Pharmacological treatments for low back pain in adults: an overview of Cochrane Reviews

Key messages

For acute low back pain 

• NSAIDs and muscle relaxants may provide small benefits on pain, however muscle relaxants may be associated with unwanted effects. Paracetamol had no effect on pain or unwanted effects.

For chronic low back pain

• Opioids may reduce pain but may be associated with unwanted effects. NSAIDs may reduce pain without unwanted effects and antidepressants may make little or no difference on pain. 

Physicians should discuss the possibility for a small effect on pain and increased risk for unwanted effects when considering different medicines for treating low back pain. Funders and researchers should prioritise identification of medicines that provide clinically meaningful benefits to people with low back pain.

What is low back pain and how is it treated?

Low back pain (LBP) is a common and debilitating health condition. In most cases, the cause or causes of low back pain cannot be reliably identified and is described as ‘non-specific’ LBP. Physicians commonly prescribe medicines to treat LBP. There are many types of medicines and medicine classes available, for example, opioid analgesics, nonsteroidal anti-inflammatory drugs (NSAIDs), and paracetamol. With so many options available, there is a need to determine which medicines are best and safest.

What did we want to find out?

We wanted to summarise the evidence from Cochrane Reviews on the most effective and safest medicines for adults with non-specific LBP.

What did we do?

We searched for all Cochrane systematic reviews that assessed the benefits and harms of medicines for adults with non-specific LBP to produce an overview of Cochrane evidence.

What did we find?

We found seven reviews (that included 103 studies on a total of 22,238 people). Five reviews were assessed as having high quality. The included reviews reported data on six distinct medicines or medicine classes: paracetamol, NSAIDs (for example, ibruprofen), muscle relaxants (for example, cyclobenzaprine), benzodiazepines (for example, diazepam), opioids (for example, tapentadol), and antidepressants (for example, paroxetine). Five reviews included participants reporting LBP lasting longer than six weeks. The confidence in the evidence ranged from very low to high.

For people with acute LBP, we found that NSAIDs and muscle relaxants may reduce pain in the short-term (≤ three months postintervention). However, muscle relaxants may be associated with unwanted effects. Paracetamol had no effect on pain or unwanted effects and no reviews looked at opioids or antidepressants. For chronic LBP, we found that opioids may reduce pain in the short-term but may be associated with unwanted effects such as nausea, headache, constipation, and dizziness. NSAIDs may reduce pain in the intermediate term (> 3 months and ≤ 12 months postintervention) without unwanted effects. Antidepressants had no effect on chronic LBP and no review looked at paracetamol for chronic LBP.

What are the limitations of the evidence?

We have reduced confidence in the evidence because one review was of moderate quality and one review was of low quality and six reviews were published more than five years ago. There is a need to update these Cochrane Reviews following recommended guidance.

For acute LBP, we are at least moderately confident about the effects of paracetamol, NSAIDs and muscle relaxants on short-term pain and function. For other time points and other medicines (e.g. opioids, antidepressants), we have no evidence to inform treatment decisions.

For chronic LBP, we are at least moderately confident about the effects of paracetamol and opioids on short-term pain and function but less confident about the effects of other medicines (e.g. NSAIDs, antidepressants, muscle relaxants, and benzodiazepines). Factors that decreased confidence in findings included flaws in how the studies were designed (patients were not assigned to treatments randomly, allocation to treatment assignment was not concealed, patients were not compliant to their prescribed treatment), not having enough studies or participants to be certain about the results, and variations between treatment delivery.

The definition and reporting of unwanted effects for each medicine within each review was limited, making it difficult to assess safety for each pharmacological intervention. There remains clear gaps in the evidence base for the safety of medicines for LBP. 

How up to date is this evidence?

This evidence is up to date to June 2021

Authors' conclusions: 

We found no high- or moderate-certainty evidence that any investigated pharmacological intervention provided a large or medium effect on pain intensity for acute or chronic LBP compared to placebo. For acute LBP, we found moderate-certainty evidence that NSAIDs and muscle relaxants may provide a small effect on pain, and high-certainty evidence for no evidence of difference between paracetamol and placebo. For safety, we found very low- and high-certainty evidence for no evidence of difference with NSAIDs and paracetamol compared to placebo for the risk of adverse events, and moderate-certainty evidence that muscle relaxants may increase the risk of adverse events. For chronic LBP, we found low-certainty evidence that NSAIDs and very low- to high-certainty evidence that opioids may provide a small effect on pain. For safety, we found low-certainty evidence for no evidence of difference between NSAIDs and placebo for the risk of adverse events, and low-certainty evidence that opioids may increase the risk of adverse events. 

Read the full abstract...
Background: 

Pharmacological interventions are the most used treatment for low back pain (LBP). Use of evidence from systematic reviews of the effects of pharmacological interventions for LBP published in the Cochrane Library, is limited by lack of a comprehensive overview.

Objectives: 

To summarise the evidence from Cochrane Reviews of the efficacy, effectiveness, and safety of systemic pharmacological interventions for adults with non-specific LBP.

