Effectiveness of musculoskeletal manipulations in patients with neck pain: a protocol for a systematic review and network meta-analysis

Abstract

Introduction

Neck pain is a common problem that severely affects physical and mental health. While musculoskeletal manipulations are recommended as the first-line treatment for adults with neck pain, the comparative effectiveness of different musculoskeletal manipulations remains unclear. This systematic review and network meta-analysis of randomised controlled trials (RCTs) will compare the effectiveness of different types of musculoskeletal manipulations, with the overarching aim of guiding clinical practice.

Methods and analysis

Two independent reviewers will search four English electronic databases (Web of Science, Cochrane Library, EMBASE, PubMed) and three Chinese electronic databases (China National Knowledge Infrastructure, China Science and Technology Journal Database, Wanfang) for relevant RCTs published from 1 January 2013 to 30 April 2023. The Clinical Trials Registry (ClinicalTrials.gov) will be searched for completed but unpublished RCTs. English and Chinese will be used to search English databases and Chinese databases, respectively. RCTs of musculoskeletal manipulations for adults (aged ≥18 years) with neck pain will be considered eligible for inclusion. A pairwise meta-analysis and network meta-analysis will be performed, and pooled risk ratios, standardised mean differences and 95% CIs will be determined.

Ethics and dissemination

Ethics approval is not required as this study is a literature review. The results of this review will be published in peer-reviewed journals or disseminated at conferences.

PROSPERO registration number

CRD42023420775.

STRENGTHS AND LIMITATIONS OF THIS STUDY.

• By integrating direct and indirect evidence, network meta-analysis will increase the precision of treatment estimates and allow for the ranking of available musculoskeletal manipulations.

• Our study will be conducted following the Cochrane Handbook, which will ensure methodological rigour.

• The revised Cochrane risk of bias tool (V.2.0) will be used to provide a critical appraisal of the risk of bias for included studies.

• We will assess the evidence certainty using the GRADE (Grading of Recommendations Assessment, Development and Evaluation) approach, which will provide essential information for clinical decision-making.

• The quality of the studies included in the systematic review will influence the results of the network meta-analysis.

Introduction

Neck pain is a common musculoskeletal disorder and major health burden. Globally, neck pain affects 288.7 million people, with a prevalence of 3551 cases per 100 000 population.¹ The number of neck pain cases is expected to increase significantly in the near future, owing to the ageing population and changes in working habits.² As the fourth most prevalent musculoskeletal disorder,³ neck pain poses a heavy socioeconomic burden and has severe adverse impacts on quality of life and both physical and mental health.⁴ Therefore, the prevention and treatment of neck pain warrants greater attention.⁵

Several interventions are available for the treatment of neck pain, including pharmacotherapy, electrotherapy, acupuncture and musculoskeletal manipulations.⁶ Due to its comparative safety and effectiveness, musculoskeletal manipulations are the recommended first-line treatment method and are widely used in clinical practice.⁷ A previous Cochrane review of several randomised controlled trials (RCTs) confirmed the effectiveness of musculoskeletal manipulations in patients with neck pain.⁸ There is currently a wide range of different musculoskeletal manipulations available. Nevertheless, their comparative effectiveness is unclear, and there is a lack of evidence to guide clinicians in the selection of the most effective intervention.

Network meta-analysis (NMA) allows ranking of intervention effectiveness by simultaneously comparing several interventions in a network plot.⁹ To our knowledge, the comparative effectiveness of different musculoskeletal manipulations has not yet been investigated via NMA. Thus, this NMA aims to systematically evaluate and synthesise the effectiveness of different types of musculoskeletal manipulation in patients with neck pain.

Methods and analysis

Design

The protocol for this review follows the Preferred Reporting Items for Systematic Reviews and Meta-Analysis Protocols (PRISMA) statement.¹⁰ The NMA will be conducted in accordance with the guidelines of the Cochrane Handbook (V.6.3, 2022),¹¹ and reporting will adhere to the guidelines of PRISMA-NMA.¹² This protocol has been registered in PROSPERO (CRD42023420775).

Information sources

Four English electronic databases (Web of Science, Cochrane Library, EMBASE and PubMed) and three Chinese electronic databases (China National Knowledge Infrastructure, China Science and Technology Journal Database and Wanfang) will be searched for RCTs published between 1 January 2013 and 30 April 2023. We will also search the Clinical Trials Registry (ClinicalTrials.gov) for completed but unpublished RCTs. Our three-part search strategy includes terms to identify (1) RCTs pertaining, (2) musculoskeletal manipulations and (3) neck pain. The detailed search strategies are presented in online supplemental appendix 1.

Eligibility criteria

RCTs will be included in accordance with the following criteria:

1. Participants: adults (age ≥18 years) with neck pain. Neck pain is defined as a primary complaint of pain between the occiput and the first thoracic vertebra.

2. Interventions: any musculoskeletal manipulations, including upper cervical spine and cervicothoracic spine manipulations;

3. Comparison: any other musculoskeletal manipulations, sham treatment, placebo, no treatment or active control (eg, education, exercise).

4. Outcomes: the primary outcome is neck pain intensity (measured by various scales, such as Visual Analogue Scale, Numeric Pain Rating Scale); the secondary outcomes include quality of life, flexion, extension and adverse events.

Exclusion criteria

Studies will be excluded if they (1) report primary complaints other than neck pain, such as postconcussion syndrome, headache and migraine; (2) are not published in Chinese or English or (3) musculoskeletal manipulations were only used as adjuvant therapy.

