PMCCPMCCPMCC

Search tips
Search criteria 

Advanced

 
Logo of bmjclinevidLink to Publisher's site
 
BMJ Clin Evid. 2007; 2007: 1108.
Published online 2007 June 1.
PMCID: PMC2943791

NSAIDs

Peter C Gøtzsche, MD MSc, Director and Professor

Abstract

Introduction

NSAIDs are widely used. Almost 10% of people in The Netherlands used a non-aspirin NSAID in 1987, and the overall use was 11 defined daily doses per 1000 population per day. In Australia in 1994, overall use was 35 defined daily doses per 1000 population a day, with 36% of the people receiving NSAIDs for osteoarthritis, 42% for sprain and strain or low back pain, and 4% for rheumatoid arthritis; 35% of the people receiving NSAIDs were aged >60 years.

Methods and outcomes

We conducted a systematic review and aimed to answer the following clinical questions: Are there any important differences between NSAIDs? What are the effects of topical NSAIDs; and of co-treatments to reduce the risk of gastrointestinal adverse effects of NSAIDs? We searched: Medline, Embase, The Cochrane Library, and other important databases up to September 2009 (Clinical Evidence reviews are updated periodically, please check our website for the most up-to-date version of this review). We included harms alerts from relevant organisations such as the US Food and Drug Administration (FDA) and the UK Medicines and Healthcare products Regulatory Agency (MHRA).

Results

We found 36 systematic reviews, RCTs, or observational studies that met our inclusion criteria. We performed a GRADE evaluation of the quality of evidence for interventions.

Conclusions

In this systematic review we present information relating to the benefits and harms of the following interventions: alternative analgesics, H2 blockers, misoprostol, NSAIDs (systemic, topical, differences in efficacy between, dose-response relationship of), proton pump inhibitors.

Key Points

Non-steroidal anti-inflammatory drugs (NSAIDs) inhibit the cyclo-oxygenase (COX) enzyme to exert their anti-inflammatory, analgesic, and antipyretic effects.

No important differences in efficacy have been demonstrated between different systemic NSAIDs in the management of musculoskeletal disorders.

  • There seems to be a plateau for effectiveness, with recommended doses close to those required for maximal effectiveness. However, the risk of adverse effects increases with increasing dose, with no plateau.
  • Systemic NSAIDs that selectively inhibit COX-2 have a reduced risk of causing gastrointestinal ulcers compared with less-selective NSAIDs. However, COX-2 inhibitors increase the risk of myocardial infarction and other cardiovascular events.
  • Paracetamol is less effective than systemic NSAIDs at reducing pain in osteoarthritis, but similarly effective for acute musculoskeletal pain.

Misoprostol reduces serious NSAID-related gastrointestinal complications and symptomatic ulcers compared with placebo, but is itself associated with adverse effects including diarrhoea, abdominal pain, and nausea.

  • Proton pump inhibitors and H2 antagonists have been shown to reduce endoscopic ulcers in people taking NSAIDs, but their clinical benefits are less clear.
  • We don't know which treatment is the most effective at reducing gastrointestinal adverse effects from systemic NSAIDs.

We don't know whether topical NSAIDs are beneficial.

About this condition

Definition

Non-steroidal anti-inflammatory drugs (NSAIDs) have anti-inflammatory, analgesic, and antipyretic effects, and they inhibit platelet aggregation. This review deals specifically with the use of NSAIDs for the treatment of the symptoms of musculoskeletal conditions. NSAIDs have no documented effect on the course of musculoskeletal diseases. NSAIDs inhibit the enzyme cyclo-oxygenase (COX), which has two known isoforms: COX-1 and COX-2. NSAIDs are often categorised according to their ability to inhibit the individual isoforms, with newer NSAIDs often predominantly inhibiting the COX-2 isoform and older NSAIDs often being less specific inhibitors.

Incidence/ Prevalence

NSAIDs are widely used. Almost 10% of people in The Netherlands used a non-aspirin NSAID in 1987, and the overall use was 11 defined daily doses per 1000 population per day. In Australia in 1994, overall use was 35 defined daily doses per 1000 population a day, with 36% of the people receiving NSAIDs for osteoarthritis, 42% for sprain and strain or low back pain, and 4% for rheumatoid arthritis; 35% of the people receiving NSAIDs were aged >60 years.

Aims of intervention

To reduce symptoms in rheumatic disorders; to avoid severe gastrointestinal adverse effects.

Outcomes

Primary outcomes: pain intensity; personal preference for one drug over another; global efficacy; clinically significant gastrointestinal complications. Secondary outcomes: number of tender joints; perforation; gastrointestinal haemorrhage; dyspepsia; and ulcer detected by routine endoscopy; other adverse effects.

