Diaphragmatic dysfunction and activation interventions for chronic nonspecific low back pain: A systematic review and meta-analysis

Abstract

Background

Chronic nonspecific low back pain (CNLBP) is a leading cause of disability, with diaphragm dysfunction implicated in impaired spinal stability. However, systematic evidence on targeted interventions remains limited.

Objective

To quantify the efficacy of diaphragm-focused interventions on pain and function in CNLBP.

Methods

The systematic review/meta-analysis followed PRISMA-Cochrane guidelines. Analyzed randomized controlled trials (RCTs) assessing diaphragmatic interventions (manual therapy, respiratory training, neuromuscular re-education). Databases were searched until February 2025. Random-effects models and subgroup analyses were applied.

Results

Meta-analysis of 7 RCTs (n = 283) showed diaphragmatic interventions significantly reduced pain (overall SMD = −1.11, 95%CI: −1.52–−0.71, p < 0.01). Diaphragmatic training (SMD = −1.37) and manual therapy (SMD = −1.25) were most effective, while neuromuscular re-education was not significant (SMD = −0.62). For disability, overall SMD was −0.83（95%CI: −1.14–−0.52, p < 0.01, with manual therapy showing the largest effect size (SMD = −1.30), followed by diaphragmatic training (SMD = −0.69) and neuromuscular re-education (SMD = −0.47). Excluding one 4-week DNS study eliminated heterogeneity and increased pain effect to SMD = −1.24. Evidence certainty was low (GRADE).

Conclusion

Diaphragm-targeted interventions improve short-to-midterm pain and function. Future high-quality RCTs are needed to confirm long-term efficacy and standardize protocols.

Keywords

chronic nonspecific low back pain diaphragmatic function meta-analysis systematic review spinal stability

Introduction

Background: Chronic nonspecific low back pain (CNLBP) is the leading cause of disability worldwide, exhibiting annual occurrence rates of 17–94% and lifetime prevalence estimates between 33–84% and a CNLBP prevalence of approximately 23%, of which 11–12% of the population is disabled due to CNLBP.^1,2 Compared to CNLBP with specific identifiable causes, CNLBP may result from the interaction of biological, psychological, and social factors, accounting for about 80% to 90% of all CNLBP cases.³ Despite its complex etiology, lumbar segmental instability is considered one of the core mechanisms,⁴ and dysfunction of local muscles (such as the transversus abdominis, pelvic floor muscles, and diaphragm) is closely associated with reduced spinal stability.⁵ Supported by the study of Silfies et al.,⁶ insufficient lumbar stability can lead to compensatory increases in the activity of other systems, creating a vicious cycle of pain-compensation.

Lumbar stability relies on a closed-loop system of “core muscle co-contraction and intra-abdominal pressure (IAP) regulation, “where the diaphragm serves dual roles in both respiratory and postural control functions, may serve as a key hub for spinal stabilization through the modulation of IAP.^7–9 The generation of IAP depends on the synergistic activation of the diaphragm and transversus abdominis, and its deficiency can increase lumbar load, exacerbating instability.^10–13

Notably, the co-contraction of the diaphragm with the transversus abdominis and pelvic floor muscles can enhance spinal rigidity through the “hydraulic effect”.^14,15 Although such studies have revealed the physiological basis of the diaphragm in spinal stabilization, critical questions remain unresolved, the evidence base remains fragmented and requires systematic synthesis. Several smaller-scale studies have explored the therapeutic effects of diaphragmatic activation in people with CNLBP (e.g., evaluated specific interventions like breathing exercises or manual therapy in CNLBP^8,16,17). However, existing systematic reviews focusing specifically on diaphragmatic dysfunction and its clinical management in CNLBP are limited. Key gaps persist: Existing meta-analyses have predominantly concentrated on determining which specific exercise modalities¹⁸ or manual therapy techniques¹⁹ demonstrate efficacy in improving functional outcomes for people with CNLBP, while the efficacy of distinct types of diaphragmatic activation interventions (e.g., manual therapy, specific breathing training, neuromuscular re-education like DNS) on core outcomes like pain and function has not been rigorously compared through meta-analysis. Furthermore, heterogeneity in intervention protocols (e.g., duration, specific techniques) and outcome measures across existing studies complicates the interpretation of overall effectiveness.

Thus, this meta-analysis seeks to: Quantify the therapeutic efficacy of diaphragmatic interventions on pain and functional outcomes via randomized controlled trial (RCT) analysis. The findings may provide stronger evidence for the potential benefits of diaphragmatic activation in people with CNLBP.

Materials and methods

Study protocol

This meta-analysis was conducted in accordance with the guidelines of the Cochrane Back Review Group (CBRG)²⁰ following the PRISMA 2020 statement.²¹ This systematic review and meta-analysis was prospectively registered on the PROSPERO international prospective register of systematic reviews (registration number: CRD420251030332) on April 29, 2025. The final conduct of this review, including the research question, eligibility criteria (participants, interventions, comparators, outcomes, study designs), search strategy, data extraction items, risk of bias assessment tools, primary outcomes, and planned statistical analysis methods (including the use of random-effects models and subgroup analyses based on intervention type), adhered strictly to the pre-specified protocol registered in PROSPERO. No substantive deviations from the registered protocol occurred during the review process. All data generated or analyzed during this study are included in this manuscript and its supplementary files. The full dataset is available from the corresponding author upon request.

Study design

This study employs a systematic review and meta-analysis design. Analyzes RCTs evaluating the efficacy of diaphragmatic activation interventions on pain and functional outcomes in CNLBP populations. This design approach was selected to determine the therapeutic effects of diaphragmatic activation interventions in people with CNLBP.

Data sources and search strategy

The search timeframe included all entries from database creation until February 13, 2025. The primary databases included the Cochrane Library, PubMed, and Embase. Adopting a strategy combining subject headings (MeSH terms) with free-text terms. The search strategy employed is outlined in Table 1. Study selection was restricted to English-language publications only; no limitations were imposed on study design (covering RCTs, CCSs, and other relevant designs), while coverage was extended to grey literature through supplementary searches of the Cochrane Central Register of Controlled Trials (CENTRAL), ClinicalTrials.gov, and reference lists of identified publications.

Table 1.

Search strategy.

