Table 1.

Observational studies of gastrointestinal adverse drug events from NSAIDs in older adults^32–36

Study	NSAIDs assessed	Design	Patient Population	Safety Outcome	Findings
Chang	Celecoxib nsNSAIDs: indomethacin, sulindac, diclofenac, ketorolac, piroxicam, flurbiprofen, mefenamic acid	Case-crossover study	Patients aged ≥20 years in 2006 in Taiwan (mean age 61.4 years)	Hospitalization for upper GI adverse events (peptic ulcer and bleeding, gastritis, and duodenitis)	Adjusted OR (relative to no NSAID use): Celecoxib use in those aged 65–79: 1.97 (1.53–2.54) Celecoxib use in those aged 80+: 1.63 (1.18–2.24) Oral nsNSAID use in those aged 65– 79: 3.42 (3.14–3.72) Oral nsNSAID use in those aged 80+: 4.35 (3.85–4.93)
Mamdani	nsNSAIDs (e.g., naproxen, ibuprofen, diclofenac) Diclofenac + misoprostol Rofecoxib Celecoxib	Observational cohort	Older adults (aged 66+) in Ontario, Canada who initiated therapy with one of the four NSAID categories; control group of older adults who did not use any NSAID	Hospitalization for upper GI bleed	Adjusted Risk Ratios (relative to control): nsNSAIDs 4.0 (2.3–6.9), diclofenac + misoprostol 3.0 (1.7–5.6), rofecoxib 1.9 (1.3–2.8), and celecoxib 1.0 (0.7– 1.6) Adjusted Risk Ratios (relative to celecoxib): nsNSAIDs 4.4 (2.3– 8.5), diclofenac + misoprostol 3.2 (1.6–6.5), rofecoxib 1.9 (1.2–2.8) Adjusted Risk Ratios (relative to rofecoxib): nsNSAIDs 1.9 (1.0– 3.5), diclofenac + misoprostol 1.4 (0.7–2.7)
Rahme (a)	tNSAIDs (not specified)	Observational cohort	Older adults (aged 65+) in Quebec, Canada who filled a prescription for APAP or a tNSAID between January 1998 and December 2004	Hospitalization for upper and lower GI events	Adjusted Hazard Ratios (relative to APAP ≤3g/day): APAP >3g/day 1.20 (1.03–1.40), tNSAIDs 1.63 (1.44– 1.85), and APAP + tNSAIDs 2.55 (1.98– 3.28) Adjusted Hazard Ratio (relative to tNSAIDs): APAP + tNSAIDs 1.55 (1.20–2.00)
Rahme (b)	nsNSAIDs: ibuprofen, diclofenac, naproxen Rofecoxib Celecoxib	Observational cohort	Older adults (aged 65+) in Quebec, Canada who filled a prescription for one of the analgesic medications specified between April 1999 and December 2002	Hospitalization for GI bleeding	Adjusted Hazard Ratios (relative to APAP alone): Naproxen 2.75 (2.05–3.69), celecoxib + ASA 1.85 (1.48–2.31), diclofenac + ASA 3.06 (2.16–4.35), naproxen + ASA 2.37 (1.40–3.99), APAP + ASA 1.56 (1.31–1.87)
				Combined outcome of AMI/GI bleeding	Adjusted Hazard Ratios (relative to APAP alone): Naproxen 1.59 (1.31–1.93), celecoxib + ASA 1.34 (1.19–1.52), diclofenac + ASA 1.69 (1.35–2.10), APAP + ASA 1.29 (1.17–1.42)
Solomon	Non-selective NSAIDs: diclofenac, etodolac, flurbiprofen, ketorolac, ibuprofen, indomethacin, meloxicam, naproxen, piroxicam, sulindac COX-2 selective NSAIDs: celecoxib, rofecoxib, valdecoxib	Observational cohort	Older Medicare beneficiaries from Pennsylvania and New Jersey who qualified for pharmaceutical assistance for low- income older adults and who initiated therapy with an nsNSAID, a COX-2 selective NSAID, or an opioid from January 1, 1999 through December 31, 2005	Upper or lower GI tract bleeding	Adjusted HR (COX- 2 selective NSAID vs. nsNSAIDs) 0.60 (0.35–1.00)

Abbreviations: AMI: acute myocardial infarction; APAP: acetaminophen; ASA: aspirin; COX: cyclooxygenase; GI gastrointestinal; HR: hazard ratio; NSAID: non-steroidal anti-inflammatory drug; nsNSAID: non-selective non-steroidal anti-inflammatory drug; OR: odds ratio; tNSAID: traditional non-steroidal anti-inflammatory drug