Methods

Clinical Evidence search and appraisal September 2009. The following databases were used to identify studies for this systematic review: Medline 1966 to September 2009, Embase 1980 to September 2009, and The Cochrane Database of Systematic Reviews 2009, Issue 3 (1966 to date of issue). An additional search within the Cochrane Library was carried out for the Database of Abstracts of Reviews of Effects (DARE) and Health Technology Assessment (HTA). We also searched for retractions of studies included in the review. Abstracts of the studies retrieved from the initial search were assessed by an information specialist. Selected studies were then sent to the contributor for additional assessment, using pre-determined criteria to identify relevant studies. Study design criteria for inclusion in this review were: published systematic reviews of RCTs. We included systematic reviews of RCTs where harms of an included intervention were studied applying the same study design criteria for inclusion as we did for benefits. For comparisons between paracetamol and an NSAID for musculoskeletal disorders, we included individual RCTs, as very few trials have been published. This review focuses on the use of NSAIDs for the treatment of the symptoms of musculoskeletal disorders. We have presented data regarding our primary outcomes in preference to data on secondary outcomes where available. More than 100 systematic reviews and thousands of RCTs have compared various NSAIDs. Owing to the large volume of data, in the question on different systemic NSAIDs compared with each other, we assess NSAIDs for the indication of rheumatoid arthritis only as results in this population are generalisable to other musculoskeletal conditions. We assess the use of topical NSAIDs for any musculoskeletal condition. We have not included any systematic reviews that assess single NSAIDs — again, owing to the larger volume of data. We have included systematic reviews and RCTs of the harms of NSAIDs (e.g., gastrointestinal toxicity) not solely in people with musculoskeletal disorders if the evidence would otherwise have been insufficient. Many RCTs are unpublished or published in sources not indexed in publicly available databases. The quality of the RCTs is variable and bias is common, both in the design and analysis of the RCTs, to such an extent that one systematic review identified false significant findings favouring new drugs over control drugs in 6% of RCTs. We have favoured systematic reviews that have not been sponsored or authored by the pharmaceutical industry, because bias in such reviews has repeatedly been shown but may be difficult to detect. For example, it is easy seemingly to follow the rules for systematic reviews and yet use inclusion and exclusion criteria that omit inconvenient studies. Furthermore, one systematic review found that industry-supported meta-analyses of drugs are of lower methodological quality, pay less attention to bias, and have fewer reservations than Cochrane reviews of the same drugs.We have also favoured higher-quality reviews, based on search rigour and appropriate meta-analysis, over poorer-quality reviews with more-recent search dates. In addition, we use a regular surveillance protocol to capture harms alerts from organisations such as the US Food and Drug Administration (FDA) and the UK Medicines and Healthcare products Regulatory Agency (MHRA), which are added to the review as required. To aid readability of the numerical data in our reviews, we round many percentages to the nearest whole number. Readers should be aware of this when relating percentages to summary statistics such as relative risks (RRs) and odds ratios (ORs). We have performed a GRADE evaluation of the quality of evidence for interventions included in this review (see table ). The categorisation of the quality of the evidence (into high, moderate, low, or very low) reflects the quality of evidence available for our chosen outcomes in our defined populations of interest. These categorisations are not necessarily a reflection of the overall methodological quality of any individual study, because the Clinical Evidence population and outcome of choice may represent only a small subset of the total outcomes reported, and population included, in any individual trial. For further details of how we perform the GRADE evaluation and the scoring system we use, please see our website (www.clinicalevidence.com). We believe that we have presented sufficient data to answer the question on the efficacy and safety of NSAIDs (both oral and topical) and that reporting further confirmatory data is not beneficial to our readers. We will continue to scan the literature, including drug safety monitoring, and will update the text of this review only should any high-quality evidence be published that reflects a change to the evidence we currently present.

Table
GRADE evaluation of NSAIDs.

Glossary

Defined daily dose
The assumed average daily dose for the main indication of a specified drug. The defined daily dose per 1000 population a day is an estimate of the proportion of that population receiving treatment with that drug.
High-quality evidence
Further research is very unlikely to change our confidence in the estimate of effect.
Moderate-quality evidence
Further research is likely to have an important impact on our confidence in the estimate of effect and may change the estimate.
Very low-quality evidence
Any estimate of effect is very uncertain.

Notes

Disclaimer

The information contained in this publication is intended for medical professionals. Categories presented in Clinical Evidence indicate a judgement about the strength of the evidence available to our contributors prior to publication and the relevant importance of benefit and harms. We rely on our contributors to confirm the accuracy of the information presented and to adhere to describe accepted practices. Readers should be aware that professionals in the field may have different opinions. Because of this and regular advances in medical research we strongly recommend that readers' independently verify specified treatments and drugs including manufacturers' guidance. Also, the categories do not indicate whether a particular treatment is generally appropriate or whether it is suitable for a particular individual. Ultimately it is the readers' responsibility to make their own professional judgements, so to appropriately advise and treat their patients.To the fullest extent permitted by law, BMJ Publishing Group Limited and its editors are not responsible for any losses, injury or damage caused to any person or property (including under contract, by negligence, products liability or otherwise) whether they be direct or indirect, special, incidental or consequential, resulting from the application of the information in this publication.