Database	Search terms for all databases	Filters
PubMed Embase Cochrane library	#1 (‘low back pain’ OR ‘lower back pain’ OR ‘lumbago’ OR ‘chronic low back pain’ OR ‘CLBP’ OR ‘mechanical back pain’ OR ‘sciatica’ OR ‘non-specific low back pain’ OR ‘lumbar pain’ OR ‘backache’) #2 MeSH descriptor: [Low Back Pain] #3 (‘diaphragm’ OR ‘phrenic nerve’ OR ‘hiatal hernia’ OR ‘crus of diaphragm’ OR ‘paradoxical breathing’ OR ‘diaphragm flutter’ OR ‘hiccup’) #4 MeSH descriptor: [Diaphragm] #5 (#1 OR #2) AND (#3 OR #4)	Language: English Date: Inception to February 13, 2025

Inclusion and exclusion criteria

To enhance methodological rigor and reproducibility, this systematic review explicitly defined eligibility criteria using the PICOS framework: (1) Population: Adults (≥18 years) with strictly defined CNLBP; pain duration ≥12 weeks, localized between ribs and gluteal folds without specific etiology); studies with mixed acute/chronic populations were included only if separable CNLBP data were provided; (2) Intervention: Diaphragm-targeted therapies (manual techniques, respiratory training emphasizing diaphragmatic activation, or neuromuscular re-education [e.g., DNS]); (3) Comparator: Active controls (conventional therapy, core stabilization) or passive controls (placebo, waitlist); (4) Outcomes: Primary—pain intensity, assessed using the Numeric Pain Rating Scale (NPRS) or Visual Analog Scale (VAS); Functional disability, evaluated through validated instruments including the Roland-Morris Disability Questionnaire (RMDQ), Oswestry Disability Index (ODI), Quebec Back Pain Disability Scale (QBPDS), or Core Outcome Measures Index (COMI); (5) Study design: RCTs assessing interventions in CNLBP versus healthy controls. Non-English studies and animal research were excluded. This structured approach minimized ambiguity, particularly regarding population heterogeneity and intervention specificity.

Literature screening

The literature screening process was conducted on the Rayyan platform and carried out independently by two reviewers (MBQ and XLJH). Initially, titles and abstracts of the retrieved records were screened to exclude duplicate entries and studies that clearly did not meet the eligibility criteria. Subsequently, full-text articles of the preliminarily included studies were thoroughly reviewed to further eliminate research that did not satisfy the inclusion criteria. Any discrepancies encountered during the process were resolved through discussion with a third reviewer (SHJ) until a consensus was reached.

Outcome measures

The outcome measures for the RCT studies were (1) Low back pain intensity measured using either the VAS or NPRS; and (2) functional impairment status assessed employing either the ODI, RMDQ or COMI, to assess whether diaphragmatic intervention alleviates pain and improves function in people with CNLBP.

For the evaluation of functional disability, although COMI differs from other scales in specific items and scoring methods, they collectively measure the same core construct related to functional limitations and quality of life dimensions in patients with chronic low back pain.²² Standardized Mean Difference (SMD) was used to pool these data to eliminate dimensional differences. Therefore, combining COMI with RMDQ and ODI data for meta-analysis is methodologically and conceptually appropriate.

Data extraction

A data collection form was created using Excel. Two independent reviewers (MBQ and XLJH) extracted data using a pilot-tested Excel sheet, resolving discrepancies through discussion, with unresolved disagreements arbitrated by a third reviewer (SHJ). The extracted elements included: lead author, publication year, study methodology, sample size, demographic characteristics (sex, age), intervention parameters (specific technique, duration (weeks), weekly frequency), and outcome variables.

Statistical analysis

Effect size calculation: For continuous outcome measures (e.g., pain scores, disability indices), due to the use of different measurement standards, we calculated SMDs and 95% CIs to integrate data from different evaluation systems. Model selection: effect sizes were pooled using a DerSimonian-Laird random-effects model to accommodate between-study heterogeneity. Heterogeneity was assessed using the I² statistic, we interpreted I² values according to conventional benchmarks: <25% (minimal heterogeneity), 25–50% (moderate), >50% (substantial). If I² ≥ 50%, sensitivity analysis (involving the stepwise exclusion of single studies) and subgroup analysis (grouped by intervention types in the experimental arm) were carried out to verify result robustness and investigate heterogeneity. Given the anticipated clinical heterogeneity arising from variations in intervention protocols (e.g., type, duration, frequency), particularly between approaches like manual therapy/respiratory training and neuromuscular re-education, we planned a priori subgroup analyses based on the primary intervention type employed in the experimental arm (manual therapy, respiratory/diaphragmatic training, neuromuscular re-education/DNS). Intervention duration was also considered as a potential effect modifier, although the limited number of studies within each subgroup constrained formal meta-regression analyses. Funnel plot symmetry was visually inspected to evaluate publication bias. Due to the limited number of studies, funnel plot asymmetry tests should be interpreted with caution.

Software tool: All analyses were performed using RevMan 5.3 (Cochrane Collaboration).

Risk of bias and quality assessment

Methodological quality and risk of bias of included studies was independently assessed by two investigators (MBQ and XLJH). Disagreements were resolved by consensus or third-reviewer arbitration (SHJ). The original Cochrane Risk of Bias tool (RoB 1.0) was used to assess bias across seven domains: sequence generation, allocation concealment, blinding of participants and personnel, blinding of outcome assessment, incomplete outcome data, selective outcome reporting, and other sources of bias. Each domain was judged as “low risk,” “high risk,” or “unclear risk” according to the tool's guidance. Random sequence generation was rated “low risk” if a truly random method was described; “high risk” if non-random methods were used. Allocation concealment was considered “low risk” if adequate methods were used; otherwise, it was rated “high risk” or “unclear risk” if insufficient information was provided. Blinding of participants and personnel was rated “low risk” if blinding was implemented and unlikely to be broken; “high risk” if no blinding or incomplete blinding was present, and the outcome was likely to be influenced; “unclear risk” if information was insufficient to judge. Blinding of outcome assessment was judged “low risk” if outcome assessors were blinded; otherwise, “high risk” or “unclear risk.” Incomplete outcome data were rated “low risk” if missing data were minimal and balanced across groups with reasons provided; “high risk” if missing data were likely to introduce bias. Selective outcome reporting was considered “low risk” if all prespecified outcomes were reported; “high risk” if some key outcomes were not reported. Other sources of bias were evaluated based on factors such as baseline imbalance, funding source, or other specific biases. The overall risk of bias for each RCT was categorized based on the most severe rating across all domains, following the conservative approach recommended by the Cochrane Handbook.