Methods: 

The Cochrane Database of Systematic Reviews was searched from inception to 3 June 2021, to identify reviews of randomised controlled trials (RCTs) that investigated systemic pharmacological interventions for adults with non-specific LBP. Two authors independently assessed eligibility, extracted data, and assessed the quality of the reviews and certainty of the evidence using the AMSTAR 2 and GRADE tools. The review focused on placebo comparisons and the main outcomes were pain intensity, function, and safety.

Main results: 

Seven Cochrane Reviews that included 103 studies (22,238 participants) were included. There is high confidence in the findings of five reviews, moderate confidence in one, and low confidence in the findings of another. The reviews reported data on six medicines or medicine classes: paracetamol, non-steroidal anti-inflammatory drugs (NSAIDs), muscle relaxants, benzodiazepines, opioids, and antidepressants. Three reviews included participants with acute or sub-acute LBP and five reviews included participants with chronic LBP.

Acute LBP

Paracetamol

There was high-certainty evidence for no evidence of difference between paracetamol and placebo for reducing pain intensity (MD 0.49 on a 0 to 100 scale (higher scores indicate worse pain), 95% CI -1.99 to 2.97), reducing disability (MD 0.05 on a 0 to 24 scale (higher scores indicate worse disability), 95% CI -0.50 to 0.60), and increasing the risk of adverse events (RR 1.07, 95% CI 0.86 to 1.33).

NSAIDs

There was moderate-certainty evidence for a small between-group difference favouring NSAIDs compared to placebo at reducing pain intensity (MD -7.29 on a 0 to 100 scale (higher scores indicate worse pain), 95% CI -10.98 to -3.61), high-certainty evidence for a small between-group difference for reducing disability (MD -2.02 on a 0-24 scale (higher scores indicate worse disability), 95% CI -2.89 to -1.15), and very low-certainty evidence for no evidence of an increased risk of adverse events (RR 0.86, 95% CI 0. 63 to 1.18).

Muscle relaxants and benzodiazepines

There was moderate-certainty evidence for a small between-group difference favouring muscle relaxants compared to placebo for a higher chance of pain relief (RR 0.58, 95% CI 0.45 to 0.76), and higher chance of improving physical function (RR 0.55, 95% CI 0.40 to 0.77), and increased risk of adverse events (RR 1.50, 95% CI 1. 14 to 1.98).

Opioids

None of the included Cochrane Reviews aimed to identify evidence for acute LBP.

Antidepressants

No evidence was identified by the included reviews for acute LBP.

Chronic LBP

Paracetamol

No evidence was identified by the included reviews for chronic LBP.

NSAIDs

There was low-certainty evidence for a small between-group difference favouring NSAIDs compared to placebo for reducing pain intensity (MD -6.97 on a 0 to 100 scale (higher scores indicate worse pain), 95% CI -10.74 to -3.19), reducing disability (MD -0.85 on a 0-24 scale (higher scores indicate worse disability), 95% CI -1.30 to -0.40), and no evidence of an increased risk of adverse events (RR 1.04, 95% CI -0.92 to 1.17), all at intermediate-term follow-up (> 3 months and ≤ 12 months postintervention).

Muscle relaxants and benzodiazepines

There was low-certainty evidence for a small between-group difference favouring benzodiazepines compared to placebo for a higher chance of pain relief (RR 0.71, 95% CI 0.54 to 0.93), and low-certainty evidence for no evidence of difference between muscle relaxants and placebo in the risk of adverse events (RR 1.02, 95% CI 0.67 to 1.57).

Opioids

There was high-certainty evidence for a small between-group difference favouring tapentadol compared to placebo at reducing pain intensity (MD -8.00 on a 0 to 100 scale (higher scores indicate worse pain), 95% CI -1.22 to -0.38), moderate-certainty evidence for a small between-group difference favouring strong opioids for reducing pain intensity (SMD -0.43, 95% CI -0.52 to -0.33), low-certainty evidence for a medium between-group difference favouring tramadol for reducing pain intensity (SMD -0.55, 95% CI -0.66 to -0.44) and very low-certainty evidence for a small between-group difference favouring buprenorphine for reducing pain intensity (SMD -0.41, 95% CI -0.57 to -0.26).

There was moderate-certainty evidence for a small between-group difference favouring strong opioids compared to placebo for reducing disability (SMD -0.26, 95% CI -0.37 to -0.15), moderate-certainty evidence for a small between-group difference favouring tramadol for reducing disability (SMD -0.18, 95% CI -0.29 to -0.07), and low-certainty evidence for a small between-group difference favouring buprenorphine for reducing disability (SMD -0.14, 95% CI -0.53 to -0.25).

There was low-certainty evidence for a small between-group difference for an increased risk of adverse events for opioids (all types) compared to placebo; nausea (RD 0.10, 95% CI 0.07 to 0.14), headaches (RD 0.03, 95% CI 0.01 to 0.05), constipation (RD 0.07, 95% CI 0.04 to 0.11), and dizziness (RD 0.08, 95% CI 0.05 to 0.11).

Antidepressants

There was low-certainty evidence for no evidence of difference for antidepressants (all types) compared to placebo for reducing pain intensity (SMD -0.04, 95% CI -0.25 to 0.17) and reducing disability (SMD -0.06, 95% CI -0.40 to 0.29).