The Effectiveness of Trigger Point Dry Needling for Musculoskeletal Conditions by Physical Therapists: A Systematic Review and Meta-analysis

Search Strategy

Eligible studies in this systematic review included human subjects with musculoskeletal conditions who had been treated by a physical therapist with dry needling, compared with a control, sham, or other intervention. Only randomized controlled trials were included. Studies were excluded if patients were less than 18 years of age and if the full text was not published in English.

The electronic databases MEDLINE, AMED, CINAHL, and Embase were searched independently by the primary investigator in consultation with a biomedical librarian. The terms “dry needling” or “intramuscular stimulation,” paired with “random,” “group,” “trial,” “randomized controlled trial,” or “controlled clinical trial,” were used to search the electronic databases. Results were limited to human studies. An example of the search strategy is included in the APPENDIX (available at www.jospt.org). Bibliographic reference lists from identified articles were hand searched for any other potential study not identified during the database searches. Search results are displayed in FIGURE 1.

Download FigureDownload PowerPoint

Study Selection

After the duplicate articles retrieved from the different databases were removed, 2 independent reviewers (E.G. and Kelly Lavallee) screened titles and abstracts to determine which studies met the inclusion and exclusion criteria. Studies that appeared to meet the inclusion criteria or whose eligibility could not be determined from the title/abstract screening were retrieved for full-text review by 2 independent reviewers (E.G. and Sebastian Sabadis). Disagreements between reviewers were resolved by consulting a third reviewer (J.C.) who was blind to other reviewers' decisions on whether the study should be included. Once study selection was complete, primary authors of included studies were e-mailed and asked if they were aware of any other studies that would satisfy eligibility requirements.

Data Extraction and Quality Assessment

Data extraction was performed by the primary investigator (E.G.), and the data were compiled into a standardized data-extraction form. Data included sample size, diagnosis, inclusion/exclusion criteria, duration of symptoms, type of needling intervention (location, technique, and duration), main outcomes, time to outcome, and harm reported.

Included studies were analyzed by 2 independent reviewers (E.G. and Sebastian Sabadis), using the PEDro (Physiotherapy Evidence Database) quality scale. The PEDro scale is based on 11 criteria, of which 10 contribute to the score, representing methodological quality and risk of bias. The first item is not included in the score, as it relates to external validity of the study. The PEDro scale has been shown to have fair to good interrater reliability, with an intraclass correlation coefficient of 0.55 (95% confidence interval [CI]: 0.41, 0.72)²⁶ and higher scores indicating higher methodological quality. Disagreements between the reviewers were resolved by consulting another reviewer (Jodi Young) who was blind to previous assessment scores.

Quantitative Data Synthesis and Analysis

Interrater agreement between the reviewers who screened the studies for inclusion was performed using kappa statistics.²¹

Meta-analyses of study outcomes were performed wherever possible using RevMan (The Nordic Cochrane Centre, Copenhagen, Denmark). For analysis of continuous data, standardized mean differences (SMDs) with 95% CIs were used, as this method has been reported to be more generalizable³⁹ and allows for assessment of studies utilizing different scales to assess the same outcome. The random-effects model was used to account for variability between studies and its effect on the intervention. The I² statistic was used to measure the heterogeneity between trials.¹⁵ An I² value of 25% represents a small, 50% a moderate, and 75% a large degree of heterogeneity.¹⁶ Effect size was interpreted using Cohen's criteria for pooled estimates.⁸ Cohen described 0.2 as small, 0.5 as moderate, and 0.8 as large effect sizes.⁸

To assess for risk of publication bias, funnel plots were constructed.²² A symmetrical funnel plot indicates a lower risk of publication bias, whereas an asymmetrical funnel plot indicates a high risk of publication bias.

A 2-point change in pain on a visual analog scale (VAS) ranging from 0 to 10 was considered a clinically meaningful change in pain for between-group comparisons, as this criterion was also used in a previous systematic review of dry needling,²⁰ enabling comparisons with the previous literature. As we were unable to find evidence describing the minimal clinically important difference for PPT, between-group comparisons were not performed for PPT.