The certainty of evidence for the primary outcomes—pain intensity and functional disability in the RCTs—was assessed using the GRADE approach. The initial rating was set as “high” for RCTs based on study design. The evidence was rigorously evaluated across five domains: (1) Risk of bias, assessed using RoB 1.0 for RCTs; the evidence was downgraded by one level if the majority of the included studies had serious limitations. (2) Inconsistency, evaluated not only by the I² statistic (>50% indicating substantial heterogeneity) but also by considering the direction and magnitude of effects, overlap of confidence intervals confidence intervals, and clinical heterogeneity (e.g., differences in intervention types or population characteristics). (3) Indirectness, determined by comparing the relevance of the study PICOS elements to the clinical question of this review; downgrading occurred if important differences existed in populations, interventions, comparators, or outcomes. (4) Imprecision, downgraded if the 95% confidence interval of the pooled effect estimate crossed either the line of no effect or a predetermined minimal clinically important difference (MCID; e.g., a reference value of 1.5 points for pain intensity); and (5) Publication bias, assessed primarily through visual inspection of funnel plot symmetry and supplemented by Egger's test (p < 0.05 suggestive of potential bias); however, in cases with a limited number of included studies (e.g., <10), results were interpreted cautiously without relying solely on statistical tests. The overall certainty of evidence could be downgraded if serious concerns were identified in any domain, and was categorized as “High,” “Moderate,” “Low,” or “Very low.”

Results

Literature screening results

Systematic searches were conducted across multiple databases (from inception to March 2025) and supplemented by manual reference screening. This process identified seven eligible RCTs (Figure 1) examining the impact of diaphragmatic activation in patients with low back pain. These RCTs evaluated the effects of diaphragmatic activation interventions on pain and function in people with CNLBP (providing the primary source of evidence).

Figure 1.

The interventional study section follows the flow chart of the PRISMA systematic review guidelines.

Study characteristics

Table 2 summarizes the key characteristics of the included randomized controlled trials. RCT section participants: The seven eligible RCTs^{8,16,17,23–26} enrolled a total of 283 people with CNLBP, with individual study sample sizes varying between 22 and 66 participants. Participants were predominantly aged 18–60 years. All included studies enrolled patients meeting standardized diagnostic criteria for CNLBP (persisting ≥3 months). The VAS and Numeric Rating Scale (NRS) served as pain intensity measures, with eligible participants typically demonstrating moderate-to-severe pain (scores 3–7) at baseline. One study targeted amateur athletes (n = 24),¹⁶ requiring weekly exercise frequency of 2–4 sessions. Owing to the limited total RCT sample size (<400), the GRADE guidelines²⁷ indicate a risk of imprecision in these outcomes; consequently, results should be interpreted in conjunction with sensitivity analysis.

Table 2.

Characteristics of 7 randomized controlled experiments.

Author	Year	Research Type	Sample Size	Age (Mean ± Sd）	Gender (Male/ Female)	Intervention	Duration And Weekly Frequency	Control	Outcome
Otadi	2021	RCT	N = 24	35.2 ± 9.73	12/12	Diaphragmatic training + electrical stimulation	4 weeks, 3 sessions/week	Electrical stimulation	Pain intensity and Disability severity
Vicente-Campos	2021	RCT	N = 40	23.5 ± 6.04	Unavailable	HAG	8 weeks, 2 sessions/week	Not	Pain intensity and Disability severity
Huanjie	2018	RCT	N = 60	38.3 ± 7.91	15/45	DNS	4 weeks, 3 sessions/week	Core stability	Pain intensity and Disability severity
Masroor	2023	RCT	N = 22	25.4 ± 3.73	Unavailable	Breathing training + Core stability	4 weeks, 3 sessions/week	Core stability	Pain intensity and Disability severity
Siglan	2022	RCT	N = 42	38.38 ± 8.3	20/22	DMT	4 weeks, 1 sessions/week	Placebo	Pain intensity and Disability severity
Salvador	2018	RCT	N = 66	42.55 ± 10.51	29/37	OMT	4 weeks, 5 sessions/week	Not	Pain intensity and Disability severity
Rabieezadeh	2024	RCT	N = 29	40.31 ± 6.33	10/19	DNS	8 weeks, 3 sessions/week	Not	Pain intensity and Disability severity

Table 3 summarizes the stratification of the RCT section into three subgroups according to distinct therapeutic mechanisms, as follows: Manual therapy refers to interventions primarily involving hands-on techniques applied directly to the diaphragm or related structures (e.g., osteopathic manipulative treatment, myofascial release) to improve diaphragmatic mobility and function (classification criteria: practitioner-applied manual techniques targeting the diaphragm). Both diaphragmatic training and neuromuscular re-education involve “breathing exercises,” yet they differ in theoretical foundations theoretical foundations and implementation focus. Diaphragmatic training is more localized, emphasizing direct training of diaphragmatic function, whereas neuromuscular re-education adopts a more holistic approach, focusing on integrating the diaphragm's postural stabilizing role within overall movement patterns. For instance, in Dynamic Neuromuscular Stabilization (DNS), breathing is regarded as one key component for initiating core stability but not as the sole objective. The primary emphasis lies in restoring neurocontrol strategies according to physiological developmental kinesiology sequences, rather than merely enhancing isolated diaphragmatic function. Specifically, if a study primarily describes “breathing exercises/training performed by patients” with the main goal of improving breathing patterns or diaphragmatic mobility, it should be categorized as “diaphragmatic training.” Conversely, if a study explicitly references a developmental kinesiology framework (such as DNS) and emphasizes the diaphragm's role as a postural muscle, it should be classified under “neuromuscular re-education.”

Table 3.

Subgroup information classified by treatment mechanism.

Author	Intervention Description	Subgroup	Classification Basis
Siglan	DMT	Manual Therapy	Practitioner-applied manual techniques targeting the diaphragm
Salvador	OMT	Manual Therapy	Manual therapy applied to diaphragm-related structures
Otadi	Diaphragmatic Training + Electrical Stimulation	Diaphragmatic Training	Patient-performed breathing exercises emphasizing diaphragmatic activation
Vicente-Campos	HAG	Diaphragmatic Training	Focused breathing technique to enhance diaphragmatic control
Masroor	Breathing Training + Core Stability	Diaphragmatic Training	Combined respiratory and core exercises targeting diaphragmatic contraction
Huanjie	DNS	Neuromuscular Re-education	Developmental kinesiology-based protocol integrating diaphragmatic postural function
Rabieezadeh	DNS	Neuromuscular Re-education	Comprehensive neuromuscular reprogramming including diaphragmatic stabilization

Note: OMT, Osteopathic manipulative treatment; HAG, Hypopressive abdominal gymnastics; DNS, Dynamic neuromuscular stabilitation; DMT, Diaphragm myofascial techniques

Furthermore, (Table 4) within the neuromuscular re-education subgroup, due to its multidimensional nature: although both included DNS intervention studies explicitly required diaphragmatic activation, only one employed an objective method to specifically assess diaphragmatic recruitment (evaluating diaphragm thickness). This may lead to an inability to attribute the effects solely to diaphragmatic mechanisms, making it susceptible to influence from other confounding factors.

Table 4.

Introduction to the intervention method of neuromuscular re-education subgroup.