Selection process

All retrieved records will be imported into the literature management software. After the removal of duplicate records, two reviewers will independently screen titles and abstracts. Full texts of records meeting the inclusion criteria will be read, and all reasons for exclusion will be reported. Any disagreements will be discussed and resolved by the two reviewers; a third reviewer will be consulted when necessary. The study selection process will be reported using a PRISMA flow diagram (figure 1).

Figure 1.

Study screening process. RCTs, randomised controlled trials.

Data extraction

The following data for each included study will be extracted by two independent reviewers using a standardised extraction table agreed on by all authors: study identifiers (eg, first author, publication year, country or region, sample size, setting), details of intervention and comparator (eg, components, frequency, duration), patient information (eg, mean age, sex percentage) and outcomes (eg, neck pain scale scores at baseline and after the intervention, measurement tools).

Risk of bias assessment of individual studies

The risk of bias assessment tool used in this review will be the revised Cochrane risk of bias tool for RCTs (V.2.0).¹³ Studies will be evaluated in five domains: randomisation process, deviations from intended interventions, missing outcome data, measurement of the outcome and selection of the reported result. The overall risk of bias will be considered as ‘low’ if all five domains having the assessment of ‘low risk’; as ‘some concerns’ if at least one domain having the assessment of ‘some concerns’ and as ‘high’ if at least one domain having the assessment of ‘high risk’. The risk of bias assessment will be independently conducted by two reviewers. The third reviewer will resolve any conflicts through discussion, as needed.

Data synthesis

Measures of intervention effect

Dichotomous outcomes will be calculated as the risk ratio (RR) and 95% CI, whereas the standardised mean difference (SMD) and 95% CI will be reported when outcomes are continuous. If the available data are not suitable for meta-analysis, a systematic review will be performed.

Pairwise meta-analysis

Prior to undertaking a pairwise meta-analysis, Cochran’s Q test will be used to qualitatively evaluate whether there is heterogeneity among the RCTs, with p<0.1 indicating significant heterogeneity. I² values will be used to quantitatively measure the level of heterogeneity, which will be classified as follows: low (<25%), moderate (25%–50%) or high (>50%).¹⁴ A random-effects model will be used due to potential heterogeneity between the studies.¹⁵ Tables and forest plots will be used to report quantitative data. Publication bias will be assessed using both a funnel plot and the Egger’s test when 10 or more RCTs are included in the analysis.¹⁶

Network meta-analysis

NMA will be performed using a frequentist approach, and a random-effects model with maximum likelihood estimation will be used to assess the comparative effectiveness of eligible interventions via STATA (V.14).¹⁷ Pooled RR and SMD with 95% CI will be used to assess intervention effects on dichotomous and continuous outcome data, respectively.

An NMA must adhere to the assumptions of transitivity and consistency.¹⁸ Transitivity is assessed by comparing clinical and methodological characteristics between RCTs. When these characteristics are similar, transitivity is considered satisfactory.¹⁹ Inconsistency includes the global inconsistency and local inconsistency, with each component being evaluated by different methods. The design-by-treatment interaction model is used to detect global inconsistency,^{20 21} while the node-splitting method is used to evaluate local inconsistency.²⁰ The absence of inconsistency between indirect and direct comparisons is indicated by p>0.05 (for both the overall inconsistency test and local inconsistency test). In such cases, the consistency model will be selected to compare all interventions.^{22 23}

All comparisons in the network will be displayed in a network plot. The results of the interventions in the direct and indirect comparisons will be reported in a league table. The surface under the cumulative ranking curve will be used to rank the effectiveness of the interventions for each outcome.²⁴ Publication bias will be evaluated using a comparison-adjusted funnel plot if at least 10 studies are included.²⁵ A symmetrical distribution of scatter points around the zero line on the funnel plot indicates the absence of publication bias.

Sensitivity analysis

If sufficient RCTs are included, we will perform a sensitivity analysis by excluding RCTs with a high risk of bias.

Assessment of evidence certainty

The evidence certainty in the NMA will be evaluated using the Grading of Recommendations, Assessment, Development and Evaluation guidelines.^{26 27} First, based on the risk of bias, inconsistency, publication bias and indirectness, the evidence certainty of effect estimates about direct comparisons will be rated as high, moderate, low or very low. Second, the evidence certainty of indirect comparisons will be rated as high, moderate, low or very low based on the higher order loop (loops with the least common comparator connecting the comparison of interest). We will rate down the certainty of the indirect comparison one further level after taking into account intransitivity. Finally, the higher rating between direct and indirect comparison will be identified. And the evidence certainty about network comparisons will be rated after taking into account incoherence and imprecision. A third reviewer will resolve any disagreement until a consensus is reached. Detailed instruction of above-mentioned factors is shown in online supplemental appendix 2.

Patient and public involvement

None.

Discussion

This NMA will provide a comprehensive assessment of the available musculoskeletal manipulations for adults with neck pain. The results of our comparative analysis and assessment of evidence quality will be used to establish a hierarchy of interventions according to effectiveness, thereby guiding the selection of optimal musculoskeletal manipulation in clinical practice. Furthermore, our risk of bias assessment will inform the development of future studies with improved study designs and methodological quality.

Ethics and dissemination

Ethics approval is not required for this study, as it is a literature review and does not involve direct contact with patients or patient privacy issues. The results of this review will be disseminated to national and international scientific communities, healthcare professionals and the general population through publication in a peer-reviewed journal and presentation at scientific conferences.

Ethics statements

Patient consent for publication

Not applicable.