Two reviewers (E.G. and J.C.), using the Grading of Recommendations Assessment, Development and Evaluation (GRADE) approach,³ performed an analysis of the studies included in each meta-analysis independently. After appraising the evidence, each meta-analysis was classified as 1 of the following levels of evidence to support its findings¹⁴:

• High-quality evidence: further research is very unlikely to change our confidence in the estimate of effect
• Moderate-quality evidence: further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate
• Low-quality evidence: further research is very likely to have an important impact on our confidence in the estimate of effect and is likely to change the estimate
• Very low-quality evidence: any estimate of effect is very uncertain

Randomized controlled trials began with a high-quality evidence classification but were downgraded based on 5 domains: (1) study design and risk of bias (downgraded if greater than 25% of the participants were from studies with a high risk of bias, which we defined as PEDro scale scores of less than 6); (2) inconsistency of results (downgraded if significant heterogeneity was present on visual inspection or the I² value was greater than 50%); (3) indirectness (generalizability of the findings downgraded if greater than 50% of the participants were outside the target group); (4) imprecision (downgraded if fewer than 400 participants were included in the comparison for continuous data); and (5) other (publication bias).³⁷ We reduced the quality of evidence by 1 level for each domain not met in the comparison to determine the overall quality rating of the evidence for each meta-analysis performed.

Study Selection

The database search returned a total of 218 articles after duplicates were removed. Following title and abstract screening, 88 manuscripts were selected for full-text review. FIGURE 1 illustrates the flow of papers through the review. After the full review process, 12 studies were selected for inclusion.^{2,7,12,24,27,28,32,33,35,36,38,43} After contact with primary authors of included studies, 1 further study was identified for inclusion,⁶ resulting in a total of 13 studies. The absolute degree of rater agreement for the first and second stages of the study selection was 86% and 93%, respectively. The chance-corrected degree of agreement was good for screening by title and abstract (κ = 0.717; 95% CI: 0.629, 0.806) and good for screening by full text (κ = 0.759; 95% CI: 0.575, 0.943). For the title and abstract screening, the third reviewer resolved 38 disagreements, most commonly over whether the study met the inclusion criteria for dry needling (31%). For the full-text screening, the third reviewer was needed to resolve 3 disagreements, all regarding the type of health practitioner who provided the needling.

Study Characteristics

This systematic review focused on musculoskeletal pain and included 13 studies: 6 on neck pain (5 on mechanical neck pain^{6,24,28,32,43} and 1 on chronic whiplash-associated disorder³⁸), 1 on postoperative shoulder pain,² 1 on chronic lower back pain,³³ 1 on total knee arthroplasty,²⁷ 1 on chronic ankle instability,³⁵ 2 on myofascial pain,^12,36 and 1 on fibromyalgia.⁷ Inclusion and exclusion criteria varied greatly across the studies. The characteristics of the 13 studies are included in TABLE 1.

The 13 trials in this systematic review included a total of 723 participants. The majority (85%) of the included studies examined the effects of dry needling in participants with chronic musculoskeletal conditions. Two trials (15%) examined dry needling either coinciding with surgical intervention or after surgical intervention.

Two of the studies utilized control groups that did not receive dry needling, 3 used control groups that received sham dry needling, 6 compared dry needling to other treatments, and 2 used a variety of comparison groups (TABLE 1). All dry needling and comparison treatments were performed by physical therapists. Follow-up periods ranged widely from immediate to 12 months.

The risk of bias within studies was assessed with PEDro scale scores (TABLE 2). The absolute percent of rater agreement for PEDro scale scoring was 94%, and the chance-corrected degree of agreement was very good (κ = 0.863; 95% CI: 0.771, 0.955). The PEDro scale scores for included studies ranged from 4 to 9 (out of a maximum score of 10), with a median score of 7 and an interquartile range of 2 (6–8). None of the trials were able to blind the treating therapist, and only 3 (23%) of the trials were able to blind subjects through the use of sham dry needling. Only 8 (62%) studies included concealed allocation. All studies specified eligibility criteria, randomized patients, provided results of between-group statistical comparisons for at least 1 key outcome, and provided both point measures and measures of variability for at least 1 key outcome. Most studies (n = 12, 92%) scored well on having similar groups at baseline. Nine (69%) of the studies collected measures of at least 1 key outcome from more than 85% of the subjects initially allocated to groups. The risk of bias across studies is displayed in FIGURE 2.