Author (Year)	Intervention Description	Diaphragm Activation Requirement*	Independent Diaphragm Assessment	Current Subgroup
Huang (2018)	DNS	Yes (Emphasize activating the diaphragm)	Yes (Assessment of Diaphragm Thickness)	Neuromuscular re-education subgroup
Rabieezadeh (2024)	DNS	Yes (Emphasize activating the diaphragm)	No (Diaphragm function not assessed)	Neuromuscular re-education subgroup

Note: *Both studies emphasized the role of the diaphragm in postural control and integrated the coordinated activation of the diaphragm, transversus abdominis, and pelvic floor muscles during functional activities.

Characteristics of interventions

The intervention characteristics of the RCT section: All seven included RCTs targeted CNLBP, with intervention durations ranging from 4 to 8 weeks: predominantly 4 weeks, as in five studies,^{16,17,23–25} while only two studies^8,26 employed an 8-week duration, in the two studies investigating DNS—the intervention durations were 4 weeks²³ and 8 weeks.²⁶ The frequency was 2–3 sessions per week. The experimental interventions were categorized into three types: Manual activation interventions,^17,25 diaphragmatic training interventions,^8,16,24 and neuromuscular re-education interventions.^23,26 Specific interventions included: diaphragmatic training and TENS,¹⁶ DNS training,^23,26 low-pressure abdominal gymnastics (HAG),⁸ diaphragmatic breathing and core training,²⁴ osteopathic manipulative treatment (OMT),¹⁷ and myofascial release (MFR).²⁵ Control group designs: Two studies^23,24 employed conventional training interventions, one¹⁶ used TENS intervention, one²⁵ utilized a placebo intervention, while three studies^8,17,26 had no intervention. The primary goal of these interventions was to activate diaphragmatic function and improve core stability.

Outcome characteristics

A comprehensive analysis of the seven studies^{8,16,17,23–26} revealed that diaphragmatic-related interventions for CNLBP demonstrated significant effects in pain relief, functional improvement, muscle activity, and respiratory function. Pain assessment: Most studies used the NPRS,^8,16,24,25 three employed the VAS,^17,23,26 and the Short-Form McGill Pain Questionnaire (SF-MPQ) was utilized in one trial.¹⁷ For dis1ability evaluation, the following measures were implemented: The majority adopted the ODI^17,23–26 and RMDQ,^8,23–25 while one study¹⁶ independently used the Core Outcome Measures Index (COMI).

Risk of bias

Risk of bias was evaluated separately for RCTs (Figure 2), with study-level domain risks summarized in Figures 3. Across seven RCTs, overall risk of bias was moderate/unclear due to inadequate methodological reporting: Three trials^23,24,26 were rated “higher risk.” Key issues included poor allocation concealment (potential selection bias in four studies^16,24–26), insufficient blinding of outcome assessment (detection bias in two studies^16,24), and incomplete outcome data handling (attrition bias in two studies,^24,26 notably: 19.4% dropout without intention-to-treat analysis). Performance bias (inherent unblinding) affected five studies.^16,23–26 Other concerns encompassed retrospective trial registration,^16,26 convenience sampling, and adherence monitoring gaps.

Figure 2.

Risk of bias assessment results for the RCT section using the Cochrane RoB 1.0.

Figure 3.

Percentage of RCTs rated as low risk of bias across individual domains.

Primary outcome measures

Pain improvement data

Seven RCTs^{8,16,17,23–26} were included (total sample=283: intervention group=140, control group=143). The pooled analysis demonstrated a significant overall effect size, the SMD was −1.11 (95% CI: −1.52 to −0.71, p < 0.01), with the confidence interval not crossing zero. The absolute SMD >1 indicated a large effect size. Moderate heterogeneity was observed (I²=58%, p = 0.03).

Subgroup analysis (Figure 4): Manual therapy subgroup (2 studies,^17,25 n = 108), the analysis demonstrated a significant effect size (SMD = −1.25, 95% CI: −1.89 to −0.62, p < 0.01), with moderate between-study heterogeneity observed (I² = 54%, p = 0.14). Diaphragmatic training subgroup (3 studies,^8,16,24 n = 86), the meta-analysis revealed a large effect size (SMD = −1.37, 95% CI: −1.85 to −0.89, p < 0.01), with no heterogeneity (I² = 0%, p = 0.49), indicating consistent evidence for pronounced analgesic effects (large effect size). Neuromuscular re-education subgroup (2 studies,^23,26 n = 89), the pooled analysis showed a non-significant effect size (SMD = −0.62, 95% CI: −1.39 to 0.15, p = 0.11). Considerable between-study heterogeneity was observed (I² = 64%, p = 0.10). Inter-subgroup differences were non-significant (I² = 24.0%, p = 0.27). However, diaphragmatic training (SMD = −1.37) yielded a markedly larger effect than neuromuscular re-education (SMD = −0.62).

Figure 4.

Subgroup analysis results and funnel plot of pain improvement in different intervention types in the experimental group.

Sensitivity analysis (Figure 5): Sequential exclusion of six studies^{8,16,17,24–26} did not alter pooled results. Notably, removing the study by Huanjie et al.²³ (utilizing 4-week DNS) eliminated heterogeneity (I²=0%, p = 0.54 vs. original 58%, p = 0.03), yielding a more consistent and substantially larger outcome (SMD = −1.24; 95% CI: −1.54 to −0.95, p < 0.01), reflecting an 11.7% enhancement over the initial estimate.This strongly suggests that the 4-week DNS intervention employed by Huanjie et al.²³ was a primary source of heterogeneity and yielded a smaller effect size compared to other interventions and the longer-duration DNS study.

Figure 5.

Sensitivity analysis results and funnel plot of pain improvement after excluding studies (Huanjie et al.).

Disability improvement data

The total sample comprised 283 cases (140 in the experimental group and 143 in the control group). The analysis revealed a moderate-to-large effect size (SMD = −0.83, 95% CI: −1.14 to −0.52, p < 0.01), with an absolute SMD > 0.8, indicating a large effect size, with moderate heterogeneity present (I² = 36%, p = 0.16).

Subgroup analysis (Figure 6) revealed that manual interventions (2 studies;^17,25 n = 108), the analysis demonstrated a large and highly significant effect size (SMD = −1.30, 95% CI: −1.72 to −0.88, p < 0.01), with no observed heterogeneity (I² = 0%, p = 0.71), indicating a large effect size (SMD > 1.2). The diaphragmatic training intervention subgroup (three studies^8,16,24 with a sample size of 86) showed an effect size (SMD = −0.54, 95% CI: −0.87 to −0.21, p < 0.01), with no heterogeneity (I² = 0%, p = 0.85), and a moderate effect size (SMD ≈ 0.5). The neuromuscular re-education intervention subgroup (two studies,^23,26 n = 89), the analysis revealed a small-to-moderate effect (SMD = −0.47, 95% CI: −0.89 to −0.05, p = 0.03), also with no heterogeneity (I² = 0%, p = 0.38). Inter-subgroup differences were statistically significant, with high heterogeneity (I² = 75.3%, p = 0.02). The results suggest that all subgroups showing that experimental interventions were significantly superior to the control group (p < 0.05). The effect sizes were ranked as follows: manual intervention (SMD = −1.30) > diaphragmatic training (SMD = −0.69) > neuromuscular re-education (SMD = −0.47).

