Comparative efficacy of treatments for osteoarthritis (Part 1)

As far as I can recall, paracetamol (or what is known as acetaminophen in other parts of the world) has always been recommended as the initial drug of choice for patients with mild-to-moderate knee osteoarthritis by Harrison’s Principles of Internal Medicine – THE bible for internal medicine training programs in the Philippines (the very first edition I had was the 14th, residents nowadays carry the 19th ed.). Even local practice guidelines (2008 by Penserga et al for the Philippine Rheumatology Association) made the same recommendation.

However, it is very well known that NSAIDs are 25% better in reducing pain compared to paracetamol. But NSAIDs also have more adverse events even when used in recommended doses. This, I believe, is the major point for starting treatment for OA with paracetamol.

When I started speaking for industry, I’d always have a slide showing that comparable levels of pain relief can be obtained when using NSAIDs in the following daily doses:

  • Naproxen 1000 mg
  • Diclofenac 150 mg
  • Ibuprofen 2400 mg
  • Celecoxib 200 mg
  • Etoricoxib 60 mg

This was based on an earlier AHCRP review for treatment of knee OA. And I’d emphasize that the choice of analgesic then would be based on what’s safe for the patient given his/ her other comorbids (No NSAIDs in patients with a history of MI or CVD-infarcts. Coxibs or  NSAIDs with proton pump inhibitors in patients with dyspepsia).

But in early 2015, an article piqued my interests as it was the first I’ve read comparing systemic and intra-articular interventions for knee OA. Up to that time, I was just aware of the utility of viscosupplementation for knee OA but had an incomplete idea of how it fared compared to other treatments. Instead of a forrest plot, the article presented the effect size for various treatments which would give us an idea of how much better a particular treatment is compared to others. This challenged my prior concept of equipotent analgesia for OA as the article suggests that some agents are better than others in effecting pain relief.

Briefly, the effect size (and their 95% confidence intervals) for pain at 3 months of various treatments for osteoarthritis (compared to placebo) are as follows:

  • Paracetamol 0.18 (0.04 – 0.33)
  • Intra-articular placebo 0.29 (0.04 – 0.54)
  • Celecoxib 0.33 (0.25 – 0.42)
  • Naproxen 0.38 (0.27 – 0.49)
  • Ibuprofen 0.44 (0.25 – 0.63)
  • Diclofenac 0.52 (0.34 – 0.69)
  • Intra-articular steroids 0.61 (0.32 – 0.89)
  • Hyaluronic acid 0.63 (0.39 – 0.88)

Looking at the confidence intervals, all treatments were much better than oral placebo in producing analgesia. The above treatments are in the order of increasing efficacy – meaning that the least effective in reducing pain was paracetamol while the most effective is intra-articular hyaluronate.

Additionally, in interpreting effect size, the difference between groups is considered small if <=0.20, moderate if <=0.50 and large if <=0.80. So all treatments are better than giving placebo BUT the difference in pain relief afforded by either placebo and paracetamol is small. Of all oral agents listed, diclofenac afforded the best pain relief at 3 months. However, the best analgesic of all is actually visco-supplementation with hyaluronate. Interestingly, simply performing joint aspiration (i.e. intra-articular placebo) resulted in better pain relief compared to paracetamol – so the procedure even without steroids is therapeutic aside from being diagnostic.

An interesting conclusion of the article is the comparison of other interventions to paracetamol as an analgesic. The results were as follows:

  • Intra-articular placebo 0.11 (-0.17 – 0.38)
  • Celecoxib 0.15 (0 – 0.30)
  • Naproxen 0.20 (0.03 – 0.37)
  • Ibuprofen 0.26 (0.05 – 0.47)
  • Diclofenac 0.33 (0.12 – 0.54)
  • Intra-articular steroids 0.42 (0.12 – 0.73)
  • Hyaluronic acid 0.45 (0.18 – 0.72)

Looking at the confidence intervals of IA placebo and celecoxib, it appears there was no significant difference in analgesic effect between these two interventions and paracetamol. All other agents evaluated had a significant difference (vs paracetamol) in relief afforded ranging from small (naproxen) to moderate (the rest). Again, diclofenac was the best oral agent and hyaluronate was the best intervention over-all.

All interventions EXCEPT intra-articular corticosteroids were better at improving function at 3 months. All NSAIDs (including celecoxib) were better than paracetamol in improving function.

Looking at the side effects, there were more GI adverse events and withdrawals due to adverse events in the oral non-selective NSAIDs compared to placebo and paracetamol. While, there was no difference in rates observed between paracetamol and celecoxib. The short exposure times (<= 3 months) may have been insufficient to document other AEs of interest like MI, stroke and death.

Obviously missing in the analysis are etoricoxib (approved in the Philippines for use in osteoarthritis) and tramadol (including the now ubiquitous paracetamol + tramadol combination which a lot of doctors are using). The review was done in the US where etoricoxib was not approved by their FDA for marketing. However, I do suspect that its effect size may be close to diclofenac as most (if not all) its studies had diclofenac as a reference drug. For tramadol, I’m hoping that future analyses include it as real world experience can attest to its efficacy.

What are the important messages of the study by Bannuru et al?

  • All interventions are better than oral placebo in improving pain among patients with knee OA.
  • Not all NSAIDs are better than paracetamol in providing pain relief (Yes, I’m looking at you, celecoxib) but the trade off is a similar safety profile up to 3 months.
  • I can no longer speak of equi-potent analgesia (Time to delete that slide?)
  • Intra-articular agents (steroids and hyaluronate) are better than oral agents for pain relief.
  • There may be a disconnect between analgesia and improvements in function. All NSAIDs (including celecoxib) were better than paracetamol in improving function. But IA steroids, while better than ALL oral agents in producing relief, was no better than oral placebo in improving function. It may be that function is not entirely dependent of absence of pain and other mechanisms may be in play.

Comparative effectiveness of pharmacologic interventions for knee osteoarthritis: a systematic review and network meta-analysis.


Author: Sids Manahan MD 🇵🇭

Rheumatology. Internal Medicine. Educating Patients and Colleagues. Curating Rheumatology. Bloggero-Wanabe.

Leave a Reply

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

You are commenting using your 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