored to select appropriate patients for SNE so that this intervention can be provided to those with the greatest likelihood of successful outcome. In addition to muscle dysfunction tests, the presence and influence of kinesiophobia, catastrophization, or dysfunction of pro- and antinociceptive mechanisms104 on treatment outcomes should be explored.
Limitations
The findings regarding the effect of exercise tailoring and the presence of motor control dysfunction on treatment outcome are based on an indirect analysis of study findings by means of subgrouping in the meta-analysis. It is possible that other unknown confounding factors may have affected the results in this analysis. Current guidelines advocate the use of a multimodal treatment approach for NSNP; however, studies included in this review assessed the effect of exercise in isolation. It is possible that the effect of exercise may differ when it is used in combination with other interventions. The large majority of studies included in this review considered only short- to medium-term effects; hence, the findings regarding long-term effects should be considered with caution.
GRADE and methodological quality assessments revealed a low quality of evidence, mostly limited to flaws in study design and small sample sizes.
SNE and AIE are effective at reducing pain and disability in NSNP, although the preferential use of SNE may be recommended to achieve better short- to medium-term outcomes. The use of the CCFT alone to select patients and tailor SNE exercises is not supported by the evidence and therefore cannot be recommend. The quality of the evidence is low, and therefore further high-quality studies assessing the effectiveness of SNE compared with other forms of exercise are required. Research that explores alternative methods of patient selection and exercise tailoring are warranted to ensure that SNE are provided to those who are likely to obtain a considerable benefit so that treatment effects are maximized.
Author Contributions
Concept/idea/research design: I. Villanueva-Ruiz, D. Falla, I. Lascurain-Aguirrebeña
Writing: I. Villanueva-Ruiz, D. Falla, I. Lascurain-Aguirrebeña
Data collection: I. Villanueva-Ruiz, I. Lascurain-Aguirrebeña
Data analysis: I. Villanueva-Ruiz, I. Lascurain-Aguirrebeña
Project management: I. Villanueva-Ruiz, I. Lascurain-Aguirrebeña
Providing facilities/equipment: I. Villanueva-Ruiz
Consultation (including review of manuscript before submitting): D. Falla
Funding
There are no funders to report for this study.
Systematic Review Registration
This protocol was registered on the International Prospective Register of Systematic Reviews (PROSPERO No.: CRD42020182180).
Disclosures
The authors completed the ICMJE Form for Disclosure of Potential Conflicts of Interest and reported no conflicts of interest.
Comments