Figure 6.

Subgroup analysis results and funnel plot of disability improvement in the experimental group based on different intervention types.

Sensitivity analysis (Figure 7): Except for the individual exclusion individual exclusion of the study by Huanjie et al.,²³ the removal of each of the other six studies separately did not significantly alter the pooled heterogeneity (I² = 19–46%). After excluding the study by Huanjie et al.,²³ the analysis demonstrated a large effect size (SMD = −0.97, CI: −1.25 to −0.69, p = 0.44), representing a 14% absolute increase compared to the original analysis (SMD = −0.83 when including Huanjie et al.²³). The confidence interval narrowed, with the range decreasing from 0.62 in the original analysis to 0.56, indicating improved precision. After exclusion, heterogeneity was substantially reduced (I² = 0% vs. original 36%). This finding further supports the notion that the shorter duration and/or specific protocol of the DNS intervention in Huanjie et al.²³ contributed to heterogeneity and a comparatively smaller effect size in the disability outcome.

Figure 7.

Sensitivity analysis results and funnel plot of disability improvement after excluding studies (Huanjie et al.).

Evidence quality of included studies

Using the GRADE system (Table 5), seven randomized controlled trials^{8,17,20,21,24,25,28} (total sample size n = 283) incorporating assessments of pain intensity (VAS/NPRS) and functional disability (ODI/RMDQ) were included. The starting evidence quality was high (due to RCT design) but was downgraded by one level each for: Risk of bias (small sample sizes, inadequate blinding); inconsistency (significant heterogeneity in interventions leading to substantial variation in efficacy); imprecision (wide 95% confidence interval of the pooled effect size). Consequently, the evidence was finally rated as low quality.

Table 5.

GRADE evidence for diaphragmatic activation in pain intensity and functional disability.

Key Outcomes	Number Of Studie	Risk Of Bias^∗	Inconsistency ^†	Indirectness Of Evidence^#	Imprecision^&	Publication Bias^	Quality Of Evidence	SMD (95% CI)
Pain Intensity	7 RCTs (n = 283)	Serious	Serious (i² = 58%)	Not serious	Serious	Not serious	LOW	SMD −1.11 (95% CI: −1.52 to −0.71)
Functional Disability	7 RCTs (n = 283)	Serious	No (i² = 36%)	Not serious	Serious	Not serious	LOW	SMD −0.83 (95% CI: −1.14 to −0.52)

Note: SMD = standardized mean difference. ∗ “No serious” = most information is from results at low risk of bias; “Serious” = Inadequate generation of the random sequence, insufficient allocation concealment, lack of blinding of participants and personnel, incomplete outcome data, selective outcome reporting, or other potential sources of bias. † “No serious” = I² ≤ 50%; “Serious” = I² > 50%. # No indirectness of evidence was found in any study. & Based on sample size. “No serious” = n ≥ 400 subjects; “Serious” = n < 400 subjects. ^Based on funnel plots. No publication bias was found.

Based on the funnel plot results presented in Figures 4 and 6, the Egger's tests for the RCT section indicated no significant publication bias. The regression intercepts approached zero (suggesting symmetry in the funnel plots), with p > 0.05. Furthermore, the limited number of studies included in each funnel plot analysis may have resulted in limited statistical power for these tests.

Discussion

Pain improvement: After treatment based on diaphragmatic activation, people with CNLBP experienced significant pain reduction, demonstrating advantages over other treatments. The primary mechanism by which diaphragmatic activation alleviates pain is the “hydraulic effect” achieved through coordinated contraction of the diaphragm, transversus abdominis, and pelvic floor muscles. The diaphragm and pelvic floor act as two pistons pushing against each other, while abdominal muscle contraction resists the lateral movement of intra-abdominal contents, increasing IAP.²⁹ The augmentation of IAP strengthens lumbar rigidity, thereby offloading vertebral discs and facet joints and resulting in pain reduction. The process involves functional diaphragm activation while decreasing compensatory overload of superficial stabilizers (erector spinae, quadratus lumborum), which prevents regional muscle fatigue and subsequent pain development. Inyoung et al.³⁰ found that diaphragmatic stretching reduced trunk muscle activity, indicating that relieving diaphragmatic fatigue can alleviate compensatory high-tension states in local muscles. Since individuals with recurrent CNLBP exhibit reduced perception of back proprioceptive signals during upright posture, the absence of proprioceptive information consistent with motor intention activates cortical centers to monitor sensory discrepancies, leading to pathological pain.^31,32 Diaphragmatic activation training enhances the perception of back proprioceptive signals during postural control, sufficiently stimulating muscles, joints, and tendons to strengthen proprioception in people with CNLBP.³³ Compared to healthy controls, CNLBP individuals exhibit greater diaphragmatic fatigability,¹² which can be improved through targeted training or manual therapy to enhance diaphragmatic endurance. Clinical studies have also confirmed that resistance training effectively improves muscle oxidative capacity,^34,35reducing metabolic reflex-induced inhibition of peripheral muscles.

Disability improvement: All three diaphragmatic activation interventions significantly reduced disability severity in CNLBP individuals. Among them, manual therapy yielded the most pronounced effects, while DNS was slightly less effective than the other two approaches. The improvement mechanism primarily stems from the fact that people with CNLBP, due to impaired diaphragmatic function, often alter their muscle activity or even become inactive to mitigate perceived pain. The erector spinae and other core muscles may overwork to maintain stability, leading to asymmetric stabilization,²⁸ which can impair movement quality and limit functional task performance. Thus, if diaphragmatic function is compromised, unpaired movement patterns emerge,^28,36 ultimately significantly affecting the patient's functional disability status. One study³⁷ indicated that functional diaphragm training can improve running economy and prolong endurance exercise time, thereby suggesting that high-quality functional diaphragm training may alleviate functional impairments.

Comparison with previous studies: The large effect sizes observed in our analysis for manual therapy (SMD = −1.25) and diaphragmatic training (SMD = −1.37) are broadly consistent with prior systematic reviews in this field. For instance, Zhai et al.³⁸ reported that respiratory exercises significantly reduced low back pain (SMD = −0.87) and improved functional disability (SMD = −0.79), particularly when the total intervention duration exceeded 500 min. Similarly, Shi et al.³⁹ found positive effects of breathing exercises on pain (MD = −0.50 for VAS) and function (MD = −2.46 for ODI). However, our study provides a more granular perspective by directly comparing distinct categories of diaphragm-targeting approaches. Our subgroup analysis reinforces the importance of intervention type—a dimension not extensively explored in earlier reviews. This refined categorization helps explain the heterogeneity in pooled estimates and underscores that how the diaphragm is engaged (e.g., via manual release vs. resistive breathing) may differentially influence outcomes. This contrasts with the broader inclusive approach taken by Zhai et al.,³⁸ which integrates various breathing techniques without independently analyzing diaphragmatic activation.

