Effectiveness of instrument-assisted soft tissue mobilization and neurodynamic techniques for improving hamstring flexibility and neural mechanosensitivity: A systematic review and meta-analysis

Databases and search strategy

Two reviewers independently searched PubMed, Scopus, physiotherapy evidence database (PEDro), and the Cochrane Library for studies published between January 2010 and March 2025. The search strategy used a combination of medical subject headings terms and relevant keywords, including “instrument-assisted soft tissue mobilization,” “IASTM,” “neurodynamic technique,” “nerve gliding,” “hamstring flexibility,” “hamstring tightness,” straight leg raise (SLR), active knee extension (AKE), “slider technique,” and “tensioner technique.” Boolean operators (“AND,” “OR”) were applied to refine the search results.

Eligibility criteria

Randomized controlled trials (RCTs) involving human participants aged ≥18 years were included if they evaluated the effects of IASTM or NDTs on hamstring flexibility or neural mechanosensitivity in healthy individuals or those with mild musculoskeletal conditions, such as hamstring tightness or knee osteoarthritis. Eligible participants included sedentary, athletic, or recreationally active adults with limited hamstring flexibility or signs of neural mechanosensitivity. Interventions included IASTM methods (e.g., Graston, Gua Sha) and NDTs (e.g., nerve glides, slump stretches), applied alone or in combination with stretching or exercise. Comparison groups could include no treatment, placebo, static stretching, proprioceptive neuromuscular facilitation (PNF), or muscle energy techniques (METs).

Primary outcomes included hamstring flexibility, measured using tests such as the straight leg raise (SLR), active knee extension test, or Sit-and-Reach. Secondary outcomes included neural mechanosensitivity, muscle strength, and pressure pain threshold (PPT). Only full-text, peer-reviewed studies published in English were considered. Exclusion criteria included non-human studies, non-intervention articles (e.g., reviews, case reports), studies unrelated to the interventions or outcomes of interest, or those lacking accessible full text. Both authors independently screened the titles and abstracts, followed by a full-text review. Any disagreements on eligibility were resolved through consensus.

Data extraction

Two reviewers independently extracted the data, each using separate Excel spreadsheets (Microsoft Office 2010). Any discrepancies identified between the two reviewers during data extraction were discussed in detail, and a consensus decision was reached.

Risk of bias and methodological quality

Risk of bias was assessed with the Cochrane (RoB 2.0)^[23] tool across five domains, rated as low, some concerns, or high risk. Methodological quality was assessed using the PEDro scale (0–10), with scores ≥6 considered high quality. The overall certainty of the evidence was evaluated using the GRADE framework. Two reviewers (AT and KB) independently applied the GRADE approach using the GRADEpro Guideline development tool (GDT) software, assessing five domains: Risk of bias, inconsistency, imprecision, indirectness, and publication bias.

Data synthesis and statistical analysis

When two or more studies provided comparable data, a meta-analysis was performed to estimate pooled effects. Continuous outcomes measured on the same scale were reported as the mean difference (MD), while outcomes assessed with different instruments were reported as the standardized MD (Hedges’ g), both with 95% confidence intervals (CIs). A random effect model was used to account for anticipated heterogeneity. The I² statistic quantified heterogeneity (25%, 50%, and 75% representing low, moderate, and high, respectively); when I² ≥ 30%, leave-one-out sensitivity and subgroup analyses (by population type, intervention dose, and follow-up duration) were performed.

Statistical analyses were conducted using RevMan 5.4 (The Cochrane Collaboration) for meta-analyses, forest plots, and funnel plot assessments. Although three RCTs investigated NDTs, pooling was not feasible because of differing primary outcomes (AKE, SLR, sit-and-reach test), timepoints, and incomplete variance data. Therefore, NDT results were synthesized narratively.

Study selection

The initial database search identified 183 records. After removing 62 duplicates, 121 records were screened by title and abstract, and 76 were excluded. Forty-five full-text articles were assessed; nine RCTs met all the inclusion criteria. The process of selecting studies is summarized in the PRISMA flow diagram [Figure 1].

Close

Figure 1:

Preferred Reporting Items for Systematic Reviews and Meta-analyses flow diagram illustrating the study selection process.

Export to PPT

Characteristics of included studies

Table 1 summarizes the characteristics of the included studies.^{[3,9,16,24-29]} The included studies were conducted in countries including the United States, Turkey, Pakistan, India, Iran, and Belgium. Study sample sizes ranged from 18 to 60 participants overall, with group sizes typically between 15 and 35 participants per arm, and participant ages ranging from 18 to 65 years.

