Effective treatments for musculoskeletal pain

Effective treatment options for musculoskeletal pain in primary care: A systematic overview of current evidence – Jun 2017 – Free full-text PMC5480856/

These are my annotations of a very long, very thorough PubMed article on pain treatments.

Musculoskeletal pain, the most common cause of disability globally, is most frequently managed in primary care.  People with musculoskeletal pain in different body regions share similar characteristics, prognosis, and may respond to similar treatments.

This overview aims to summarise current best evidence on currently available treatment options for the five most common musculoskeletal pain presentations (back, neck, shoulder, knee and multi-site pain) in primary care.  


A systematic search was conducted. Initial searches identified clinical guidelines, clinical pathways and systematic reviews. Additional searches found recently published trials and those addressing gaps in the evidence base.


Moderate to strong evidence suggests that exercise therapy and psychosocial interventions are effective for relieving pain and improving function for musculoskeletal pain

NSAIDs and opioids reduce pain in the short-term, but the effect size is modest and the potential for adverse effects need careful consideration

Corticosteroid injections were found to be beneficial for short-term pain relief among patients with knee and shoulder pain.

However, current evidence remains equivocal on optimal dose, intensity and frequency, or mode of application for most treatment options.


This review presents a comprehensive summary and critical assessment of current evidence for the treatment of pain presentations in primary care. The evidence synthesis of interventions for common musculoskeletal pain presentations shows moderate-strong evidence for exercise therapy and psychosocial interventions, with short-term benefits only from pharmacological treatments. Future research into optimal dose and application of the most promising treatments is needed.


Research evidence suggests that in the general population and those presenting to primary care, localised musculoskeletal pain frequently coexists in more than one body region

and that those with different regional pains share similar underlying attributes, course of symptoms and prognostic factors

A more holistic view is perhaps difficult to obtain since trials and systematic reviews usually focus on a specific musculoskeletal pain site, comparing only two or three treatment options

To our knowledge there are no published reviews in which evidence regarding the comparative effectiveness of a wide range of treatments is systematically synthesised for the most common musculoskeletal pain presentations.

The aim of this study was to critically appraise current best evidence regarding the effectiveness of treatments to reduce pain and /or improve function for people with the five most common musculoskeletal pain presentations in primary care (i.e., back, neck, shoulder, knee and multi-site pain as indicated by Jordan et al.[25]). The specific objectives of this review were to:

  1. identify effective treatment options for the five most common musculoskeletal pain presentations and
  2. highlight gaps in evidence and priorities for policy or future research.

The review also identified, where available, evidence regarding patient subgroups most likely to respond to different treatment options


Sources of data and search strategy

Integrated information from higher levels of evidence has been suggested as an “ideal source of evidence for clinical decision-making” by the Evidence Based Practice group (http://hsl.mcmaster.libguides.com/ebm). Therefore, using national clinical guidelines, policy documents, care pathways such as Map of Medicine (MoM), and clinical evidence summaries as a starting point, the search for evidence for this overview followed a pyramidal tract through a hierarchy of available evidence

All Cochrane reviews matching the inclusion criteria were included in the synthesis.

Relevant publications (guidelines, systematic reviews and meta-analyses of RCTs as well as recent RCTs which are yet to be summarised in reviews) were obtained and assessed against predefined eligibility criteria according to the study protocol by two reviewers.

But guidelines are a political/legal response to current assumptions, so they cannot possibly be evidence of anything except the political era they were written in.

Inclusion criteria

  • Study populations: Reviews/studies of adults (18 years and over) presenting with at least one of the five most common musculoskeletal pain presentations: back, neck, shoulder, knee and multi-site pain (the latter defined as musculoskeletal pain in more than one area of the body).
  • Type of treatments: Reviews/studies of currently available treatments
  • Outcomes: Reviews/studies had to report outcomes of pain (e.g. intensity, widespreadness, bothersomeness, number of episodes, duration), and/or functional disability.

These were considered primary outcomes for this review.

Quality appraisal

In order to weigh the conclusion of reviews within our evidence summaries, the methodological quality of non-Cochrane systematic reviews was assessed using the 11-item ‘assessment of multiple systematic reviews’ (AMSTAR) checklist [26].

The guidelines and care pathways which were included in this evidence synthesis were not quality assessed as they all made use of published development processes based on explicit methodology.

Yet the CDC opioid prescribing guidelines are seriously distorted, use biased data, and misinterpreted results.

Such politically expedient guidelines are the opposite of evidence.

Extraction of data

Data were extracted by one reviewer using a data collection form and independently checked for consistency and completeness by a second reviewer.