Limitations (heterogeneity): Subgroup and sensitivity analyses were conducted for the included RCTs to explain the heterogeneity in the outcomes of pain improvement and functional disability. Subgroup analysis stratified by confounding factors (different intervention methods in the experimental groups) showed that for pain improvement, the test for between-subgroup differences did not reach statistical significance (I² = 24%, p = 0.27), indicating insufficient evidence that the type of intervention modifies the analgesic effect. However, the effect sizes for both the diaphragmatic training subgroup and the manual therapy subgroup were larger than that of the neuromuscular re-education subgroup. This suggests that intervention strategies directly targeting the activation or release of the diaphragm or related structures may yield more consistent and pronounced analgesic effects in the short term. The manual therapy subgroup exhibited higher heterogeneity, which is likely attributable to differences in treatment frequency and total duration. For instance, Marti-Salvador et al.¹⁷ administered treatment twice a week for a total of 5 sessions, whereas Siglan et al.²⁵ conducted three sessions per week for a total of 12 sessions. Variations in treatment protocols (frequency and number of sessions) may have contributed to the divergent results. Substantial heterogeneity was observed within the neuromuscular re-education subgroup analysis. Sensitivity analysis indicated that the study by Huanjie et al.²³ was the primary source of this heterogeneity, mainly due to an insufficient intervention duration. DNS aims to reprogram dysfunctional movement patterns by promoting neuromuscular coordination and core stability through principles of developmental kinesiology.^23,26,34,40 This process of neuromotor learning and functional integration is inherently time-dependent, relying on cortical plasticity to establish new motor engrams.⁴⁰ The contrasting outcomes within the neuromuscular re-education subgroup—where the 8-week program²⁶ demonstrated a significant effect (SMD = −1.09 for pain after removal of the 4-week study), while the 4-week program²³ did not—strongly support this notion. Previous research also indicates that significant neuromuscular adaptations and pain reduction typically require intervention durations exceeding 6 weeks.^41,42 Therefore, the shorter 4-week duration in Huanjie et al.²³ was likely insufficient to elicit the full therapeutic potential of DNS, contributing to its smaller effect size and the observed heterogeneity. Furthermore, although the incorporated DNS studies all emphasize diaphragmatic activation, their essence lies in integrative neuromuscular re-education rather than diaphragm-targeted intervention. Their effects are more likely to result from the reconstruction of overall movement patterns (e.g., diaphragmatic activation plus core coordination) rather than from a singular diaphragmatic mechanism. This may also be a contributing factor to the observed variability in their effectiveness.

For functional disability improvement, the between-subgroup differences were more definitive (I² = 75.3%, p = 0.02), suggesting that the type of intervention itself may be a key modifier influencing the degree of functional improvement. The manual therapy subgroup showed the largest effect size, followed by the diaphragmatic training subgroup, with the neuromuscular re-education subgroup having the lowest effect size. All subgroups exhibited low within-group heterogeneity. The differential patterns observed in the subgroup analyses for these two outcome measures may stem from variations in the mechanisms and pathways of action of the different interventions.^38,39 Pain improvement may be more readily influenced by the immediate physiological effects conferred by diaphragmatic training and manual therapy, such as increased intra-abdominal pressure enhancing lumbar rigidity, reduced compensatory tension in superficial stabilizer muscles, and possible autonomic nervous system modulation.^{8,16,17,24,25} Consequently, diaphragmatic training and manual therapy can produce consistent and significant analgesic outcomes. In contrast, functional disability is a multidimensional outcome; its improvement depends not only on pain reduction but also on the immediate availability of trunk mobility, soft tissue compliance, and functional movement patterns.⁴³ Manual therapy, by directly enhancing the mobility of the diaphragm and related fascial and articular structures and alleviating myofascial restrictions and kinesiophobia,^17,25 may translate more directly into improvements in daily activity capabilities, thereby exhibiting the largest effect size on disability scales.

In addition to the heterogeneity of interventions, another potentially significant methodological factor attenuating the observed pooled effect size in our meta-analysis was the inclusion of studies utilizing active control groups. Three included RCTs^16,23,24 provided control group participants with interventions possessing inherent therapeutic potential. Compared to studies^8,17,25,26 using placebo or no-treatment controls, the utilization of active controls reduces the net difference between the experimental and control interventions. While this approach is sometimes necessary for ethical reasons or to enhance the pragmatic feasibility of trials, it inherently dilutes the effect size, thereby introducing a source of bias. Furthermore, this meta-analysis found that studies employing active controls generally yielded lower effect sizes for improvements in pain and disability than those using passive controls. For instance, in the study by Huanjie et al.²³ involving DNS, the control group received core stabilization exercise administered three times per week for 30 min per session, an intensity comparable to the experimental DNS intervention (twice per week for 45 min per session), this significantly attenuated the between-group difference.

Limited number of studies and sample size: Only seven RCTs were included in the analysis, with a total pooled sample size of fewer than 400 participants, which increased the imprecision of our effect estimates. Moreover, the number of studies within individual subgroups was also limited, particularly for the manual therapy and neuromuscular re-education subgroups, each of which was based on only two studies. Therefore, the interpretation of results for these two subgroups requires particular caution. The limited number of included studies resulted in imprecise effect estimates, as reflected in wide confidence intervals, and renders the conclusions highly unstable. Consequently, the current evidence for these two types of interventions should be considered preliminary. In contrast, the diaphragmatic training subgroup, which included three studies, provides relatively more robust evidence.

Clinical implications: Clinicians may consider incorporating manual techniques (e.g., OMT, MFR) aimed at improving diaphragmatic mobility, as well as structured diaphragmatic breathing retraining into their treatment plans. These interventions appear to provide robust short-term benefits for pain and function. Regarding target selection, if rapid pain relief is the goal, diaphragm training or diaphragm-related manual therapy may be prioritized; if functional improvement is the primary focus, manual therapy may hold greater advantage, while neuromuscular re-education requires adequate intervention duration and a standardized protocol to more consistently translate into reductions in pain and functional disability. The effectiveness of these targeted interventions underscores the importance of evaluating and treating the diaphragm as a key stabilizer, not merely as a respiratory muscle.