References

1. Leufkens HG, Ameling CB, Hekster YA, et al. Utilization patterns of non-steroidal anti-inflammatory drugs in an open Dutch population. Pharm Weekbl Sci 1990;12:97–103. [PubMed]
2. McManus P, Primrose JG, Henry DA, et al. Pattern of non-steroidal anti-inflammatory drug use in Australia 1990–1994. A report from the drug utilization sub-committee of the pharmaceutical benefits advisory committee. Med J Aust 1996;164:589–592. [PubMed]
3. Gøtzsche PC. Methodology and overt and hidden bias in reports of 196 double-blind trials of nonsteroidal anti-inflammatory drugs in rheumatoid arthritis. Control Clin Trials 1989;10:31–56. Search date 1985; primary source Medline. [Erratum in Control Clin Trials 1989;10:356] [PubMed]
4. Smith GD, Matthias E. Meta-analysis: unresolved issues and future developments. BMJ 1998;316:221–225. [PMC free article] [PubMed]
5. Jørgensen AW, Hilden J, Gøtzsche PC. Cochrane reviews compared with industry supported meta-analyses and other meta-analyses of the same drugs: systematic review. BMJ 2006;333:782. [PMC free article] [PubMed]
6. Gøtzsche PC. Patients' preference in indomethacin trials: an overview. Lancet 1989;i:88–91. Search date 1985; primary sources Medline, personal contact with companies that marketed proprietary products, and hand searches of reference lists of collected articles.
7. Garner S, Fidan D, Frankish R, et al. Celecoxib for rheumatoid arthritis (Cochrane review). In: The Cochrane Library, Issue 4, 2002. Chichester, UK: John Wiley & Sons, Ltd. Databases searched up to August 2002: Medline, Embase, The Cochrane Library, Cochrane Controlled Trials Register, National Research Register, NHS Economic Evaluation Database, Health Technology Assessment Database, bibliographies of retrieved papers, and content experts were consulted for additional references.
8. Garner S, Fidan D, Frankish R, et al. Rofecoxib for rheumatoid arthritis. (Cochrane review). In: The Cochrane Library, Issue 4, 2004. Chichester, UK: John Wiley & Sons, Ltd. Electronic databases searched up to December 2000: Medline, Embase, The Cochrane Library, Cochrane Controlled Trials Register, National Research Register, NHS Economic Evaluation Database, Health Technology Assessment database, and bibliographies of retrieved papers were scanned for additional references. The manufacturers of rofecoxib, MSD, were also approached by the UK National Institute for Clinical Excellence to submit additional evidence to inform it's appraisal on the use of cyclo-oxygenase inhibitors for arthritis.
9. Simon LS, Weaver AL, Graham DY, et al. Anti-inflammatory and upper gastrointestinal effects of celecoxib in rheumatoid arthritis: a randomized controlled trial. JAMA 1999;282:1921–1928. [PubMed]
10. Emery P, Zeilder H, Kvein TK, et al. Celecoxib versus diclofenac in long-term management of rheumatoid arthritis; randomised double-blind comparison. Lancet 1999;354;2106–2111. [PubMed]
11. Goldstein LJ, Correa P, Zhao WW, et al. Reduced incidence of gastroduodenal ulceration with celecoxib, a novel cyclooxygenase-2 inhibitor, compared to naproxen in patients with arthritis. Am J Gastroenterol 2001;96;1019–1027. [PubMed]
12. Gøtzsche PC. Meta-analysis of NSAIDs: contribution of drugs, doses, trial designs, and meta-analytic techniques. Scand J Rheumatol 1993;22:255–260. Search date 1988; primary source Medline. [PubMed]
13. Towheed TE, Hochberg MC, Shea BJ, et al. Analgesia and non-aspirin, non-steroidal anti-inflammatory drugs for osteoarthritis of the hip. In: The Cochrane Library, Issue 4, 1997. Oxford: Update Software.
14. Watson M, Brookes ST, Faulkner A, Kirwan J. Non-aspirin, non-steroidal anti-inflammatory drugs for treating osteoarthritis of the knee. In: The Cochrane Library, Issue 1, 1997. Oxford: Update Software.
15. Ogilvie Harris DJ, Gilbart M. Treatment modalities for soft tissue injuries of the ankle: a critical review. Clin J Sport Med 1995;5:175–186. Search date 1993; primary sources Medline and Embase. [PubMed]
16. Green SE, Buchbinder R, Forbes A, et al. Interventions for shoulder pain. In: The Cochrane Library, Issue 2, 1999. Oxford: Update Software. Search date 1998; primary sources Cochrane Musculoskeletal Group trials register, Cochrane Controlled Trials Register; Medline; Embase; Cinahl; Science Citation Index) up to May 1998; and hand searches of major textbooks, bibliographies of relevant literature, the fugitive literature, and the subject indices of relevant journals including: American College of Rheumatology, British College of Rheumatologists, the Biennial Conference of the Manipulative Physiotherapy Association of Australia, International Federation of Manual Therapists conference proceedings, British Orthopaedic Association, and American Orthopaedic Association.
17. Woo WWK, Man S-Y, Lam PKW, et al. Randomized double-blind trial comparing oral paracetamol and oral nonsteroidal antiinflammatory drugs for treating pain after musculoskeletal injury. Ann Emerg Med 2005;46:352–361. [PubMed]
18. Dalton JD, Schweinle JE. Randomized controlled noninferiority trial to compare extended release acetaminophen and ibuprofen for the treatment of ankle sprains. Ann Emerg Med 2006;48:615–623. [PubMed]
19. Hooper L, Brown TJ, Elliott RA, et al. The effectiveness of five strategies for the prevention of gastrointestinal toxicity induced by non-steroidal anti-inflammatory drugs: systematic review. BMJ 2004;329:948–952. [PMC free article] [PubMed]
20. Jüni P, Nartey L, Reichenbach S, et al. Risk of cardiovascular events and rofecoxib: cumulative meta-analysis. Lancet 2004;364:2021–2029. [PubMed]
21. Caldwell B, Aldington S, Weatherall M, et al. Risk of cardiovascular events and celecoxib: a systematic review and meta-analysis. J R Soc Med 2006;99:132–140. [PMC free article] [PubMed]
22. Bresalier RS, Sandler RS, Quan H, et al. Adenomatous Polyp Prevention on Vioxx (APPROVe) Trial Investigators. Cardiovascular events associated with rofecoxib in a colorectal adenoma chemoprevention trial. N Engl J Med 2005;352:1092–1102. [PubMed]
23. Jüni P, Rutjes AWS, Dieppe PA. Are selective COX 2 inhibitors superior to traditional non steroidal anti-inflammatory drugs? BMJ 2002;325:163–164. [PMC free article] [PubMed]
24. Kearney PM, Baigent C, Godwin J, et al. Do selective cyclo-oxygenase-2 inhibitors and traditional non-steroidal anti-inflammatory drugs increase the risk of atherothrombosis? Meta-analysis of randomised trials. BMJ 2006;332:1302–1308. [PMC free article] [PubMed]
25. Salpeter SR, Gregor P, Ormiston TM, et al. Meta-analysis: cardiovascular events associated with nonsteroidal anti-inflammatory drugs. Am J Med 2006;119:552–559. [PubMed]
26. Jüni P, Reichenbach S, Dieppe PA, et al. Discontinuation of Vioxx. Lancet 2005;365:26–27.
27. Huskisson EC, Woolf DL, Balme HW, et al. Four new anti-inflammatory drugs: responses and variations. BMJ 1976;i:1048–1049. [PMC free article] [PubMed]
28. Cooperating Clinics Committee of the American Rheumatism Association. A seven-day variability study of 499 patients with peripheral rheumatoid arthritis. Arthritis Rheum 1965;8:302–334. [PubMed]
29. Gøtzsche PC. Review of dose–response studies of NSAIDs in rheumatoid arthritis. Dan Med Bull 1989;36:395–399. Search date 1985; primary sources Medline, hand searches of reference lists, and contact with companies marketing proproctary preparations. [PubMed]
30. Eisenberg E, Berkey CS, Carr DB, et al. Efficacy and safety of nonsteroidal antiinflammatory drugs for cancer pain: a meta-analysis. J Clin Oncol 1994;12:2756–2765. Search date 1992; primary source Medline. [PubMed]
31. Cappelleri JC, Lau J, Kupelnick B, et al. Efficacy and safety of different aspirin dosages on vascular diseases in high-risk patients. A metaregression analysis. Online J Curr Clin Trials 1995; Doc. No. 174. Search date 1992; primary sources Medline and Current Contents. [PubMed]
32. Henry D, Lim LL, Garcia Rodriguez LA, et al. Variability in risk of gastrointestinal complications with individual non-steroidal anti-inflammatory drugs: results of a collaborative meta-analysis. BMJ 1996;312:1563–1566. Search date 1994; primary source Medline. [PMC free article] [PubMed]
33. Henry D, McGettigan P. Epidemiology overview of gastrointestinal and renal toxicity of NSAIDs. Int J Clin Pract 2003;135(suppl):43–49. Search date 2001; primary sources Medline, Embase, The Cochrane Library, Healthstar, Cinahl, IDIS, reference lists, and hand searches.
34. Rostom A, Dube C, Wells G, et al. Prevention of NSAID-induced gastroduodenal ulcers. In: The Cochrane Library, Issue 4, 2002. Oxford: Update Software. Search date 2004; primary sources Medline, Current Contents, Embase, Cochrane Controlled Trials Register, Biosis Previews, ADIS LMS Drug Alerts, Pharmaceutical News Index. New articles since the last search update were evaluated. Recent conference proceedings were reviewed and content experts and companies were contacted.
35. Hawkey CJ, Karrasch JA, Szczepanski L, et al. Omeprazole compared with misoprostol for ulcers associated with nonsteroidal antiinflammatory drugs. Omeprazole versus misoprostol for NSAID-induced ulcer management (OMNIUM) study group. N Engl J Med 1998;338:727–734. [PubMed]
36. Graham DY, Agrawal NM, Campbell DR, et al. Ulcer prevention in long-term users of nonsteroidal antiinflamatory drugs: results of a double-blind, randomized, multicentre, active- and placebo-controlled study of misoprostol and lansoprazole. Arch Intern Med 2002;162:169–175. [PubMed]
37. Chalmers TC, Berrier J, Hewitt P, et al. Meta-analysis of randomized controlled trials as a method of estimating rare complications of non-steroidal anti-inflammatory drug therapy. Aliment Pharmacol Ther 1988;2(suppl 1):9–26. Search date not reported; primary source Medline. [PubMed]
38. Yeomans ND, Tulassay Z, Juhasz L, et al. A comparison of omeprazole with ranitidine for ulcers associated with nonsteroidal antiinflammatory drugs. Acid suppression trial: ranitidine versus omeprazole for NSAID associated ulcer treatment (ASTRONAUT) study group. N Engl J Med 1998;338:719–726. [PubMed]
39. Mason L, Moore RA, Edwards JE, et al. Topical NSAIDs for acute pain: a meta-analysis. BMC Fam Pract 2004;5:10. [PMC free article] [PubMed]
40. Lin J, Zhang W, Jones A, Doherty M. Efficacy of topical non-steroidal anti-inflammatory drugs in the treatment of osteoarthritis: meta-analysis of randomised controlled trials. BMJ 2004;329:324–326. [PMC free article] [PubMed]
41. Bjordal JM, Klovning A, Ljunggren AE, et al. Short-term efficacy of pharmacotherapeutic interventions in osteoarthritic knee pain: A meta-analysis of randomised placebo-controlled trials. Eur J Pain 2007;11:125–138. [PubMed]
42. Prepublication history for: Mason L, Moore RA, Edwards JE, et al. Topical NSAIDs for acute pain: a meta-analysis. BMC Fam Pract 2004;5:10. Available online at: http://www.biomedcentral.com/1471-2296/5/10/prepub (last accessed 13 May 2010). [PMC free article] [PubMed]
2007; 2007: 1108.
Published online 2007 June 1.

