Generated August 2026 · Clinical Deep Dive
Beyond COX Selectivity: What PRECISION and a Decade of Follow-Up Data Mean for Choosing an NSAID
0 Comments
NSAIDs rank among the most heavily used drugs in medicine, available by prescription and over the counter, yet every one of them, from the corner-drugstore ibuprofen tablet to the injectable ketorolac given in the emergency department, now carries a single boxed warning spanning two organ systems at once: cardiovascular thrombotic events and gastrointestinal bleeding. DailyMed ibuprofen
For a decade, most clinicians managed that risk with a shortcut: reach for naproxen, widely believed to be the cardiovascular-safe NSAID, and reserve cyclooxygenase-2 (COX-2)-selective celecoxib for patients who could not tolerate a nonselective agent’s gastrointestinal (GI) burden. The Prospective Randomized Evaluation of Celecoxib Integrated Safety versus Ibuprofen or Naproxen (PRECISION) trial dismantled that shortcut. In 24,081 patients with osteoarthritis (OA) or rheumatoid arthritis (RA) and elevated cardiovascular risk, moderate-dose celecoxib was noninferior to both ibuprofen and naproxen for the composite of cardiovascular death, nonfatal myocardial infarction, and nonfatal stroke, and was associated with fewer GI and renal events than either comparator. N Engl J Med
“These findings challenge the widely held view that naproxen provides superior cardiovascular safety,” principal investigator Steven E. Nissen, MD (Cleveland Clinic), said when the results were presented. Medscape
PRECISION’s conclusions come with limitations clinicians should weigh alongside its size. Only celecoxib was capped by design at a comparatively moderate dose (100 mg twice daily, with dose escalation to 200 mg twice daily permitted only for rheumatoid arthritis and taken up by just 5.8% of celecoxib patients overall), whereas ibuprofen and naproxen patients could and did up-titrate freely, with roughly 55% of the osteoarthritis subgroup escalating to the highest labeled dose of each. DailyMed Celebrex The trial’s own published report confirms a 68.8% treatment discontinuation rate and 27.4% loss to follow-up, attrition a formal discussant at the time called concerning enough to strain confidence in the dataset. N Engl J Med +1 None of this overturns PRECISION’s core finding, but it means the trial compared a fixed, moderate celecoxib dose against nonselective agents that were frequently pushed to their labeled ceiling, and did so with attrition rates that leave a meaningful share of outcomes unaccounted for.
PRECISION reframed the question clinicians should be asking. Selectivity for COX-2 over COX-1 is a useful organizing concept, but it does not cleanly predict an individual agent’s cardiovascular, GI, or renal risk. Diclofenac is the clearest illustration: a nationwide Danish cohort study found a 50% increased risk of major adverse cardiovascular events in the 30 days after starting diclofenac compared with not starting an NSAID, a risk comparable to or exceeding that seen with COX-2-selective agents despite diclofenac being nonselective. BMJ Choosing among the seven agents examined in depth below (ibuprofen, naproxen, diclofenac, celecoxib, meloxicam, ketorolac, and indomethacin) requires weighing each patient’s cardiovascular, renal, and GI risk factors against agent-specific data rather than defaulting to a class reputation.
FIGURE. COX-1/COX-2 Selectivity Spectrum and Relative Risk Signal, 7 Featured Agents
| Agent | COX-1/COX-2 Selectivity Position | Cardiovascular Signal | GI / Renal Signal | Selectivity Holds Up To… |
| Ketorolac | Potent nonselective (COX-1-leaning) | Not established as elevated vs class at labeled short-term dosing DailyMed ketorolac tromethamine | Highest GI bleed risk in class; boxed 5-day limit DailyMed ketorolac tromethamine | Not applicable; nonselective across its labeled dose range |
| Indomethacin | Nonselective | Class-wide boxed warning; no distinguishing outcome trial DailyMed Indocin | Highest CNS/GI tolerability burden historically reported vs naproxen N Engl J Med | Not applicable; nonselective across its labeled dose range |
| Ibuprofen | Nonselective, mid-range | Noninferior to celecoxib in PRECISION N Engl J Med; highest stroke signal among NSAIDs in a separate network meta-analysis BMJ | More GI and renal events than celecoxib in PRECISION N Engl J Med | Not applicable; nonselective across its labeled dose range |
| Naproxen | Nonselective, mid-range | Noninferior to celecoxib in PRECISION, overturning the “safest NSAID” assumption N Engl J Med | More GI and renal events than celecoxib in PRECISION N Engl J Med | Not applicable; nonselective across its labeled dose range |
| Meloxicam | Preferentially COX-2 at low dose (degree, not absolute) DailyMed Mobic | Class-wide boxed warning; not directly compared in PRECISION DailyMed Mobic | Labeled GI risk similar to class; no dedicated outcome trial DailyMed Mobic | Preferential COX-2 inhibition is lost at the 15 mg/day ceiling, where meaningful COX-1 inhibition emerges; the 7.5 mg/day dose is more clearly COX-2-preferential Emerging Evidence in NSAID Pharmacology |
| Diclofenac | Preferentially COX-2 (degree, not absolute) DailyMed Voltaren | Cardiovascular signal comparable to or exceeding COX-2-selective agents in several large analyses BMJ | Class-wide GI boxed warning across all branded forms DailyMed Voltaren | At standard therapeutic dosing (50 mg 3 times daily) diclofenac already shows meaningful concurrent COX-1 inhibition alongside strong COX-2 inhibition, unlike celecoxib, which spares COX-1 across its labeled range Emerging Evidence in NSAID Pharmacology |
| Celecoxib | COX-2 selective | Noninferior to naproxen and ibuprofen in PRECISION N Engl J Med | Fewer GI and renal events than naproxen or ibuprofen in PRECISION N Engl J Med | Holds across the full labeled dose range (up to 400 mg/day) DailyMed Celebrex |
Note: “selectivity” here reflects in vitro/ex vivo COX-1:COX-2 inhibition ratios, not a guarantee of clinical outcome; diclofenac’s biochemical COX-2 preference is real but does not, on its own, predict its comparatively unfavorable cardiovascular signal. Emerging Evidence in NSAID Pharmacology
DEEP DIVE:
Seven Agents, One Class: Rebuilding the Case for Deliberate NSAID Selection
Medscape Medical News © 2026 WebMD, LLC
Send comments and news tips to news@medscape.net.
Cite this: Beyond COX Selectivity: What PRECISION and a Decade of Follow-Up Data Mean for Choosing an NSAID – Medscape– August 31, 2026.
