Sunday, September 11, 2011

What Works Best for Acute Pain? Evidence Review.

Acute Pain A Cochrane Review of data from about 45,000 patients involved in approximately 350 individual clinical trials provides evidence of what to expect from commonly used pain relievers taken at specific doses for acute postoperative pain. The review also identifies analgesics for which there is only poor or no reliable evidence. Topping the list for pain relief were etoricoxib and oxycodone-plus-acetaminophen, while codeine scored the worst.

Working at the Oxford Pain Research Unit at Oxford University in the UK, R. Andrew Moore, MD, and colleagues analyzed the findings of 35 Cochrane Reviews of randomized clinical trials testing how well different analgesics work when used for moderate to severe postoperative pain [Moore et al. 2011]. Their review-of-reviews, or overview, is reported in the latest edition of The Cochrane Collaboration Library, which provides some of the most extensive and quantitatively rigorous investigations of clinical evidence.

The researchers identified 35 relevant Cochrane Reviews with 38 analyses of single-dose oral analgesics tested in acute postoperative pain models. Data included results from about 45,000 adult participants studied in approximately 350 individual trials judged to be of standardized design and high quality. Reliable results were obtained for 46 drug/dose combinations used to treat postsurgical conditions of at least moderate pain intensity; 45 were used for dental pain (largely after third molar extractions) and 14 were tested in non-dental settings for painful conditions (eg, episiotomy, orthopedic and abdominal surgery, etc). All comparisons were between active analgesic treatment and placebo.

Consolidated postoperative outcomes for each active treatment, reported as numbers-needed-to-treat, or NNTs, varied from about 1.5 to 20.0 for at least 50% maximum pain relief — that is, attaining at least half the maximum possible pain relief on whatever pain scale was used in the trial — during 4 to 6 hours, compared with placebo. The proportion of participants achieving this level of benefit varied from about 30% to over 70%, and the time to needing a second dose of medication, or re-medication, varied from 2 hours (placebo) to more than 20 hours.

The following figure lists in rank order NNTs (and their Confidence Intervals) for the 46 drug/dose combinations, reflecting at least 50% maximum pain relief during 4 to 6 hours following all types of surgery as compared with placebo. [Note: paracetamol is the same as acetaminophen.] NNTs Acute Pain

As a reminder, NNT as a concept has been discussed previously [eg, see Pain-Topics UPDATE here] and is based on the inverse of the Absolute Risk Reduction (1/ARR). In this Cochrane Review, the ARR, or benefit of therapy, is the proportion of patients achieving up to 50% maximum pain relief for 4 to 6 hours in the active treatment group minus the proportion similarly benefitting in the placebo group. So, the NNT represents how many patients would need to receive the respective analgesic in order for one of them to benefit, as compared with all of them having received placebo instead. For example, an NNT=2.0 suggests that 1 of every 2 actively treated patients would benefit from the analgesic being considered. Lengthier bars, denoting wider Confidence Intervals, reflect less precision in the NNT estimates due to smaller numbers of participants assessed in the respective reviews.

In their overview, Moore et al. considered an NNT≥10.0 as being clinically irrelevant or unacceptable, and codeine 60 mg was the only drug exceeding this limit. The several drug/dose combinations with the best (lowest) NNTs were (in order): etoricoxib 180 mg or 240 mg (NNT=1.5; 95% Confidence Interval 1.3 to 1.7), oxycodone 10 mg + paracetamol 1000 mg (1.8; 1.6 to 2.2), etoricoxib 120 mg (1.9; 1.7 to 2.1), and ketoprofen 25 mg (2.0; 1.8-2.3).

Those with the longest duration of action (≥8 hours) included etoricoxib 120 mg, diflunisal 500 mg, oxycodone 10 mg + paracetamol 650/1000 mg, naproxen 500/550 mg, and celecoxib 400 mg. Whereas, oxycodone 5 mg, codeine 60 mg, and paracetamol 600/650 exhibited the shortest duration of time to re-medication (<4 hours).

Participants reporting at least one adverse effect were few and generally no different between active drug and placebo. There were a few exceptions as follows (along with their NNH, or number needed to treat for one patient to experience an adverse effect, or harm): aspirin 600/650 mg (NNH=44), diflunisol 1000 mg (NNH=7.7), dihydrocodeine 30 mg (NNH=7.4), codeine + paracetamol (NNH=6.0), and oxycodone + paracetamol combinations (NNH= 3.5 to 4.5).