Differences in efficacy between systemic NSAIDs

Summary

PAIN INTENSITY COX-2 inhibitors compared with older NSAIDs: COX-2 inhibitors are as effective at reducing pain scores as older NSAIDs ( high-quality evidence ). Other systemic NSAIDs versus each other: Different systemic NSAIDs seem as effective as each other for pain relief in people with acute and chronic musculoskeletal disorders (high-quality evidence). NSAIDs compared with paracetamol: NSAIDs may be no more effective than paracetamol at reducing pain in people with acute musculoskeletal disorders (high-quality evidence). ADVERSE EFFECTS COX-2 inhibitors compared with older NSAIDs: COX-2 inhibitors, or COX-2 selective NSAIDs, reduce symptomatic ulcers compared with older, non-selective NSAIDs (high-quality evidence). COX-2 inhibitors increase the risk of cardiovascular events compared with placebo or older NSAIDs (high-quality evidence).

Benefits

Oral indometacin (indomethacin) versus newer oral NSAIDs:

We found one systematic review (search date 1985, 37 crossover RCTs, 1416 people with rheumatoid arthritis), which compared indometacin versus 10 newer NSAIDs (naproxen in 8 cases) for a median of 2 weeks with each drug. Four of the RCTs included a placebo period, and one RCT compared four drugs. The review found that 5% more people (95% CI 0% to 10%) preferred the newer NSAID to indometacin.

Oral cyclo-oxygenase-2 (COX-2) inhibitors versus older oral NSAIDs:

We found two systematic reviews (search dates 2002 and 2000), which found that celecoxib and rofecoxib were no more effective for clinical outcomes than older NSAIDs in people with rheumatoid arthritis or osteoarthritis (see table 1 ). Rofecoxib has been withdrawn from the market because of adverse effects (see harms below).

Table 1
Cyclo-oxygenase-2 (COX-2) inhibitors versus older NSAIDs for the symptoms of musculoskeletal disorders.

Other comparisons of systemic NSAIDs versus each other:

We found five other systematic reviews comparing different systemic NSAIDs. The first of these systematic reviews (search date 1988, 88 RCTs, each comparing 2 NSAIDs, 6440 people with rheumatoid arthritis) found no significant differences in the number of tender joints improved between 17 different NSAIDs. The second and third reviews (search dates 1994 and 1996) found no clear differences between various NSAIDs used to treat osteoarthritis of the hip (39 RCTs) or the knee (16 RCTs; see NSAIDs in review on osteoarthritis of the knee). The fourth and fifth systematic reviews were in people with acute musculoskeletal syndromes and identified generally poor-quality RCTs. The fourth systematic review (search date 1993, 84 RCTs, 32,025 people with soft-tissue injuries of the ankle) was unable to meta-analyse data. The fifth review (search date 1998, 17 RCTs for shoulder pain) was inconclusive.

