Selective COX-2 inhibitors reduced pain interference after surgery with no detected safety signal

Design

Systematic review and meta-analysis of randomised trials

Population

38 RCTs, 5,424 adults having surgery

Primary outcome

Pain-related interference with daily function within one month

Effect

BPI −1.1 (−1.4 to −0.8); chronic pain OR 0.44 (0.21–0.93)

This meta-analysis, published in June 2026, pooled 38 randomised trials (5,424 adults) comparing perioperative selective COX-2 inhibitors with placebo, opioids or usual care. Its primary outcome was pain-related interference with daily living — how much pain stops patients doing things — measured with tools such as the Brief Pain Inventory.

COX-2 inhibitors reduced acute pain interference by 1.1 points (95% CI 0.8–1.4; moderate certainty), at the minimal clinically important difference. Chronic postsurgical pain incidence was lower (OR 0.44, 0.21–0.93; low certainty), quality of recovery improved modestly and blood loss was slightly lower. No differences were detected in acute kidney injury, gastrointestinal bleeding, impaired bone healing, myocardial infarction, stroke or death.

The case for including a selective COX-2 inhibitor in multimodal analgesia, unless contraindicated, now rests on function as well as pain scores. Trials typically exclude high-risk patients, so the safety data apply to the selected, and renal, cardiovascular and bleeding risks still need individual judgement.

• Consider a selective COX-2 inhibitor as part of multimodal analgesia unless contraindicated
• Screen for renal impairment, active cardiovascular disease and bleeding risk before prescribing
• Use the shortest effective course after surgery
• Ask patients about pain interfering with walking, sleep and self-care, not only pain scores

Why it matters

Moves the argument for COX-2 inhibitors from pain scores to whether patients can function.

Don't overread it

Trials excluded high-risk patients; safety in renal or cardiovascular disease is not established here.

The statistics, in plain English

A 1.1-point reduction on a 0–10 interference scale sits right at the 1-point threshold patients notice. The chronic pain result rests on five trials and low-certainty evidence, so it is suggestive rather than settled. Absence of a detected safety difference is not proof of safety, because trials were short and excluded high-risk patients.

Read the rest in the app

You have read your two free briefings this month. The app carries all 27 specialties, every morning, free — and this finding is waiting in it.

Get Daily Doctor — free

Scan to keep reading on your phone. No account needed to start.