More specifically, data were extracted regarding:

  • population characteristics (e.g. age, gender, symptom duration, musculoskeletal pain site and where possible musculoskeletal pain condition/diagnoses,
  • treatments (type/intensity/dosage),
  • primary and secondary outcome measures (as stated above),
  • estimates of treatment effect (where pooled, and as presented in the systematic reviews),
  • estimates of treatment effect for patient subgroups (where available),
  • treatment setting (e.g. primary care), and
  • sources of evidence.

Treatments were assessed for short-term (up to 3 months) and long-term (greater than 6 months) effectiveness based on the primary outcomes of pain and function.

Grading of evidence

Summaries of the overall evidence for the effectiveness of treatment options and strength of recommendations for each pain site were assessed based on (a modified) GRADE rating (http://www.gradeworkinggroup.org/).

For each treatment option, evidence was graded as:

  1. “Very weak evidence”—based solely on expert opinion or consensus in guidelines only or in the absence of systematic review evidence
  2. “Limited evidence”—in the presence of little evidence from systematic reviews/evidence-based guidelines AND when there were small, inconsistent, or non-significant treatment effect sizes
  3. “Moderate evidence”–in the presence of little evidence from systematic reviews/evidence-based guidelines (as in 2) but showing a medium to large treatment effect OR in the presence of strong evidence from high quality systematic reviews, but with small or inconsistent treatment effect sizes
  4. “Strong evidence”—in the presence of strong evidence from high quality systematic reviews and evidence-based clinical guidelines AND medium or large effect sizes.

Evidence synthesis

A narrative synthesis approach was undertaken.

the evidence was summarised at a high level (using systematic reviews and guidelines where available), and therefore no new meta-analyses were conducted.

However, pooled estimates of treatment effectiveness from systematic reviews, as well as comments on the consistency and magnitude of treatment effects were extracted and reported

Additional information from policy documents and guidelines on treatment recommendations and priorities, including the type of evidence from which it was generated was also noted.

Guidelines are NOT evidence.


A total of 3,588 unique citations (including Cochrane reviews) were retrieved from the electronic bibliographic databases.

71 Cochrane systematic reviews met the selection criteria and were included. Non-Cochrane systematic reviews (n = 75) were only included where a gap not already covered by Cochrane reviews was identified

As Cochrane reviews followed a generic protocol specifying methods and review protocols go through a comprehensive peer review process prior to publication, the methodological quality of most Cochrane reviews included in this evidence synthesis was satisfactory

Evidence synthesis

Effectiveness of available treatments for musculoskeletal pain was highlighted in the following order:

  1. self-management advice and education,
  2. exercise therapy,
  3. manual therapy,
  4. pharmacological interventions,
  5. aids and devices,
  6. other treatments (including ultrasound, TENS, laser, acupuncture, ice / hot packs), psychosocial interventions and surgery.

A summary of the findings is presented in Table 1. 

1. Self-management advice and education

Evidence base: Evidence was extracted from two clinical guidelines, one clinical pathway and eight reviews about the effectiveness of self-management advice and education.

Magnitude of effects:

Where estimated, summary effect sizes were usually small and/or not statistically or clinically significant.


Bottom line:

Despite the limited evidence-base, there were strong recommendations for the use of self-management advice and education as a first line treatment option for musculoskeletal pain.

2. Exercise therapy

Magnitude of effects: Exercise therapy was determined to be beneficial for pain, function and quality of life in all five pain presentations

Exercise therapy led to clinically significant improvements in pain, function and quality of life for shoulder, knee, back and multi-site pain.

In addition, medium to large summary effect sizes were reported in favour of exercise across the body of evidence

There appears to be little empirical evidence in favour of any particular exercise type, programme or mode of delivery, either as structured individual or group treatment for musculoskeletal pain

although functional exercises (which adapt patients’ exercises to their activities of daily living, and enables them to perform such activities more easily and without injuries) appear to be more beneficial than exercises not specifically targeting function

Bottom line:

Current evidence shows significant positive effects in favour of exercise on pain, function, quality of life and work related outcomes in the short and long-term for all the musculoskeletal pain presentations (compared to no exercise or other control) but the evidence regarding optimal content or delivery of exercise in each case is inconclusive.