Oral NSAIDs versus oral paracetamol:

We found no systematic review but found two RCTs. The first RCT included 300 people with painful limb injuries. People were randomised to four arms: daily intake for 3 days of paracetamol 4 g, diclofenac 75 mg, indometacin 75 mg, or paracetamol 4 g plus diclofenac 75 mg. The trial was double blind for the comparisons of paracetamol versus indometacin, and diclofenac versus paracetamol plus diclofenac. An imbalance in numbers in the four groups was caused by simple randomisation being used, rather than randomisation in blocks (Rainer TH, personal communication). It is therefore not an indication that the process was biased. The RCT found no significant or clinically relevant difference in pain reduction between the groups (on a 100 mm visual analogue scale [VAS] [minimally relevant difference was 13 mm, based on other studies]; difference in pain reduction between indometacin and paracetamol: +5.7 mm, 95% CI –1.5 mm to +12.8 mm). In the second RCT (non-inferiority, 260 people with lateral ankle sprains), people were randomised to take either paracetamol 3.9 g or ibuprofen 1200 mg daily for 9 days. The RCT found no significant difference in pain reduction on walking between the groups after 4 and 9 days (on a 100 mm VAS [0 = no pain; 100 = very severe pain]; day 4 mean change from baseline: –37 with paracetamol v –35 with ibuprofen; P = 0.24; day 9 mean change from baseline: –56 with paracetamol v –57 with ibuprofen; P = 0.73).

Harms

Oral indometacin (indomethacin) versus newer oral NSAIDs:

The review that compared indometacin versus newer NSAIDs did not report on harms separately, but combined benefits and harms in one overall outcome (patient preference).

Oral COX-2 inhibitors versus older oral NSAIDs:

We found two systematic reviews of the benefits and harms in people with rheumatoid arthritis or osteoarthritis; for harms see table 1 . We found three systematic reviews specifically on the harms of COX-2 inhibitors. The first systematic review of harms (search date 2002) found that COX-2 inhibitors and older NSAIDs with substantial COX-2 inhibiting effects significantly reduced the occurrence of symptomatic ulcers compared with other older NSAIDs (COX-2 inhibitors v older NSAIDs: 17 RCTs, 25,564 people; RR 0.49, 95% CI 0.40 to 0.60; older NSAIDs with COX-2 inhibitory activity v other older NSAIDs: 51 RCTs, 28,178 people; RR 0.41, 95% CI 0.30 to 0.70).

The second systematic review (search date 2004, 18 RCTs, 25,273 people) examined the risk of cardiovascular events with rofecoxib in people with chronic musculoskeletal disorders. It found that rofecoxib significantly increased the risk of myocardial infarction (MI) compared with placebo or older NSAIDs (RR 2.24, 95% CI 1.24 to 4.02). There was no evidence that results would differ according to comparator (P for interaction = 0.41). Rofecoxib was withdrawn from the market in 2004 after a placebo-controlled RCT found that it increased the risk of cardiovascular adverse effects.

The third systematic review (search date 2005, 6 RCTs, 12,780 people) looked at the cardiovascular adverse effects of celecoxib; one of the RCTs was misleadingly reported as it was not one RCT, but contained the combined results of two RCTs. A meta-analysis of four of the RCTs (4422 people) found that celecoxib significantly increased the risk of MI compared with placebo (OR 2.26, 95% CI 1.00 to 5.10). When studies with diclofenac, ibuprofen, and paracetamol as comparators were included in the meta-analysis (total 6 RCTs,12,780 people), the review still found that celecoxib significantly increased the risk of MI (OR 1.88, 95% CI 1.15 to 3.08). The available evidence suggests that the increase in MI is a class effect of the COX-2 inhibitors.

Other comparisons of systemic NSAIDs versus each other:

We found one systematic review (search date 2005,13 RCTs, 7718 people) of cardiovascular events with non-selective NSAIDs. It found no significant difference in MI with non-selective NSAIDs compared with placebo (OR 1.3, 95% CI 0.8 to 2.1).

Oral NSAIDs versus oral paracetamol:

The RCT found that adverse effects occurred in <7% of cases and were not severe.

Comment

One of the RCTs identified by the review of the cardiovascular adverse effects of celecoxib, which compared celecoxib versus ibuprofen and diclofenac, has been criticised because the publication differs from the trial protocols in objectives, primary outcomes, statistical analysis, and trial duration.

Clinical guide:

Important differences in adverse effects exist between different NSAIDs. The reduction in ulcers with COX-2 inhibitors should be weighed against the increase in cardiovascular risk compared with the older NSAIDs. In contrast, the beneficial effects of NSAIDs seem similar. People's preferences for particular drugs have not been replicated and could therefore be due to chance or natural fluctuations in disease activity. The evidence suggests that, if the NSAID is unsatisfactory, then switching to another NSAID will not solve the problem. Likewise, doubling the dose of an NSAID leads to only a small increase in effect, which may not be clinically relevant, but results in increased adverse effects (see below). For acute musculoskeletal problems, it is doubtful whether NSAIDs have any clinically relevant anti-inflammatory effect; we found one large double-blind RCT comparing an NSAID versus paracetamol, which did not show any significant or clinically relevant differences, and another large RCT that did not find a significant difference in ankle swelling. Paracetamol has been studied in osteoarthritis, where it had less effect than NSAIDs (see simple analgesics v NSAIDs in review on osteoarthritis of the knee).

Substantive changes

No new evidence

2007; 2007: 1108.
Published online 2007 June 1.

Dose–response relationship of systemic NSAIDs

Summary

PAIN INTENSITY High-dose compared with lower-dose systemic NSAIDs: High-dose NSAIDs have a very low benefit in terms of extra reduction in pain scores compared with moderate-dose NSAIDs ( high-quality evidence ). Recommended doses are close to creating the maximum benefit even in the low-dose range. ADVERSE EFFECTS High-dose compared with lower-dose systemic NSAIDs: Adverse effects from NSAIDs increase in an approximately linear fashion with dose, with no ceiling (high-quality evidence).