Download FigureDownload PowerPoint

TABLES 3 and 4 provide the summary of findings and quality of evidence for all comparisons and outcomes included in this review.

Dry Needling Versus Control/Sham: Immediate to 12-Week Effects

Seven studies^{7,12,27,28,32,36,38} examined the immediate to 12-week effects of dry needling, and 6^{7,12,27,28,32,38} were able to be grouped for meta-analyses to determine the effect on pain (FIGURE 3). There is low-quality evidence suggesting a moderate effect⁸ (SMD, −0.7; 95% CI: −1.06, −0.34) favoring dry needling over control/sham. Heterogeneity was high (I² = 78%). All 7 studies favored dry needling over control/sham at all assessment points for reducing pain, except for Sterling et al³⁸ at 6 weeks. The 2 studies with the largest treatment effect were Pecos-Martín et al³² at 1 and 4 weeks and Mejuto-Vázquez et al²⁸ at 1 week, with a raw treatment effect size of 2.7, 3.0, and 2.6 on a VAS, respectively, which are considered clinically meaningful changes in pain. The raw treatment effect sizes for the remaining studies^7,12,27,38 on pain (1.7, 0.35, 0.85, 1.5, 0.58, 0, 1.2, 0.47, 0.3) are of questionable clinical meaningfulness (FIGURE 3).

Download FigureDownload PowerPoint

Five studies^{7,12,28,32,38} examined the immediate to 12-week effects of dry needling on PPT and were able to be meta-analyzed (FIGURE 4). There is very low-quality evidence suggesting a moderate effect (SMD, 0.8; 95% CI: 0.32, 1.27) favoring dry needling over control/sham. Heterogeneity was high (I² = 87%). On 9 of 11 occasions that PPT was assessed in the 12 weeks after intervention, dry needling increased PPT to a greater degree than control/sham.

Download FigureDownload PowerPoint

Five studies^{7,27,32,36,38} assessed functional outcomes versus control/sham during the immediate to 12-week follow-up (FIGURE 5). There is low-quality evidence suggesting a small effect⁸ (SMD, −0.44; 95% CI: −0.85, −0.04) favoring dry needling over control/sham. Heterogeneity was high (I² = 79%). Three out of the 5 studies (60%) found that dry needling improved functional outcome scores more than control or sham treatment.

Download FigureDownload PowerPoint

Dry Needling Versus Control/Sham: 6- to 12-Month Effects

Two studies^27,38 examined the long-term effect of dry needling and were meta-analyzed to determine effect on pain (FIGURE 6). There is moderate-quality evidence suggesting a small effect⁸ (SMD, −0.26; 95% CI: −0.58, 0.06) favoring dry needling over control/sham in the long term. The 95% CI crosses the line of no difference, and heterogeneity was low (I² = 11%). Two time points analyzed by Sterling et al³⁸ were in favor of dry needling, whereas that of Mayoral et al²⁷ was in favor of control/sham. The raw VAS scores in the Sterling et al³⁸ trial indicate a reduction in pain in favor of dry needling at 6 months of 1.2 points, and at 12 months of 0.5 points. The raw VAS score at 6 months in the Mayoral et al²⁷ trial was 0.26 in favor of the control (FIGURE 6).

Download FigureDownload PowerPoint

Two studies^27,38 assessed long-term effectiveness of dry needling on functional outcomes versus control/sham (FIGURE 7). There is low-quality evidence suggesting a small effect (SMD, −0.32; 95% CI: −0.62, −0.02) favoring dry needling over sham/control. Heterogeneity was low (I² = 0%).

Download FigureDownload PowerPoint

Dry Needling Versus Other Treatment: Immediate to 12-Week Effects

Eight studies^{2,6,12,24,33,35,36,43} examined the immediate to 12-week effects of dry needling, and 6 were able to be included in meta-analyses to determine the effect on pain (FIGURE 8). There is moderate-quality evidence suggesting a small effect⁸ (SMD, −0.43; 95% CI: −0.77, −0.10) favoring dry needling over other treatment. Heterogeneity was moderate (I² = 67%). On 9 of 12 occasions when pain was assessed, results favored dry needling compared to other treatment (FIGURE 8). Only Salom-Moreno et al³⁵ found a raw effect size on pain of greater than 2 points (2.1) at 12 weeks, in favor of dry needling. All other raw effect sizes on pain are of questionable clinical meaningfulness when compared to other treatments (FIGURE 8).