The overview authors conclude that there is a wealth of reliable evidence on the analgesic efficacy of select single-dose oral analgesics in postoperative settings and these generally exhibit favorable safety profiles. These data should help to inform choices by professionals and consumers; however, not all participants in the many studies examined had good pain relief and for many drug/dose combinations half or more did not achieve at last 50% maximum pain relief during four to six hours.

Perhaps most important, Moore states, “If the first [pain reliever] a person tries doesn’t seem to be working, then a doctor should look to find an alternative reliable drug and see if it is more effective in that individual patient. There are plenty of options that have a solid evidence base.”

COMMENTARY: This sort of Cochrane Review — a review-of-reviews, or overview — was a massive undertaking and provides a portrayal of best evidence within certain limitations. Moore and colleagues note that their overview focuses on results from all available, high quality trials, but they concede that some drugs/dosages are not typically used clinically, such as ibuprofen 100 mg or etoricoxib 180/240 mg (which provided the best NNT). And, in some cases, certain drugs are no longer marketed (eg, rofecoxib or dextropropoxyphene) while other drugs are not available worldwide; for example, etoricoxib in approved in 70 countries but not yet in the United States. As the authors note, “All data are presented so that readers can use that which is relevant for them.”

However, certain analgesics were not adequately examined by prior studies for Moore et al. to include them in their analyses. For example, there were review studies but no extensive trial data available for acemetacin, meloxicam, nabumetone, nefopam, sulindac, tenoxicam, and tiaprofenic acid. Inadequate amounts of data also were available for dexibuprofen, dextropropoxyphene 130mg, diflunisal 125mg, etoricoxib 60 mg, fenbufen, and indometacin. The authors noted two newer, successful studies of ibuprofen 200 mg + paracetamol 500 mg for dental pain (NNT, 95% CI; 1.6, 1.4-1.8), but these were not part of Cochrane reviews and, hence, were omitted from their analyses.

As would be expected, in all analgesic trials there also were responses to placebo. These effects — the percentage achieving at least 50% maximum pain relief over 4 to 6 hours with placebo— were consistent across the various studies, with most falling between 5% and 19%, but much greater placebo responses were observed in reviews of ibuprofen, flurbiprofen, paracetamol, and prioxicam. However, the higher rates (>20%) might have been artifacts of study design.

Only orally-administered analgesics were considered for this overview, and in adult populations, which excluded pediatrics. Furthermore, while adverse events did not appear to be a major issue, the focus was on single-dose administration of the drugs, which would not reflect potential adverse effects with longer-term use, and this needs careful consideration.

It also is important to carefully examine how dosages and combinations affect NNTs. In some cases, higher doses of the same agent greatly improve response. And, for example, while codeine 60 mg alone was largely ineffective, codeine 60 mg + paracetamol 800 mg or 1000 mg was fairly high on the list (NNT≈2.3).

Some may question whether 50% maximum pain relief is a clinically important target that is acceptable to most patients; since many patients might expect nearly complete relief. Another limitation is that the included research trials examined a somewhat narrow model of acute pain, eg, following dental and select operative procedures, whereas acute pain as a category may encompass much more varied situations. Therefore, it might be inappropriate to generalize the findings of this overview to all patients and/or to other types of pain expected to be of an acute nature — eg, minor sports injuries, other types of aches and strains, menstrual pain, headache, etc. — or even to other types of postoperative pain that where not examined in the trials.

Still, the information here might serve as a starting point, with the understanding that individual responses can vary and if one oral analgesic is ineffective there are many others worth trying. As noted above, the results for moderate to severe acute pain with single-dose administration of the 46 agents examined ranged from 30% to 70% of patients achieving pain relief, and the duration of pain relief varied from about 2 hours to about 20 hours.

We noted in previous UPDATES [here] and [here] that the inadequate treatment of postsurgical pain is somewhat of a worldwide crisis. Clearly, no particular analgesic produces high levels of acute pain relief in ALL patients and reasons for this are varied, but patients in pain (and their healthcare providers) should not be surprised if drugs they are given do not work for them and alternative analgesic drugs should be tried. Or, as Moore notes, “Pain relief doesn’t have to be a mystery. There is a body of reliable evidence about how well 46 different drug/dose combinations work against acute pain, but the review also shows there are many examples of drugs for which there is insufficient evidence, and [those] drugs in question should probably not be used to treat acute pain.”

REFERENCE: Moore RA, Derry S, McQuay HJ, Wiffen PJ. Single dose oral analgesics for acute postoperative pain in adults. Cochrane Database of Systematic Reviews. 2011;9(CD008659) [available here].