Finally, future research should aim to definitively establish the efficacy of specific diaphragmatic activation therapies (such as DNS) should ideally employ well-designed placebo/sham controls or carefully selected non-inferiority/superiority comparisons with established active treatments, accompanied by adequate trial duration and well-designed independent intervention and assessment of diaphragmatic training within DNS studies. Standardization of both intervention parameters (frequency, duration, progression) and control group design is paramount for generating robust and comparable evidence. Until such studies are conducted, the overall efficacy estimates derived from this meta-analysis should be interpreted with the understanding that they may be conservative, particularly influenced by studies utilizing active controls and shorter intervention durations for neuromotor retraining approaches like DNS. Larger, higher-quality RCTs addressing these methodological limitations are urgently needed.

Conclusion

Diaphragmatic activation interventions (manual therapy, breathing exercises, neuromuscular re-education) can improve pain and function in the short-to-medium term. Manual therapy and breathing exercises demonstrate superior efficacy, while neuromuscular re-education (e.g., DNS) requires ≥8 weeks to achieve significant effects and faces challenges in isolating its impact from other interventions. Although diaphragmatic activation interventions represent a promising clinical pathway, future high-quality randomized controlled trials are warranted to validate long-term outcomes and clarify the mechanisms underlying multifactorial interactions.

Footnotes

Acknowledgments

The authors gratefully acknowledge all primary investigators whose published works were included in this meta-analysis, as their original research provided the essential data foundation for our study. We also extend our appreciation to peer reviewers in this field, whose critical evaluations of prior studies have contributed valuable scholarly insights to our work.

Author contributions

MBQ conceptualized the study design and drafted the manuscript; MBQ, SHJ and XLJH performed study screening and data extraction; MBQ, SHJ and XLJH conducted the risk of bias assessment and methodological quality evaluation; LXF provided senior oversight and manuscript revisions. All authors reviewed and approved the final version.

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Data availability

All data are sourced from references and publicly accessible.

Supplemental material

Supplemental material for this article is available online.

References

Farahbakhsh

Rostami

Noormohammadpour

, et al. Prevalence of low back pain among athletes: a systematic review. J Back Musculoskelet Rehabil 2018; 31: 901–916. PMID: 29945342.

Balagué

Mannion

Pellisé

, et al. Non-specific low back pain. Lancet 2012 Feb 4; 379: 482–491. Epub 2011 Oct 6. PMID: 21982256.

Chiarotto

Koes

. Nonspecific low back pain. N Engl J Med 2022 May 5; 386: 1732–1740. PMID: 35507483.

Panjabi

. Clinical spinal instability and low back pain. J Electromyogr Kinesiol 2003 Aug; 13: 371–379. PMID: 12832167.

Bruno

. The use of “stabilization exercises” to affect neuromuscular control in the lumbopelvic region: a narrative review. J Can Chiropr Assoc 2014 Jun; 58: 119–130. PMID: 24932016; PMCID: PMC4025082.

Silfies

Squillante

Maurer

, et al. Trunk muscle recruitment patterns in specific chronic low back pain populations. Clin Biomech (Bristol 2005 Jun; 20: 465–473. PMID: 15836933.

Calvo-Lobo

Almazán-Polo

Becerro-de-Bengoa-Vallejo

, et al. Ultrasonography comparison of diaphragm thickness and excursion between athletes with and without lumbopelvic pain. Phys Ther Sport 2019 May; 37: 128–137. Epub 2019 Mar 28. PMID: 30954705.

Vicente-Campos

Sanchez-Jorge

Terrón-Manrique

, et al. The main role of diaphragm muscle as a mechanism of hypopressive abdominal gymnastics to improve non-specific chronic low back pain: a randomized controlled trial. J Clin Med 2021 Oct 27; 10: 4983. PMID: 34768502; PMCID: PMC8584898.

Mohan

Paungmali

Sitilerpisan

, et al. Respiratory characteristics of individuals with non-specific low back pain: a cross-sectional study. Nurs Health Sci 2018 Jun; 20: 224–230. Epub 2018 Feb 8. PMID: 29421851.

10.

Neumann

Gill

. Pelvic floor and abdominal muscle interaction: EMG activity and intra-abdominal pressure. Int Urogynecol J Pelvic Floor Dysfunct 2002; 13: 125–132. PMID: 12054180.

11.

Mitchell

Robinson

Brinkman

, et al. Diaphragm fatigue does not impact breathing mechanics or function in people with chronic low back pain. J Pain Res 2025 Feb 26; 18: 963–974. PMID: 40027215; PMCID: PMC11872084.

12.

Janssens

Brumagne

McConnell

, et al. Greater diaphragm fatigability in individuals with recurrent low back pain. Respir Physiol Neurobiol 2013 Aug 15; 188: 119–123. Epub 2013 May 31. PMID: 23727158.

13.

Janssens

Brumagne

Polspoel

. The effect of inspiratory muscles fatigue on postural control in people with and without recurrent low back pain. Spine (Phila Pa 1976) 2010 May 1; 35: 1088–1094. PMID: 20393397.

14.

Miyamoto

Shimizu

Masuda

. Fast MRI used to evaluate the effect of abdominal belts during contraction of trunk muscles. Spine (Phila Pa 1976) 2002 Aug 15; 27: 1749–1755. PMID: 12195066.

15.

Silfies

Squillante

Maurer

, et al. Trunk muscle recruitment patterns in specific chronic low back pain populations. Clin Biomech (Bristol) 2005 Jun; 20: 465–473. PMID: 15836933.

16.

Otadi

Nakhostin Ansari

Sharify

, et al. Effects of combining diaphragm training with electrical stimulation on pain, function, and balance in athletes with chronic low back pain: a randomized clinical trial. BMC Sports Sci Med Rehabil 2021 Mar 4; 13: 20. PMID: 33663607; PMCID: PMC7934526.

17.

Martí-Salvador

Hidalgo-Moreno

Doménech-Fernández

, et al. Osteopathic manipulative treatment including specific diaphragm techniques improves pain and disability in chronic nonspecific low back pain: a randomized trial. Arch Phys Med Rehabil 2018 Sep; 99: 1720–1729. Epub 2018 May 19. PMID: 29787734.

18.

Fernández-Rodríguez

Álvarez-Bueno

Cavero-Redondo

, et al. Best exercise options for reducing pain and disability in adults with chronic low back pain: pilates, strength, Core-Based, and Mind-Body. A Network Meta-Analysis. J Orthop Sports Phys Ther 2022 Aug; 52: 505–521. Epub 2022 Jun 19. PMID: 35722759.

19.

Farra F

Risio

Vismara

, et al. Effectiveness of osteopathic interventions in chronic non-specific low back pain: a systematic review and meta-analysis. Complement Ther Med 2021 Jan; 56: 102616. Epub 2020 Nov 13. PMID: 33197571.

