Article Text

Download PDFPDF

Strong opioids for osteoarthritis in primary care?


For some patients with osteoarthritis, commonly used drugs taken alone or in combination, offer insufficient pain relief, or are inappropriate. One issue this raises is whether strong opioids might then have a place in treatment. Some GPs may be reluctant to prescribe strong opioids for osteoarthritis. This may be due to concerns about addiction or long-term use of controlled drugs for non-cancer pain.1 Here we review the evidence on strong opioids as adjunctive treatment in osteoarthritis.

Statistics from

Management options for osteoarthritis

Management options for osteoarthritis involve a combination of non-drug and drug interventions.2,3 The non-drug, or ‘core’, interventions include ongoing access to appropriate information; appropriate exercise to include local muscle strengthening and general aerobic fitness; and interventions to facilitate loss of excess weight in those with associated functional limitations.2 Use of adjunctive drug options (such as paracetamol, oral or topical NSAIDs, capsaicin and/or intra articular corticosteroid injections) may be appropriate, depending on the patient's needs and preferences.2,3 Joint replacement surgery is an option if pain relief and functional improvements are inadequate with other treatments, and there is a significant impact on quality of life.2,3

What is a strong opioid?

Strong or stronger opioids (e.g. buprenorphine, fentanyl, hydromorphone, morphine, oxycodone) differ from weak opioids (e.g. codeine) in having a much broader dose range, and a proportionately greater effect can be achieved by increasing the dose in opioid-sensitive pain.46 A licensed formulation exists for most of these strong opioids for use in non-cancer pain. Tramadol is considered as either a weak opioid5,6 or a strong opioid, depending on the administered dose.4

Efficacy of strong opioids

A recently published Cochrane systematic review identified 10 randomised controlled trials (involving a total of 2,268 patients) which compared opioid use with either placebo or no treatment in osteoarthritis of the knee or hip.7 The median treatment duration was 4 weeks. Four of the trials involved oral oxycodone; three oral codeine; two oral oxymorphone; one oral morphine; and one transdermal fentanyl. Trials involving tramadol were not included since these had already been the subject of an earlier Cochrane review, which concluded that benefits of tramadol were comparable to paracetamol. However, adverse events with tramadol greatly disadvantaged it compared to other treatments for osteoarthritis, except where use of slow titration regimens substantially reduced such events.8

Opioids (10 trials) were more effective in pain reduction than control interventions (standardised mean difference [SMD] 0.36, 95% CI 0.47 to 0.26).8 This corresponded to a difference in the pain scores of 0.9cm on a 10cm visual analogue scale (VAS) between opioids and placebo. The number-needed-to-treat (NNT) to produce one additional treatment response on pain was 25. Opioids also produced a larger improvement on function (seven trials, total of 1,794 patients; SMD 0.33, 95% CI 0.45 to 0.12). This corresponds to a difference in function scores of 0.7 units on a standardised Western Ontario and McMaster Universities (WOMAC) osteoarthritis disability scale (range 0–10). Thirty patients would need to be treated to produce one additional treatment response with regards to function. The reviewers found little evidence of association between the size of response and analgesic potency; route of administration; type of control intervention; treatment duration; use of “analgesic co-interventions”; or concealment of treatment allocation. They calculated that overall, patients on opioids received a median dose equivalent to morphine 51mg daily (range 13–160mg) and there was no evidence for a linear association between equivalent daily doses and pain reduction, and little evidence for an association with improvement in function.

In a meta-analysis of four of the trials (involving a total of 1,080 patients) that reported any adverse events, patients using opioids were more likely to experience such events compared with placebo (relative risk [RR] 1.55, 95% CI 1.41 to 1.70; number-needed-to-harm [NNH] 12, 95% CI 10 to16).7 Also, those receiving opioids (10 trials, total of 2,403 patients) were four times more likely to be withdrawn from treatment or drop out due to adverse events (RR 4.05, 95% CI 3.06 to 5.38; NNH 19, 95% CI 13 to 29). Oxycodone had the highest pooled risk ratio of adverse events compared to placebo (RR 7.75, 95% CI 3.76 to 15.97). The reviewers considered that the small to moderate beneficial effects of non-tramadol opioids were outweighed by large increases in the risk of adverse events. They advised that even in patients with severe osteoarthritic pain, clinicians should use non-tramadol opioids “cautiously” and consider alternatives, such as surgery. They stated that clinicians should inform patients about the substantial risks and only moderate benefits of opioid treatment and therapeutic alternatives.