Benefits

We found three systematic reviews. The first review (search date 1985, 19 RCTs in which people were randomised to more than 1 dose of 9 different NSAIDs) found a dose–response relationship that saturated at high doses. The differences in effect between the studied doses were very small or non-existent within the commonly used dose ranges (daily doses: phenylbutazone 50–300 mg, indometacin (indomethacin) 45–105 mg, ibuprofen 800–3200 mg, naproxen 250–1500 mg, piroxicam 10–30 mg, diclofenac 75–150 mg, diflunisal 500–1000 mg, and tolfenamic acid 300–600 mg); only for aspirin was there a difference (between 2.6 and 5.3 g) but this particular drug has never been recommended at the lower dose. This and the second systematic review (search date 1992, 1545 people) found that the recommended dosages were close to providing a ceiling effect. The third of these reviews (search date 1988, 115 RCTs) found no significant differences between various doses of drugs.

Harms

Four systematic reviews (search dates 1992, 1994, and 2001) found no ceiling effect for adverse effects, and found that the incidence of adverse effects increased in an approximately linear fashion with dose.

Comment

Clinical guide:

Doubling the dose of an NSAID leads to only a small increase in beneficial effect, which may not be clinically relevant, but it also doubles the risk of harms.

Substantive changes

No new evidence

2007; 2007: 1108.
Published online 2007 June 1.

Misoprostol in people who cannot avoid systemic NSAIDs

Summary

GASTROINTESTINAL COMPLICATIONS Misoprostol compared with placebo: Misoprostol is more effective at reducing serious gastrointestinal complications and symptomatic ulcers compared with placebo ( high-quality evidence ). Misoprostol compared with proton pump inhibitors: The effects of misoprostol are unclear compared with proton pump inhibitors ( moderate-quality evidence ). ADVERSE EFFECTS Misoprostol increases withdrawals caused by adverse events — mainly diarrhoea and abdominal pain — compared with placebo. NOTE We found no clinically important results from RCTs on the effects of misoprostol compared with H 2 blockers.

Benefits

Misoprostol versus placebo:

We found one systematic review (search date 2002, 23 RCTs) that reported on clinical outcomes. It found that misoprostol significantly reduced serious gastrointestinal complications and symptomatic ulcers compared with placebo (serious gastrointestinal complications: 10 RCTs, 11,507 people; RR 0.6, 95% CI 0.4 to 0.9; symptomatic ulcers: 2 RCTs, 8913 people; RR 0.4, 95% CI 0.2 to 0.7).

Misoprostol versus proton pump inhibitor:

We found one systematic review (search date 2004), which identified two fully published RCTs comparing misoprostol versus proton pump inhibitors. The first RCT in the review (935 people treated with NSAIDs who had ulcers or >10 erosions at endoscopy) compared misoprostol (200 micrograms 4 times daily) versus omeprazole (20 or 40 mg/day) once daily. Treatment success was defined as <five erosions at each site, no ulcers, and not more than mild dyspepsia. It found no significant difference in treatment success between misoprostol and omeprazole at 8 weeks (71% with misoprostol 800 micrograms v 76% with omeprazole 20 mg v 75% with omeprazole 40 mg; P >0.2 for comparison of misoprostol v either dose of omeprazole). People having successful treatment (732 people) were re-randomised to receive maintenance treatment with misoprostol (200 micrograms twice daily), omeprazole (20 mg once daily), or placebo. Remission was defined as no ulcers, <11 gastric or duodenal erosions, only mild dyspepsia, and no adverse events requiring treatment discontinuation. Omeprazole significantly increased the proportion of people still in remission at 6 months compared with misoprostol (estimated AR: 48% with misoprostol v 61% with omeprazole; P = 0.001). The second RCT (537 people, using NSAIDs for at least 1 month and with a history of endoscopically confirmed gastric ulcer) compared four treatments: misoprostol (200 micrograms 4 times daily), two different doses of lansoprazole (15 and 30 mg) once daily, and placebo. It found that misoprostol significantly increased the length of time to recurrence compared with the other treatments (time to gastric ulcer displayed graphically; misoprostol v placebo, P <0.001; misoprostol v lansoprazole 15 mg, P = 0.01; misoprostol v lansoprazole 30 mg, P = 0.04; AR for being free of gastric ulcer at 12 weeks: 93% with misoprostol v 80% with lansoprazole 15 mg v 82% with lansoprazole 30 mg v 51% with placebo).

Misoprostol versus H2 blockers:

We found no systematic review of sufficient quality.

Harms

Misoprostol versus placebo:

The systematic review reporting on beneficial clinical outcomes did not report on the harms of misoprostol. A second systematic review (search date 2004) found that a significantly larger proportion of people receiving misoprostol than placebo withdrew because of adverse events, primarily diarrhoea, abdominal pain, and nausea (12 RCTs, 12,146 people; RR 1.41, 95% CI 1.31 to 1.51). One large RCT (8843 people) identified by the review found no significant difference in the proportion of deaths between groups (17/4404 [0.4%] with misoprostol v 21/4439 [0.5%] with placebo; ARR 0.10%; RR 0.82, 95% CI 0.43 to 1.55). One person taking placebo died as a direct result of gastrointestinal toxicity.

Misoprostol versus proton pump inhibitors:

The first review did not report on the harms of misoprostol or proton pump inhibitors. The second review found a significantly larger proportion of withdrawals with misoprostol than with proton pump inhibitors (RR 0.71, 95% CI 0.52 to 0.97).

Misoprostol versus H2 blockers:

We found no systematic review of sufficient quality.

Comment

The clinical relevance of effects of co-treatments on endoscopically detected gastrointestinal ulceration is doubtful. The rate of ulcers was more than 10 times higher in the studies in which the investigators looked for ulcers with regular endoscopy than in earlier RCTs of NSAIDs, which looked for symptomatic ulcers. The definition of ulcers varied between studies; sometimes they were defined as endoscopic lesions with a size of only 3 mm, sometimes as any lesion of an unequivocal depth, and sometimes no definition was provided at all.

