• We are sorry, but NCBI web applications do not support your browser and may not function properly. More information

Table 6Strength of Evidence by Key Question

Key QuestionStrength of evidenceConclusion
1. Are there differences in effectiveness between NSAIDs, with or without antiulcer medication, when used in adults with chronic pain from osteoarthritis, rheumatoid arthritis, soft-tissue pain, back pain, or ankylosing spondylitis?
 1a. How do oral drugs compare to one another?
  CelecoxibHigh. Evidence is available from many published trials.No clear differences in pain reduction.
  MeloxicamHigh. Consistent evidence from many published trialsNo consistent differences.
  NabumetoneModerate. Fewer RCTs/systematic reviewNo consistent differences.
  EtodolacHigh. Consistent evidence from many published trialsNo consistent differences.
  NonselectivesHigh. Consistent evidence from many published trials and several good-quality systematic reviewsNo consistent differences.
  SalsalateModerate. Limited evidence from few RCTsNo consistent differences.
  TenoxicamHigh. Many published RCTs, meta-analysisNo consistent differences.
  Tiaprofenic acidHigh. Several RCTs and 1 fair-quality reviewNo consistent differences.
 1b. How do topical drugs compare to one another?
  Diclofenac 1.5% topical solution and 1.0% topical gelLow. Indirect evidence from placebo-controlled trials.Both topical drugs had significantly greater mean changes in pain subscale scores than placebo.
  Other topical drugsInsufficientNo trials met inclusion criteria.
 1c. How do oral drugs compare to topical drugs?
  Diclofenac 1.5% topical solutionHigh. 2 head-to-head trialsCompared with oral diclofenac, diclofenac 1.5% topical solution produced similar improvement in WOMAC pain and physical function variables.
2 and 3. Are there clinically important differences in short-term (< 6 months) or long-term (≥ 6 months) harms between NSAIDs, with or without antiulcer medication, when used in adults with chronic pain from osteoarthritis, rheumatoid arthritis, soft-tissue pain, back pain, or ankylosing spondylitis?
 2a and 3a. How do oral drugs compare to one another?
  CelecoxibHigh. Evidence from many published trials and systematic reviewsGI Harms: Lower risk for celecoxib than nonselective NSAIDs in the short-term, but longer-term evidence is inconclusive.

CV Harms: No significant difference in risk of MI for celecoxib compared with nonselective NSAIDs, but evidence is primarily from short-term studies.

Other serious adverse events: No consistent differences.
  MeloxicamModerate for GI harms; low for othersShort-term and long-term GI harms: No consistent differences.

Long-term CV harms: No conclusive evidence of increased risk relative to nonselectives.

Hepatotoxicity: No evidence of increased risk relative to placebo.

Other serious adverse events: No evidence.
  NabumetoneModerate for short-term GI safety; low for othersShort-term GI harms: Decreased risk relative to nonselectives.

Other serious adverse events: No evidence
  EtodolacLow for perforation, symptomatic ulcer, or bleeding, insufficient for othersPerforation, symptomatic ulcer, or bleeding rates (duration unknown): No increased risk relative to nonuse.

Other serious adverse events: No evidence.
  NonselectivesHigh for GI safety; moderate for CV safety; low for other serious adverse eventsShort-term/long-term GI safety: All nonselectives are associated with similar increased risks relative to nonuse.

Short-term/long-term CV safety: Nonselective NSAIDs other than naproxen are associated with increased risks of CV events similar to that seen with COX-2 inhibitors (most data on high-dose ibuprofen and diclofenac). Naproxen appears to be risk-neutral with regard to cardiovascular events.

Hepatotoxicity: In short-term trials, diclofenac associated with highest rates of aminotransferase elevations >3 times upper limits of normal. Noncomparative evidence suggests similar rates in the longer term.

Fracture risk: Preliminary evidence from 1 case-control study suggestive of higher risk with ibuprofen compared with other nonselective NSAIDs.