Another systematic review assessed various types of drug intervention (including opioids and oral NSAIDs) compared with placebo in osteoarthritis of the knee.19 It included a meta analysis of six randomised placebo-controlled trials (involving a total of 1,057 patients), in which opioids (tramadol, morphine sulphate, oxycodone, oxymorphone and codeine) were more effective than placebo. At a mean of 2.8 weeks, the best mean difference of change for opioids over placebo was 10.5mm (95% CI 7.4 to 13.7) on a 100mm VAS. This appeared to be similar to the effects of oral NSAIDs (26 trials, total of 9,964 patients; a difference over placebo of 10.2mm, 95% CI 8.8 to 11.6, at a mean of 2.3 weeks). However, there were no direct comparisons between opioids and oral NSAIDs. Opioids had a high rate of drop-out or withdrawal due to adverse events: 23.6% at 2 weeks and 50.6% at 12 weeks. The reviewers did not recommend opioids for osteoarthritic knee pain.

Two other randomised controlled trials have assessed opioids in moderate to severe osteoarthritic pain. One was not included in the above meta-analyses due to insufficient number of patients with osteoarthritis of the knee or hip10 and the other11 was published since these analyses.

The first was a double-blind trial, involving a total of 133 patients, which compared oxycodone to placebo (45.9% of patients had disease of the spine or back and 30.8% disease of the knee).10 Oxycodone 20mg controlled-release tablets every 12 hours was more effective in reducing mean pain intensity (the primary outcome measure) from baseline of 2.4–2.5 points on a 4-point categorical scale than placebo or oxycodone 10mg at week 1 and week 2 (p<0.05).

The other study was a non-blinded non-inferiority* trial in which 135 patients received buprenorphine 7-day patches or prolonged-release tramadol tablets twice daily for 12 weeks.11 Doses were selected according to recommendations in the summary of product characteristics (SPC). The non-inferiority margin allowed buprenorphine to have a Box Scale 11 pain score (BS-11, said in the trial to be a “validated” measure for pain reduction) of up to 1.5 boxes below that of tramadol without being judged inferior. Mean reduction from baseline in the BS-11 pain score was 2.26 boxes with buprenorphine and 2.09 with tramadol in the intention-to-treat population (134 patients; a difference of 0.17, 95% CI 0.54 to +0.89), and 2.69 and 2.43 boxes in the per-protocol population (90 patients; difference of 0.26, 95% CI 0.59 to +1.11). The mean effective doses of each drug were not recorded in the trial report. These results showed non-inferiority of buprenorphine patches compared to tramadol tablets.

Unwanted effects

The most common unwanted effects of opioids include nausea and vomiting, constipation and dry mouth; also, larger doses can cause respiratory depression.12 Common unwanted effects include pruritus, rash, dizziness, confusion, drowsiness, hypotension, difficulty with micturition and urinary retention, visual disturbances, hallucinations and mood changes. These are often distressing to patients and may prevent many, especially older patients, from using stronger opioids appropriately.

Tolerance and dependence

A “structured evidence-based” review (including case-control, cohort, correlational and experimental data such as placebo controlled studies) assessed abuse or addiction, and aberrant drug-related behaviours (those which can indicate the development of addiction) in patients using opioid therapy for at least 1 month for non-malignant pain.13 Studies of tramadol (described as a “nonscheduled weak opioid”) and opioid antagonists were excluded. The reviewers concluded that the risk of abuse or addiction was quite low, being 3.27% (24 studies, total of 2,507 patients; an average opioid exposure time of 26.2 months). Around 17% of these studies preselected patients for no previous/current history of alcohol/illicit drug use; abuse or addiction occurred in 0.19% of patients in the preselected group compared to 5.0% in the not-selected group. Aberrant drug-related behaviours occurred in an average of 11.5% of patients (17 studies, total of 2,466 patients; average opioid exposure time of 10.8 months). Around 28% of these studies preselected patients for no previous/current history of alcohol/illicit drug use; again, fewer patients had one or more aberrant drug-related behaviours in the preselected group (0.59% vs. 11.2%).

The British National Formulary states that repeated use of opioids may cause tolerance and dependence, although this is rarely a problem with therapeutic use, and that “caution is advised” if prescribing for patients with a history of drug dependence.12 It advises avoiding abrupt withdrawal after long-term treatment. The Medicines and Healthcare products Regulatory Agency recently advised tighter control of over-the-counter sale of medicines containing codeine or dihydrocodeine, suggesting that using products containing these opioids for over 3 days exposes patients to a risk of addiction.14

What do guidelines say?

A guideline commissioned by the National Institute for Health and Clinical Excellence advises offering paracetamol for osteoarthritic pain if non-drug core interventions are inadequate.2 If paracetamol or topical NSAIDs are insufficient, then one suggested option at that stage is addition of an opioid analgesic (taking into consideration the risks and benefits, particularly in older patients). However, the guideline makes no distinction in its recommendations between weak and strong opioids. It states that the evidence for using opioids in the condition is “poor”, with virtually no good studies using opioids in peripheral joint osteoarthritis, and little evidence to suggest that dose escalation is effective. It also states that there are few data comparing different opioid formulations or routes of administration.