Download FigureDownload PowerPoint

Four studies^6,12,24,43 examined the immediate to 12-week effects of dry needling on PPT compared to other treatments and were meta-analyzed (FIGURE 9). There is very low-quality evidence suggesting a moderate effect (SMD, 0.61; 95% CI: 0.08, 1.14) favoring dry needling over other treatment. Heterogeneity was high (I² = 85%). On 7 of 10 occasions of measurement, PPT results were in favor of dry needling compared to other treatment.

Download FigureDownload PowerPoint

Six studies^{2,6,24,35,36,43} assessed functional outcomes versus other treatment over various periods ranging from 1 week to 12 weeks (FIGURE 10). There is very low-quality evidence suggesting no treatment effect (SMD, −0.01; 95% CI: −0.49, 0.47) of dry needling over other treatments. Heterogeneity was moderate (I² = 70%). Only 1³⁵ of the 6 studies (17%) that compared dry needling to other treatment reported that dry needling led to greater improvement in functional outcome.

Download FigureDownload PowerPoint

Publication Bias

Funnel plots were constructed to determine the risk of publication bias in the 8 meta-analyses performed (FIGURE 11). The funnel plots for the effects of dry needling versus control/sham on pain for immediate to 12-week follow-up (FIGURE 11A), on functional outcomes for immediate to 12-week follow-up (FIGURE 11C), on pain for 6- to 12-month follow-up (FIGURE 11D), and for the effects of dry needling versus other treatment on pain for immediate to 12-week follow-up (FIGURE 11F) are symmetrical, suggesting a lower likelihood of publication bias. However, the asymmetrical funnel plots for PPT (FIGURES 11B and 11G), the effects of dry needling versus control/sham on functional outcomes for 6- to 12-month follow-up (FIGURE 11E), and the effects of dry needling versus other treatment on functional outcomes for immediate to 12-week follow-up (FIGURE 11H) suggest a possible risk of publication bias.

Download FigureDownload PowerPoint

Download FigureDownload PowerPoint

Limitations

This systematic review has some limitations. The included studies were limited to those investigating dry needling performed by physical therapists, which might have excluded articles reporting the effects of dry needling on musculoskeletal pain performed by other practitioners. This was intentional, as the study examined evidence specific to physical therapists performing dry needling. Another limitation is the high heterogeneity in 5 of the 8 meta-analyses performed. This may be explained by the inclusion of studies investigating any type of musculoskeletal pain, where participant samples differed, comparison groups varied, and follow-up times for outcomes were different. The heterogeneity was one of the reasons we elected to use the random-effects model, allowing for improved internal validity of the results. We included all types of musculoskeletal conditions because this is typical of physical therapy clinical practice, where therapists treat a variety of clinical conditions. Also, only studies published in English were included in this review.

Though we only included randomized controlled trials, the actual overall quality of the evidence was considered to be very low to moderate using the GRADE approach. This quality of evidence suggests that further research is likely to have an important effect and is likely to change the estimate; therefore, the findings should be interpreted accordingly. Also, our results are limited to the studies included in the review. The present findings cannot be generalized to different patient populations treated by different practitioners with varied training in dry needling.

Findings

Very low- to moderate-quality evidence suggests that dry needling performed by physical therapists is more effective than no treatment, sham dry needling, or other treatments for reducing pain and improving PPT in patients presenting with musculoskeletal pain in the immediate to 12-week follow-up period. Very low- to low-quality evidence suggests superior outcomes with dry needling for functional outcomes when compared to no treatment or sham needling, but no difference in functional outcomes when compared to other physical therapy treatments. Evidence of the long-term benefit of dry needling is currently lacking.

Implications

Dry needling appears to be at least as effective as other treatments included in this review, and more effective than sham or no treatment for reducing pain and increasing PPT during the immediate to 12-week treatment period in patients with musculoskeletal pain.

Caution

The overall quality and limited number of studies reporting on dry needling performed by physical therapists at this time, combined with the high heterogeneity found in the results of the meta-analyses, indicate that readers should use caution when interpreting these results.