Clinical guide:

Only misoprostol has been shown to reduce serious gastrointestinal complications and symptomatic ulcers compared with placebo, but this drug also causes people to withdraw from treatment because of gastrointestinal adverse effects.

Substantive changes

No new evidence

2007; 2007: 1108.
Published online 2007 June 1.

H2 blockers in people who cannot avoid systemic NSAIDs

Summary

GASTROINTESTINAL COMPLICATIONS H2 blockers compared with placebo: The effects of H 2 blockers compared with placebo are unclear ( moderate-quality evidence ). H2 blockers compared with proton pump inhibitors: H 2 blockers are less effective at reducing gastrointestinal complications of NSAIDs (moderate-quality evidence). NOTE We found no clinically important results from RCTs about the effects of H 2 blockers compared with misoprostol.

Benefits

H2 blockers versus placebo:

We found one systematic review (search date 2002, 15 RCTs, 2621 people) that reported on clinical outcomes. It found insufficient evidence to compare the effects of H2 blockers versus placebo on gastrointestinal complications or symptomatic ulcers reliably. However, it found that H2 blockers significantly reduced endoscopically diagnosed ulcers compared with placebo (RR 0.6, 95% CI 0.4 to 0.7).

Proton pump inhibitors versus H2 blockers:

We found one systematic review (search date 2004), which identified one RCT (541 people with NSAID-related ulcers, or >10 gastric or duodenal erosions). The RCT compared three treatments: omeprazole 20 mg, omeprazole 40 mg, and ranitidine 300 mg daily. Treatment success was defined as no ulcers, <five erosions at each site, and only mild dyspepsia. It found that both doses of omeprazole significantly increased treatment success compared with ranitidine (80% with omeprazole 20 mg v 79% with omeprazole 40 mg v 63% with ranitidine; P ≤0.001 for comparison of ranitidine v either dose of omeprazole). People having successful treatment were re-randomised to either omeprazole 20 mg daily or ranitidine 300 mg daily. Relapse was defined as an ulcer, 10 or more gastric or duodenal erosions, moderate or severe dyspepsia, or treatment discontinuation caused by adverse events. Omeprazole significantly increased time to relapse compared with ranitidine (results displayed graphically; P = 0.004, log-rank test).

Misoprostol versus H2 blockers:

We found no systematic review of sufficient quality.

Harms

H2 blockers versus placebo:

The review did not report on the harms of H2 blockers. Another systematic review (search date 2004) found no significant difference between low- or high-dose H2 blockers and placebo in withdrawals caused by adverse effects (high dose: 1 RCT, 78 people; RR 1.00, 95% CI 0.21 to 4.65; low dose: 3 RCTs, 937 people; RR 0.86, 95% CI 0.58 to 1.28).

Proton pump inhibitors versus H2 blockers:

Few adverse events were reported in the RCT comparing omeprazole versus ranitidine. Treatment discontinuations (all causes) occurred in 10% of people taking omeprazole 20 mg, 10% taking omeprazole 40 mg, and 14% taking ranitidine (significance not reported).

Misoprostol versus H2 blockers:

We found no systematic review of sufficient quality.

Comment

See comment on misoprostol in people who cannot avoid NSAIDs.

Substantive changes

No new evidence

2007; 2007: 1108.
Published online 2007 June 1.

Proton pump inhibitors in people who cannot avoid systemic NSAIDs

Summary

GASTROINTESTINAL COMPLICATIONS Proton pump inhibitors compared with placebo: The effects of proton pump inhibitors compared with placebo are unclear ( moderate-quality evidence ). Compared with H2 blockers: Proton pump inhibitors are more effective at reducing gastrointestinal complications of NSAIDs (moderate-quality evidence). Proton pump inhibitors compared with misoprostol: The effects of misoprostol are unclear compared with proton pump inhibitors (moderate-quality evidence).

Benefits

Proton pump inhibitors (PPIs) versus placebo:

We found two systematic reviews, one reporting on symptomatic ulcers and serious gastrointestinal complications, and another examining endoscopic ulcers. The first review (search date 2002, 6 RCTs, 1358 people) found too few serious gastrointestinal complications to compare the effects of PPIs versus placebo reliably. Although initial analyses suggested a significant reduction in symptomatic ulcers with PPIs compared with placebo (2 RCTs, 343 people; RR 0.09, 95% CI 0.02 to 0.47), this significance was lost on sensitivity analysis, when the poor-quality study was removed from the analysis. The second review (search date 2004) identified five RCTs (1216 people) comparing PPIs versus placebo in people who had received NSAIDs for at least 3 months. It found that PPIs reduced endoscopically diagnosed ulcers compared with placebo (RR 0.35, 95% CI 0.29 to 0.44).

PPIs versus H2 blockers:

See benefits of H2 blockers in people who cannot avoid NSAIDs.

Misoprostol versus PPIs:

See benefits of misoprostol in people who cannot avoid NSAIDs.

Harms

Proton pump inhibitors (PPIs) versus placebo:

The review found no significant difference between PPIs and placebo in withdrawals caused by adverse effects (4 RCTs,1113 people; RR 1.20, 95% CI 0.66 to 2.15).

PPIs versus H2 blockers:

See harms of H2 blockers in people who cannot avoid NSAIDs.

Misoprostol versus PPIs:

See harms of misoprostol in people who cannot avoid NSAIDs.

Drug safety alert

FDA issues drug safety alert on the possible increased risk of fractures of the hip, wrist, and spine, associated with the use of proton pump inhibitors. (25 May 2010)

A drug safety alert has been issued on the possible increased risk of fractures of the hip, wrist, and spine, associated with the use of proton pump inhibitors. (www.fda.gov)

Comment

See comment on misoprostol in people who cannot avoid NSAIDs.

Substantive changes

No new evidence

2007; 2007: 1108.
Published online 2007 June 1.

