NSAIDs don’t work for back pain

There is an urgent need to develop clinically effective and safe drug therapies to treat back pain, say Australian researchers

A systematic review and meta-analysis conducted by The George Institute for Global Health has added NSAIDs to the pile of ineffective drugs for back pain.

Anti-inflammatories will be joining paracetamol, which the British Medical Journal, The Lancet and a Cochrane Review also ruled out as an effective treatment for back pain in 2015.

This latest study of 35 randomised placebo-controlled trials, published in the Annals of the Rheumatic Diseases, found that while NSAIDs reduced pain and disability, the effects compared with placebo were clinically unimportant.

Meanwhile, patients taking NSAIDs were 2.5 times more likely to suffer from gastrointestinal problems such as stomach ulcers and bleeding.

Lead author, Associate Professor Manuela Ferreira from the University of Sydney, says that while back pain is the leading cause of disability and is commonly managed by prescribed anti-inflammatories, the results show they only provide very limited short-term pain relief.

“They do reduce the level of pain, but only very slightly, and arguably not of any clinical significance,” says A/Prof Ferreira.

“When you factor in the side effects which are very common, it becomes clear that these drugs are not the answer to providing pain relief to the many millions of Australians who suffer from this debilitating condition every year.”

She says the study highlight an urgent need to develop new therapies to treat back pain.

Fellow author Gustavo Machado, from The George Institute and the University of Sydney’s School of Medicine, agrees.

“Millions of Australians are taking drugs that not only don’t work very well, they’re causing harm. We need treatments that will actually provide substantial relief of these people’s symptoms.

“Better still we need a stronger focus on preventing back pain in the first place. We know that education and exercise programs can substantially reduce the risk of developing low back pain.”

Most clinical guidelines currently recommend NSAIDs as the second-line analgesics after paracetamol, with opioids coming at third choice, The George Institute points out.

Previous Clinical tips: complementary medicines
Next Forum: Oh no they don't!

NOTICE: It can sometimes take awhile for comment submissions to go through, please be patient.


  1. Andrew

    Peter Gøtzsche has campaigned widely on the dodgy design and fundamentals of many of the large-scale trials that underpin pharmacotherapy. Not long ago paracetamol was declared somewhere between placebo and not great for back pain and now we have similar findings for ibuprofen.

    I wonder how many other drugs in the dispensary will have similar findings…..and the implications for pharmacy if other blockbuster molecules are found to no better than placebo?

    Any guesses on which group is next? Some antipsychotics are worth a closer look I reckon.

    • Wayne Sturges

      Andrew, I agree with you. Some products are still worth trying. People come to us for a solution to their problem and if we follow all these studies we would be sending them all away and just tell them to stretch yourself. By the way that works pretty good for dogs and cats, first thing they do after getting up from a nap. I still fell that Paracetamol/Ibuprofen combo works quite well with most, strongly recommend they stay away from codeine and yes I do explain the benefits of education and stretching exercises. No easy answer except that people expect more and more from pills and potions these days. Wind back the clock the the 70’s and 80’s when I trained and first started out, we had none of these marvelous compounds to sell. It was Panadol and Panadeine, that’s it end of story. Please don’t tell me that bad backs are worse today than 40 years ago!!!!!

      • Expat In Blighty

        Yeah, sure…until the last sentence.
        Worth consideration is the clickable graphic here:
        Summary, “In 1980, around 60% of Australian adults had a healthy weight; today
        this has almost halved to around 35%. In 1980, just 10% of adults were
        obese. In 2012, this figure tips 25%.”
        Quote from an American Journal of Epidemiology meta-analylis, “Findings indicate that overweight and obesity increase the risk of low back pain”.
        To me, that implies that bad backs are indeed worse than 40 years ago since there is a shift along the distribution curve and so the overall
        number, and hence people coming into a pharmacy seeking treatment, has
        increased. And if there is a link between body weight and pain score as implied by the following quote from a study, then the intensity of pain may indeed be increased. “Compared with participants who ranked as low-to-normal weight, the
        overweight group was 20 percent more likely to report pain yesterday,
        while the obesity-I group reported 68 percent higher rates of pain, a
        number that climbed to 136 percent and 254 percent for obesity-II and
        obesity-III groups, respectively.” Multiple studies have shown the positive relationship with increased body weight and higher pain score but inconclusive as to causation.

  2. geoff

    Without time to look at this particular study, many debunking studies claim a medication is no better than placebo so should not be supported and then state that 30-50% of participants had improved function or felt better with the placebo…..logically 30-50% of participants taking the medication must have also improved. It would be unprofessional to sell the placebo so it makes sense to sell the medication (having regard to any side effects) with positive reinforcement to increase the (possibly placebo) effect.
    Start telling people this medication won’t work and will have side effects and it will become self-fulfilling (nocebo)

  3. vixeyv

    Really interesting finding soon after having codeine upscheduled to S4…. going to cause a lot more back pain… a pain in the arse!

  4. Toorisugarino Isha

    “NSAIDS DON’T WORK FOR BACK PAIN” the headline screams. Is that the conclusion of this study? The lead author is even quoted in the body of this article saying that “they do reduce the level of pain, but only very slightly and arguably not of any clinical significance”. The operative word here is “arguably” especially since the study did find a small but statistically significant benefit of NSAIDS over placebo in reducing pain and disability.

    A few caveats I noted from reading the actual study:
    – There was significant heterogeneity in the studies being analysed (overall I-square 59% and 81% for immediate and short term pain scores respectively), meaning that you aren’t comparing apples with apples
    – The primary analysis pooled results from 35 studies including three where the comparitor NSAID was in a topical preparation. Results from studies looking at non-selective and COX2 selective NSAIDS were pooled together
    – Short median follow up (1 week) – in some studies the treatment schedule was < 24 hours
    – The threshold for clinical significance was set at 10 points (on a 0-100 scale) and appears to have been derived from a previous study for lower back pain and it appears it has been applied to all outcomes in this study
    – Despite this, there is a clear signal favouring NSAIDS over placebo for the primary outcomes (pain and disability scores in the immediate and short term) though it does not reach the pre-specified threshold
    – There was a clinically significant reduction in pain score for immediate relief of neck pain (MD -16.3 (95% CI -20.6 to -12.0))
    – Overall no difference in adverse event rates for NSAID vs placebo, however higher number of GI adverse events with NSAIDS (RR 2.5 95% CI 1.2 to 5.2)

    In summary, it's not as black and white as the headline suggests.

    Pain is a subjective beast. A 9.2 point reduction in pain score (on a 0-100 scale) may mean alot for one person, not so much for another. Also, sneaking a few studies with topical NSAIDS into the pooled analysis may have diluted the effect size. Before we write off NSAIDS completely, there's clearly room for more studies to better identify subgroups that may benefit (eg. neck pain).

    The cynic in me thinks that after the PACE study, the George Institute now has an agenda making the headlines disproving the effectiveness of pharmactherapy for back pain, hence the heavy spin put on this study.

Leave a reply