All-cause mortality/blood pressure/ CHF/edema/renal function/hepatotoxicity: No consistent difference.
  Nonselective+antiulcer medicationsLow for GI events; moderate for endoscopic ulcersClinical GI events: Misoprostol only antiulcer medication proven to reduce rates, but at expense of reduced GI tolerability.

Endoscopic ulcers: All proven to reduce rates.
  SalsalateLow for short-term overall toxicity and long-term GI harms, insufficient for othersShort-term overall toxicity: Significantly lower rates.

Long-term GI harms: No differences.

Other serious adverse events: No evidence.
  TenoxicamInsufficientNo evidence found for specific GI and CV adverse events; reporting of AEs and dropouts slightly lower with tenoxicam compared with indomethacin and piroxicam respectively.
  Tiaprofenic acidModerate for cystitis, insufficient for othersObservational studies report serious cases of cystitis.
 2b and 3b. How do topical drugs compare to one another?
  Diclofenac 1.5% topical solution and 1.0% topical gelLow. Indirect evidence from placebo-controlled trials.Withdrawals due to adverse events: Significantly greater for diclofenac 1.5% topical solution, but not for 1.0% topical gel.

Short-term GI harms: Compared with placebo, neither topical product resulted in significant increased incidence.

Application site reactions: Only diclofenac 1.5% topical solution resulted in significantly greater skin dryness.
 2c and 3c. How do oral drugs compare to topical drugs?
  Diclofenac 1.5% topical solutionHigh. 2 head-to-head trialsTopical diclofenac resulted in significantly lower incidence of GI adverse events, but higher incidence of application site skin dryness. Withdrawals due to adverse events were similar for oral and topical diclofenac.
4. Are there subgroups of patients based on demographics, other medications (e.g., aspirin), socio-economic conditions, co-morbidities (e.g., gastrointestinal disease) for which one medication is more effective or associated with fewer harms?
 4a. How do oral drugs compare to one another?
  AllModerate for concomitant use of low-dose aspirin and for NSAID use in high-risk patients with recent GI bleed. Low for others.Demographics: No differences in efficacy, but risk of certain serious harms may be lower for celecoxib than some NSAIDs in elderly patients.

History of ulcer bleeding: Recurrent ulcer bleeding significant lower for celecoxib plus esomeprazole compared with celecoxib alone. No significant difference for celecoxib alone compared with a nonselective NSAID plus a PPI.

Cardiac/renal comorbidities: Celecoxib possibly associated with decreased risk of death and recurrent heart failure compared with nonselective NSAIDs in elderly patients with a recent admission for heart failure.

Concomitant use of anticoagulants: Comparative evidence from observational studies was inconclusive. Noncomparative evidence suggested no significant increase in INR after 5 weeks of celecoxib.

Concomitant use of low-dose aspirin: Similar rates of endoscopic ulcers for celecoxib compared with naproxen plus lansoprazole in prospective RCT. Subgroup analyses also found similar endoscopic ulcer rates for celecoxib and nonselective NSAIDs.
 4b. How do topical drugs compare to one another?
  AllInsufficientNo evidence
 4c. How do oral drugs compare to topical drugs?
  AllInsufficientNo evidence

Abbreviations: AE, adverse event; COX, cyclo-oxygenase; CV, cardiovascular; GI, gastrointestinal; INR, international normalized ratio; MI, myocardial infarction; NSAID, nonsteroidal antiinflammatory drug; OARSI, Osteoarthritis Research Society International; PPI, proton pump inhibitor; RCT, randomized controlled trial; WOMAC, Western Ontario and McMaster Universities Osteoarthritis Index.

From: Summary

Cover of Drug Class Review: Nonsteroidal Antiinflammatory Drugs (NSAIDs)
Drug Class Review: Nonsteroidal Antiinflammatory Drugs (NSAIDs): Final Update 4 Report [Internet].
Peterson K, McDonagh M, Thakurta S, et al.
Portland (OR): Oregon Health & Science University; 2010 Nov.
Copyright © 2010, Oregon Health & Science University.

PubMed Health. A service of the National Library of Medicine, National Institutes of Health.