NHS Clinical Knowledge Summaries recommend that for opioids, codeine should be tried first, alone or together with paracetamol, and “specialist advice” should be obtained before prescribing stronger opioids such as morphine or fentanyl or buprenorphine patches for osteoarthritis.15

The Osteoarthritis Research Society International recommends that weak opioids “can be considered” for refractory pain from osteoarthritis of the knee or hip, where non-drug and drug options have been ineffective, or are contraindicated.3 The society recommends that stronger opioids should only be used for severe pain in exceptional cases, in addition to non-drug treatments, and that surgical treatment should be considered in these circumstances.

The European League Against Rheumatism recommends opioid analgesics (without specifying weak or strong opioid) given with or without paracetamol, as useful alternatives for those patients with osteoarthritis of the hip or knee in whom NSAIDs (including COX-2-inhibitors) are contraindicated, ineffective, and/or poorly tolerated.16,17

The Pain Society, the Royal College of Anaesthetists, the Royal College of General Practitioners and the Royal College of Psychiatrists published a joint recommendation on the appropriate use of opioids (including, possibly, strong opioids) for persistent non-cancer pain.4 This stated that hospital-based services should not start long-term opioid therapy without support from a multidisciplinary pain management service and liaison with the patient's primary care team. It also advised that in most cases, the day-to-day medical responsibility will lie with a GP once the patient is taking a stable dose of opioid, and that fixed supplies of opioids should be prescribed at fixed intervals. The recommendation stated that, although not precluding the use of opioids, patients (or those with household members) with current or relevant drug or alcohol problems, or psychiatric problems, should prompt careful assessment of the social situation and sensitive discussion about the potential for drug diversion.

Implications for practice

Blanket prescribing of strong opioids for pain management in osteoarthritis using strategies extrapolated from cancer or palliative care would raise concerns. Not least of these is that, in palliative care, there is no dose ceiling, and there is more community support to facilitate delivery and monitoring opioid use. Also, the unwanted effects of long-term usage for chronic non-cancer pain are largely unknown, which presents a major issue in osteoarthritis. Similarly, it can be argued that the World Health Organization analgesic ladder (a stepwise approach to the use of analgesic drugs) may not be appropriate in osteoarthritis.

Every effort should be made to ensure that appropriate non drug and drug management strategies for osteoarthritis have been tried, and referral to surgery has been considered before there is prolonged and established functional limitation and severe pain.2 It may be necessary to refer patients who are not candidates for joint replacement for specialist advice, for example, to a musculoskeletal team (which may comprise orthopaedic, rheumatology, physiotherapy and pain clinic input) or rheumatologists to confirm diagnosis of difficult cases. However, routine referral to secondary care specialists purely for initiation of strong opioid therapy might not be in the patients' best interest.

If a strong opioid is being considered for refractory osteoarthritic pain, GPs (through experience in prescribing strong opioids for cancer pain), are often best placed to discuss with patients this option or barriers to receiving strong opioids (for example, preconceptions that opioids are only for terminally ill people). To aid the assessment of whether patients are at high risk of opioid abuse, or for the prediction of aberrant behaviours during treatment, screening tools are available, such as those recently reviewed by the International Association for the Study of Pain.18 Treatment should be limited to short-term relief until surgery, and GPs will need to regularly assess continual need for strong opioids, and discontinue treatment or refer patients on where appropriate. Together with community pharmacists, GPs are in a good position to monitor patients for unwanted effects and check they are using it correctly, including safe storage and destruction of controlled drugs. Distressing unwanted effects such as nausea and vomiting or constipation will require concurrent treatment from the outset, with appropriate dose escalation.


Management of patients with osteoarthritic pain involves both non-drug and drug options. Commonly used analgesics may be inadequate or inappropriate and some patients will need to be referred for surgery to relieve refractory pain and restore function. There is limited evidence that strong opioids are effective in reducing pain in osteoarthritis; however, there is a lack of trials comparing strong opioids with each other in osteoarthritis, and assessing longer-term use. Unwanted effects with opioid use are common. However, the risk of developing opioid addiction with chronic use appears low, and can be reduced by carefully preselecting patients such as those without a history of abuse or addiction.

If a strong opioid is needed as adjuvantive treatment for osteoarthritic pain, its use should be relatively short-term pending surgical or other specialist referral. GPs are often in the best position to discuss such options with those patients with a low potential for abuse or addiction. Regular GP follow up will be needed to assess analgesic efficacy, minimise unwanted effects, assess the need for ongoing use or refer patients where appropriate.


    [R=randomised controlled trial; M=meta-analysis]

  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. M 7.
  8. M 8.
  9. M 9.
  10. R 10.
  11. R 11.
  12. 12.
  13. M 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.


  • * For an explanation of non-inferiority trials, see DTB 2008; 46: 55–6.

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.