Synthesis of findings

All nine studies reported improvements in hamstring flexibility, though effect sizes and outcome measures varied. IASTM generally produced a greater range of motion (ROM) gains than static stretching. For example, Anjum et al. (2023)^[25] found that IASTM improved AKE and reduced pain in knee osteoarthritis more effectively than PNF stretching, while Gunn et al. (2019)^[3] reported that both IASTM and PNF outperformed static stretching in hip flexion. NDTs, particularly sliders, increased pain-free ROM and reduced mechanosensitivity, but effects were often short-lived; a 6-week neural slider program, however, produced sustained improvements (De Ridder et al., 2019).^[16] Comparative studies have shown that IASTM, hold-relax stretching, and MET are all effective for short-term passive SLR gains, although IASTM additionally improves muscle strength and PPT (Maras et al., 2024; Nazary-Moghadam et al., 2023).^[9,27]

Analysis of included studies (AKE outcome)

Data from four studies (n = 191) reporting the AKE outcome were combined using a random-effects model, accounting for expected clinical and methodological differences. The meta-analysis demonstrated a statistically significant improvement in favor of IASTM over stretching or control, with an MD of 6.2° (95% CI: 2.5–9.9; P = 0.001) [Figure 2]. Heterogeneity was moderate (I² = 48%), indicating some variability between studies.

Close

Figure 2:

Forest plot for the active knee extension outcome. Individual study estimates, 95% confidence intervals, and the overall pooled effect (diamond) are shown. Heterogeneity statistics (I² and P-value) are also indicated. SD: Standard deviation, CI: Confidence interval

Export to PPT

Heterogeneity and sensitivity analyses

Heterogeneity for the AKE meta-analysis was moderate (I² = 48%). A leave-one-out sensitivity analysis did not substantially alter the pooled estimate (range MD 5.5°–6.8°) or heterogeneity (I² range 36–55%), suggesting that no single study unduly influenced the results. In subgroup analyses, trials involving repeated IASTM sessions (≥4 sessions) demonstrated larger and more sustained improvements in flexibility compared with single-session interventions; however, subgroup comparisons were underpowered due to the limited number of studies.

NDTs

Three RCTs evaluated the effects of NDTs on hamstring flexibility and neural mechanosensitivity. However, a quantitative meta-analysis was not conducted because the studies reported non-comparable primary outcomes (AKE, passive SLR, and sit-and-reach) and heterogeneous follow-up intervals. In addition, two studies lacked sufficient data (mean ± standard deviation) for effect-size calculation. Consequently, NDT outcomes are described narratively. All three studies demonstrated improvements in hamstring flexibility and pain thresholds following NDT interventions, although the magnitude and duration of effects varied across studies.

Methodological quality within studies

The methodological quality of the included studies, evaluated using the PEDro scale, ranged from 5 to 9 out of 10, indicating moderate to high quality. Detailed PEDro scores for each study are presented in Table 2.^{[3,9,16,24-29]}

Risk of bias within studies

Most included studies demonstrated a moderate risk of selection bias, primarily due to inadequate reporting or unclear procedures for random sequence generation and allocation concealment. Several studies had domains with “some concerns,” primarily related to allocation concealment and blinding of outcome assessors; however, no study was judged as “high risk” overall according to the Cochrane RoB 2.0 tool. Specifically, three trials were rated as low risk overall, and six trials had some concerns [Figure 3]. Common limitations included unclear allocation procedures and a lack of assessor blinding, whereas randomization methods were generally adequate.

Close

Figure 3:

Risk of bias assessment of the included studies using the Cochrane RoB 2.0 tool. Three studies were rated low risk, and six studies had some concerns; none were rated high risk. (a) Summary table showing risk of bias across five domains (D1-D5) for each study. Green (+) indicates low risk, while yellow (−) indicates some concerns. (b) Bar chart depicting the overall distribution of risk of bias across the five domains, with proportions of studies classified as low risk (green) or some concerns (yellow).

Export to PPT

GRADE evidence summary

GRADE assessment indicated moderate certainty evidence supporting IASTM improvements in AKE. Evidence for NDTs was low to very low due to the small sample size, heterogeneity, and short-term follow-up. Other outcomes (pain, PPT, strength, and neural mechanosensitivity) had low or very low certainty. Overall, both IASTM and NDTs produce short-term gains in flexibility, with longer-term benefits primarily seen with repeated IASTM interventions [Table 3].