20.

van Tulder

Furlan

Bombardier

. Bouter L; editorial board of the cochrane collaboration back review group. Updated method guidelines for systematic reviews in the cochrane collaboration back review group. Spine (Phila Pa 1976) 2003 Jun 15; 28: 1290–1299. PMID: 12811274.

21.

Page

McKenzie

Bossuyt

, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. Br Med J 2021 Mar 29; 372: n71. PMID: 33782057; PMCID: PMC8005924.

22.

Mannion

Haschtmann

Puhakka

, et al. The core outcome measures Index (COMI) comes of age: 18 years of efficient and effective patient-reported outcome measurement in the field of spinal disorders. Part 1. Development of the COMI, interpretation of scores, and use in research. Eur Spine J 2025 Jul; 34: 2618–2628. Epub 2025 May 21. PMID: 40397148.

23.

Huang

Xie

Zhang

, et al. Effects of dynamic neuromuscular stabilization training on the core muscle contractility and standing postural control in patients with chronic low back pain: a randomized controlled trial. BMC Musculoskelet Disord 2025 Mar 1; 26: 213. PMID: 40025463; PMCID: PMC11871613.

24.

Masroor

Tanwar

Aldabbas

, et al. Effect of adding diaphragmatic breathing exercises to core stabilization exercises on pain, muscle activity, disability, and sleep quality in patients with chronic low back pain: a randomized control trial. J Chiropr Med 2023 Dec; 22: 275–283. Epub 2023 Sep 2. PMID: 38205226; PMCID: PMC10774616.

25.

Sığlan

Çolak

. Effects of diaphragmatic and iliopsoas myofascial release in patients with chronic low back pain: a randomized controlled study. J Bodyw Mov Ther 2023 Jan; 33: 120–127. Epub 2022 Sep 29. PMID: 36775506.

26.

Rabieezadeh

Mahdavinejad

Sedehi

, et al. The effects of an 8-week dynamic neuromuscular stabilization exercise on pain, functional disability, and quality of life in individuals with non-specific chronic low back pain: a randomized clinical trial with a two-month follow-up study. BMC Sports Sci Med Rehabil 2024 Jul 25; 16: 61. PMID: 39054527; PMCID: PMC11271024.

27.

Guyatt

G H

Oxman

A D

Kunz

. GRADE Guidelines 6. Rating the quality of evidence-imprecision. J Clin Epidemiol 2011 Dec; 64: 1283–1293. Epub 2011 Aug 11. Erratum in: J Clin Epidemiol. 2021 Sep;137:265. doi: https://doi.org/10.1016/j.jclinepi.2021.04.014. PMID: 21839614.

28.

Abadi Marand

Noorizadeh Dehkordi

Roohi-Azizi

, et al. Effect of dynamic neuromuscular stabilization on balance, trunk function, falling, and spasticity in people with multiple sclerosis: a randomized controlled trial. Arch Phys Med Rehabil 2023 Jan; 104: 90–101. Epub 2022 Oct 4. PMID: 36206832.

29.

Novak

Jacisko

Busch

, et al. Intra-abdominal pressure correlates with abdominal wall tension during clinical evaluation tests. Clin Biomech (Bristol) 2021 Aug; 88: 105426. Epub 2021 Jul 14. PMID: 34303067.

30.

Inyoung

. The effect of the diaphragm stretching technique in the diaphragm contraction rate and trunk muscle activity in back pain patients. J Back Musculoskelet Rehabil 2024; 37: 225–231. PMID: 37718777.

31.

Anderson

Bliven

KCH

. The use of breathing exercises in the treatment of chronic, nonspecific low back pain. J Sport Rehabil 2017 Sep; 26: 452–458. Epub 2016 Aug 24. PMID: 27632818.

32.

Sakai

Watanabe

Wakao

, et al. Proprioception and geriatric low back pain. Spine Surg Relat Res 2022 Apr 12; 6: 422–432. PMID: 36348676; PMCID: PMC9605761.

33.

Janssens

McConnell

Pijnenburg

, et al. Inspiratory muscle training affects proprioceptive use and low back pain. Med Sci Sports Exerc 2015 Jan; 47: 12–19. PMID: 24870567.

34.

Zemková

Hamar

. Sport-Specific assessment of the effectiveness of neuromuscular training in young athletes. Front Physiol 2018 Apr 11; 9: 264. PMID: 29695970; PMCID: PMC5904431.

35.

Shepherd

S O

Cocks

Tipton

K D

. Resistance training increases skeletal muscle oxidative capacity and net intramuscular triglyceride breakdown in type I and II fibres of sedentary males. Exp Physiol 2014 Jun; 99: 894–908. Epub 2014 Apr 4. PMID: 24706192.

36.

Sannasi

Dakshinamurthy

Dommerholt

, et al. Diaphragm and core stabilization exercises in low back pain: a narrative review. J Bodyw Mov Ther 2023 Oct; 36: 221–227. Epub 2023 Jul 13. PMID: 37949564.

37.

Tong

McConnell

Lin

, et al. “Functional” inspiratory and core muscle training enhances running performance and economy. J Strength Cond Res 2016 Oct; 30: 2942–2951. PMID: 25162653.

38.

Zhai

Zhang

Xia

, et al. The efficiency of respiratory exercises in rehabilitation of low back pain: a systematic review and meta-analysis. J Sport Rehabil 2024 Feb 20; 33: 189–200. PMID: 38377983.

39.

Shi

Liu

Zhou

, et al. Effects of breathing exercises on low back pain in clinical: a systematic review and meta-analysis. Complement Ther Med 2023 Dec; 79: 102993. Epub 2023 Oct 10. PMID: 37827444.

40.

Frank

Kobesova

Kolar

. Dynamic neuromuscular stabilization & sports rehabilitation. Int J Sports Phys Ther 2013 Feb; 8: 62–73. PMID: 23439921; PMCID: PMC3578435.

41.

Kong

Zhang

, et al. Effects of dynamic neuromuscular stabilization/vojta therapy on pain intensity and physical function in individuals with low back pain: a systematic review and meta-analysis. Iran J Public Health 2024 Sep; 53: 1910–1923. PMID: 39429656; PMCID: PMC11490336.

42.

Abadi Marand

Noorizadeh Dehkordi

Roohi-Azizi

, et al. Effect of dynamic neuromuscular stabilization on balance and trunk function in people with multiple sclerosis: protocol for a randomized control trial. Trials 2022 Jan 21; 23: 69. PMID: 35063011; PMCID: PMC8778496.

43.

Lassfolk

Escorpizo

Korniloff

, et al. Linking the spinal function sort and functional capacity evaluation tests to the international classification of functioning, disability and health core set of vocational rehabilitation. J Occup Rehabil 2021 Mar; 31: 166–174. PMID: 32500472.