NSAIDs (topical)

Summary

PAIN INTENSITY Topical NSAIDs compared with placebo: Topical NSAIDs may be more effective at reducing pain at 1 week in people with acute musculoskeletal pain, but the effect is small, at best ( very low-quality evidence ). In people with chronic pain, topical NSAIDs may be no more effective than placebo ( very low-quality evidence ).

Benefits

We found three systematic reviews: one in people with acute musculoskeletal pain conditions and two in people with osteoarthritis

Topical NSAIDs versus placebo:

The review in people with acute musculoskeletal pain (search date 2003, 36 RCTs) found that topical NSAIDs significantly improved "clinical success" rate compared with placebo at 1 week (26 RCTs, 2853 people; "clinical success" defined as "good" or "excellent" global assessment of treatment, or "none" or "slight" pain on rest or movement; AR for "clinical success": 65% with topical NSAIDs v 39% with placebo; RR 1.6, 95% CI 1.4 to 1.7; NNT 3.8, 95% CI 3.4 to 4.4; see comment below). There was extreme statistical heterogeneity in this analysis (P <0.00001) (i.e., widely diverging results between trials), which means that these results should be interpreted with great caution as they may not be reliable.

The first review in people with osteoarthritis (search date 2003) found that, although topical NSAIDs reduced pain compared with placebo at 2 weeks, this effect was not significant at 4 weeks (2 weeks: 6 RCTs, 893 people: effect size 0.40, 95% CI 0.15 to 0.65; P ≤0.05; 4 weeks: 3 RCTs, 558 people; effect size +0.04, 95% CI –0.11 to +0.19; P >0.05; see comment below). The analysis at 2 weeks was statistically heterogeneous (P <0.001).

The second review in people with osteoarthritis (search date 2005, 8 RCTs, 6 included in the first review, 749 people receiving topical NSAIDs, number of people receiving placebo not reported) found that NSAIDs significantly improved pain compared with placebo at a mean 1.6 weeks (mean improvement of 12 mm, 95% CI 7 mm to 16 mm on a 100 mm visual analogue pain scale for NSAIDs v placebo; absolute numbers not reported; random effects model). However, although this difference between groups was statistically significant, it was below what was deemed by the reviewers to reflect the average minimal clinically important improvement, namely 20 mm on a 100 mm visual analogue pain scale. This review also found heterogeneous results (P = 0.002 for test of heterogeneity) owing to the use of different topical NSAID gels over different time frames.

The reviews found only a single RCT comparing topical NSAIDs versus placebo for >4 weeks.

Harms

Topical NSAIDs versus placebo:

The review in people with acute musculoskeletal pain found no significant difference between topical NSAIDs and placebo in withdrawals caused by adverse effects (0.8% with topical NSAIDs v 0.7% with placebo; RR 1.6, 95% CI 0.8 to 3.4).

The review in people with osteoarthritis found no significant difference between topical NSAIDs and placebo in adverse events (crude rates: 108/577 [19%] with topical NSAIDs v 85/531 [16%] with placebo; RR 1.02, 95% CI 0.62 to 1.68).

Comment

Both reviews found that the effect decreased significantly and substantially with increasing sample size of the trials, rendering interpretation of the results difficult. The trials identified by the review of NSAIDs for osteoarthritis were of poor quality, and this might also have been the case for the trials included in the review of acute pain conditions (see published peer review comments on this review).

Substantive changes

Topical NSAIDs versus placebo: One systematic review added comparing topical NSAIDs versus placebo in people with osteoarthritis confirms previous conclusions. While topical NSAIDs reduce pain compared with placebo at 1 week, the pain reduction is likely to be clinically unimportant to patients. Categorisation unchanged (Unknown effectiveness).

2007; 2007: 1108.
Published online 2007 June 1.

Topical versus systemic NSAIDs or alternative analgesics

Summary

PAIN INTENSITY Topical NSAIDs compared with oral NSAIDs: There may be no difference in pain reduction with topical NSAIDs compared with oral NSAIDs in people with acute and chronic pain ( very low-quality evidence ). NOTE We found no clinically important results from RCTs about the effects of topical NSAIDs compared with paracetamol in people with musculoskeletal conditions.

Benefits

We found two systematic reviews: one in people with acute musculoskeletal pain conditions (search date 2003, 36 RCTs) and one in people with osteoarthritis (search date 2003, 13 RCTs).

Topical NSAIDs versus oral NSAIDs:

The review in people with acute musculoskeletal pain conditions found three RCTs (433 people), two of which compared the same NSAID delivered topically and orally. The review found no significant difference between topical and oral NSAIDs in "clinical success" rate (AR: 57% with topical NSAIDs v 62% with oral NSAIDs; RR 0.9, 95% CI 0.8 to 1.1). The review did not specify length of follow-up in the RCTs identified.

The review in people with osteoarthritis found two RCTs (529 people) that compared different NSAIDs given topically and orally. The review found that topical NSAIDs were significantly less effective in reducing pain than oral NSAIDs at 1 week but not at 3 weeks (1 week: 1 RCT, 208 people; effect size –0.38, 95% CI –0.66 to –0.10; 3 weeks: 2 RCTs, 529 people; effect size –0.26, 95% CI –0.68 to +0.16).

Topical NSAIDs versus paracetamol:

We found no systematic review or RCTs.

Harms

Topical NSAIDs versus oral NSAIDs:

We found no reliable RCTs comparing adverse effects of the same NSAID in topical and oral formulations. The review in people with acute pain conditions did not report on harms with topical NSAIDs compared with oral NSAIDs. The review in people with osteoarthritis found that topical NSAIDs increased local adverse reactions compared with oral NSAIDs (crude rates: 18/243 [7%] with topical NSAIDs v 2/200 [1%] with oral NSAIDs; RR 5.3, 95% CI 1.1 to 24.5).

Topical NSAIDs versus paracetamol:

We found no RCTs.

Comment

Both reviews found that the effect decreased significantly and substantially with increasing sample size of the trials, rendering interpretation of the results difficult. The trials identified by the review of NSAIDs for osteoarthritis were of poor quality, and this might also have been the case for the trials included in the review of acute pain conditions (see published peer review comments on this review).

Substantive changes

No new evidence


Articles from BMJ Clinical Evidence are provided here courtesy of BMJ Publishing Group