3. Manual therapy

Magnitude of effects: Pooled estimates for the effectiveness of manual therapy for musculoskeletal pain were generally statistically significant, but variable in terms of size of the treatment effect

Manipulation, mobilisation and massage (where indicated) were reported to be beneficial for immediate and or short-term (4–6 weeks) improvement in range of motion and function in both acute and chronic neck pain patients as well as those with whiplash

For back pain, evidence suggests that manual therapy alone or in combination with other treatments may offer some benefit for pain and function

Compared with other treatments (e.g. general practitioner care, acupuncture, ultrasound, standard physiotherapy, analgesic therapy, exercise, or back school), manual therapy appears to confer little or no clinically important effect on pain intensity, functional status, global improvement or return to work among patients with acute, subacute or chronic back pain with or without sciatica

Strength of evidence: Despite several high quality reviews examining the effects of manual therapy on pain and function for neck, shoulder and back pain, current evidence generally shows small summary effect sizes or concludes no clinical effectiveness of manual therapy compared to sham or other active treatments. Overall strength of evidence was graded as limited.

Bottom line:

Current evidence regarding manual therapy is beset by heterogeneity across clinical trials. Due to paucity of high quality evidence, it is uncertain if the efficacy of manual therapy might be different for different patient subgroups or influenced by the type and experience of professional delivering the therapy.

4. Pharmacological treatments—Analgesics (oral & topical)

Evidence base: Thirty systematic reviews of pharmacological interventions for musculoskeletal pain examined the effectiveness of analgesics (opioids and non-opioids) in the short and long-term as well as in acute and chronic pain presentations.

Magnitude of effects: Compared to placebo, acetaminophen (paracetamol) was not more effective

for relieving knee and back pain. NSAIDs and opioid analgesics (especially for acute pain) were generally found to be effective but beneficial effects were evident mostly in the short-term

Cyclooxygenase (Cox)-2 selective inhibitors (e.g. celecoxib), were found to be effective for musculoskeletal pain relief. However, these were more likely to be associated with higher risks of adverse cardiovascular and gastrointestinal events

In the long-term and for more chronic pain presentations, stepwise analgesia according to the WHO analgesic ladder (mostly based on expert opinion) may be recommended

Strength of evidence: With consistent medium summary effect sizes reported across moderate to high quality systematic reviews and clinical guidelines, there is moderate evidence that pharmacological therapies are beneficial for the short-term relief of musculoskeletal pain. Overall strength of evidence was graded as moderate.

Bottom line:

NSAIDs, Cox-2 selective inhibitors and opioids reduce pain in the short-term, but the effect size is modest and the potential for adverse effects such as gastrointestinal bleeding and opioids-induced hyperalgesia need careful consideration.

5. Pharmacological interventions–injections

Likewise for knee pain, corticosteroid injections were found to be effective in the short-term for relieving moderate to severe pain compared to placebo

Though corticosteroid injections were found to relieve pain, there was a lack of evidence for clinically significant effects on function

For knee pain, viscosupplements such as intra-articular hyaluronate injections were found to be better than placebo

However, high clinical and statistical heterogeneity, evidence of publication bias and low quality trials preclude definitive recommendations about routine use in clinical practice

Furthermore, the available evidence did not suggest injections are effective for the management of neck pain or back pain

Overall, there was no strong evidence for the use of epidural spinal injections with or without steroids, as benefits (immediate reductions in pain) were small and not sustained

Generally, in the long-term, injections may be no more effective than non-pharmacological interventions such as exercise

Evidence also suggests that the addition of corticosteroid injections to local anaesthetic does not confer improved symptom relief in the long-term

however, expert opinion and guideline recommendations support its use prior to, or alongside, exercise and self-management advice

Bottom line:

The evidence indicates that injections offer short-term pain relief for shoulder and knee pain but effectiveness for back and neck pain is uncertain.

6. Aids & devices—Orthotics, tapes, braces, cervical collars and other support devices

Magnitude of effects: Either as stand-alone treatment or mostly in combination with other treatments, aids and devices for musculoskeletal pain have generally shown small effects

This may be attributed to marginal pain relief (in the short-term), and inclination to induce rest and inactivity hence prolonging disability.

Bottom line:

For neck, shoulder, back and knee pain presentations, available evidence does not justify routine use of aids and devices for effective improvement of pain, function, and / or work outcomes.

7. Other treatments: Acupuncture, ultrasound, TENS, laser, ice / hot packs

For acupuncture, available evidence from a good quality individual patient data meta-analysis suggests that acupuncture may be effective for short-term relief of back pain and knee pain with medium summary effect sizes

However, effects on function were reported to be minimal and not maintained at longer-term follow-up

Similarly for neck and shoulder pain, acupuncture was only found to be effective for short-term (immediately post-treatment and at short-term follow-up) symptom relief

TENS was no more effective for reducing pain than placebo in chronic back pain, neck pain, shoulder pain, knee, and chronic musculoskeletal pain

Ultrasound and shockwave therapy do not appear to significantly improve clinical outcomes for acute and chronic low back pain.

With regards to knee pain, other treatments including ultrasound, electromagnetic fields, low level laser therapy, TENS, biofeedback, neuromuscular electrical stimulation may confer added benefits to exercise and / or surgical treatment but empirical and clinical effect sizes are small and only supported by weak evidence

Bottom line:

The evidence for the clinical effectiveness of most of these other treatment options was not substantiated by strong evidence.

Either as stand alone or in combination with other treatments, the often small effect sizes as a result of these treatments for improving musculoskeletal pain and function was mostly not clinically significant.

8. Psychosocial interventions

Magnitude of effects: Reviews of psychosocial treatments for the management of musculoskeletal pain included a wide range of approaches that aimed to achieve increased self-management, behavioural and/or cognitive changes alongside biomedical management of pain S6 Table.

Interventions were often multimodal and involved multidisciplinary treatment.

At long-term follow-up, medium summary effect sizes  were reported for pain, function and/ or other psychosocial related-outcome measures such as quality of life.

Psychosocial interventions in combination with other treatment options appear to provide additional benefit for all musculoskeletal pain presentations. However, there was no consensus on specific treatment components, providers and settings for optimal outcomes

Bottom line:

Available evidence suggests beneficial effects of psychosocial interventions, particularly for patients identified as having a poor prognosis prior to treatment.


Most guidelines specify that surgical treatments are indicated in a small proportion of patients (as low as 8%) for neck, shoulder, back and knee pain presentations.

Within the body of synthesised evidence (supplementary S1 and S7 Tables), the presence of serious pathology, substantial pain and disability or symptoms which are refractory to conservative treatment were prominent indications for surgery.

Generally for neck, shoulder, knee and back pain, when indicated, there is moderate evidence that surgical intervention does provide benefits for pain, and function compared to waiting list controls or conservative treatments including analgesia and exercise in the short-term.

Available evidence suggests there are no long-term benefits of surgical procedures for clinical outcomes compared with conservative treatment. Neither was there strong evidence for a significant difference in favour of any particular surgical technique for any of the pain sites

Bottom line:

The effectiveness of surgery as a first line treatment option is not established in current literature.

The current evidence base is limited in terms of quantity, especially comparing surgical versus conservative interventions but there is moderate evidence from guidelines,

Cochrane reviews and other systematic reviews to support short-term efficacy of surgical interventions for pain and function for specific neck, shoulder, knee and back pain presentations.

Available evidence also suggests that surgery is not superior to conservative treatment options in the long-term.


The effectiveness of exercise therapy, psychosocial interventions and corticosteroid injections was consistently supported by empirical evidence of mostly medium effect sizes provided by meta-analyses of RCTs, by guidelines, and expert opinion for musculoskeletal pain

The above are all bad sources of evidence,

In this review, there was little information within the evidence base in relation to patient subgroups most likely to respond to different treatment options.

an optimal approach to management of musculoskeletal pain may involve strategic selection of treatments best suited for different patients.

It is worth noting that in many of the reviews, guidelines and trials contributing to this evidence-base, individual treatments were rarely used in isolation. Therefore, the evidence for the isolated effectiveness of treatments in some reviews was difficult to assess.

Overall completeness and applicability of evidence

As expected, given the breadth of this review there was wide heterogeneity in study populations, outcomes, and statistical methods for estimating summary effect measures in the included systematic reviews.

Interpreting findings within this overview was also complicated by variability in both the intervention and control conditions (placebo, no treatment, active treatments) examined within the reviews, making it difficult to summarise evidence regarding the magnitude of treatment effects.


Effective healthcare depends on high quality evidence.

Best available evidence shows that patients with musculoskeletal pain problems in primary care can be managed effectively with non-pharmacological treatments such as self-management advice, exercise therapy, and psychosocial interventions.

Pharmacological interventions such as corticosteroid injections (for knee and shoulder pain) were shown to be effective treatment options for the short-term relief of musculoskeletal pain and may be used in addition to non-pharmacological treatments.

NSAIDs and opioids also offer short-term benefit for musculoskeletal pain, but the potential for adverse effects must be considered.

So a doctor has to find the balance between

  1. the very effective, very immediate pain relief of opioids and 
  2. the statistically low possibility of an individual becoming addicted.

Furthermore, the optimal treatment intensity, methods of application, amount of clinical contact, and type of provider or setting, are unclear for most treatment options.


Other thoughts?

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s