NSAIDs and the Risk of Fracture Nonunion
Undertreated fracture pain pushes patients toward opioids, and avoiding NSAIDs without cause removes one of the safest tools for that pain.
Key terminology
- Multimodal pain control
- This pain plan combines drugs and methods that work by different routes, such as NSAIDs, acetaminophen, ice and nerve blocks. The aim is less opioid for the same pain control.
- COX-2
- This cyclooxygenase enzyme makes prostaglandins at sites of injury and inflammation. It is the target that links NSAIDs to slower bone repair in animals.
- Callus
- This soft bone and cartilage forms around a healing fracture, mostly from the periosteum. It shows on follow-up films as fluffy new bone bridging the gap.
- Nonunion
- This is a fracture that has stopped healing and will not unite without intervention. Films show a persistent gap with no bridging callus over months.
- Reverse causation
- This bias occurs when the outcome causes the exposure instead of the reverse. A painful nonhealing fracture leads patients to take more analgesics.
- Heterotopic ossification
- This is mature bone that forms in muscle and soft tissue where bone does not belong. It is common after acetabular surgery and can stiffen the hip.
- Coxibs
- These NSAIDs, such as celecoxib, block COX-2 and spare COX-1. They cause fewer stomach ulcers than ibuprofen or naproxen.
Takeaways
- NSAIDs belong in multimodal pain control for most fractures. The Orthopedic Trauma Association guideline recommends routine NSAID use in operative and nonoperative fracture care (strong recommendation, low-quality evidence). The panel judged the bone healing data too conflicting to decide either way and found no conclusive clinical evidence to prohibit NSAIDs (Hsu et al, J Orthop Trauma 2019;33:e158-e182, PMID 30681429).
- COX-2 prostaglandins drive the early inflammation that starts fracture healing. Prostaglandin E2 pushes mesenchymal cells toward the osteoblast line. COX-2 knockout mice healed tibial fractures slowly with a high rate of fibrous nonunion, and adding PGE2 rescued bone formation in culture. Rats given celecoxib or rofecoxib also healed poorly (Zhang et al, J Clin Invest 2002;109:1405-1415, PMID 12045254; Simon et al, J Bone Miner Res 2002;17:963-976, PMID 12054171).
- Early human data linking NSAIDs to nonunion may reflect reverse causation. Patients with a painful nonhealing fracture take more analgesics. In 9,995 humeral shaft fractures, only NSAID use at 61-90 days was linked to nonunion, and opioids in the same window showed the same signal. The authors judged this more likely reflects pain from fractures that were not healing (Bhattacharyya et al, Arthritis Rheum 2005;53:364-367, PMID 15934108; Giannoudis et al, J Bone Joint Surg Br 2000;82:655-658, PMID 10963160).
- The NSAID harm signal shrinks as study quality rises. One meta-analysis pooled an OR of 3.0 for nonunion, and its higher-quality subset (all spine fusion) was not significant. Another found OR 2.07 overall with no effect from low-dose or short exposure. A 2024 pooling of adjusted estimates found no significant difference (OR 1.11) (Dodwell et al, Calcif Tissue Int 2010;87:193-202, PMID 20552333; Wheatley et al, J Am Acad Orthop Surg 2019;27:e330-e336, PMID 30260913; Chuang et al, Front Endocrinol 2024;15:1428240, PMID 39319252).
- Long courses of indomethacin have the clearest human evidence of harm. Acetabular fracture patients given indomethacin for heterotopic ossification prophylaxis had long-bone nonunion in 26% vs 7%. In a randomized trial, 6 weeks of indomethacin had acetabular nonunion in 62% vs 19% with placebo, and 1 week had 24% (small arms, 63% compliance) (Burd et al, J Bone Joint Surg Br 2003;85:700-705, PMID 12892193; Sagi et al, J Orthop Trauma 2014;28:377-383, PMID 24625922).
- Coxibs may raise nonunion risk more than nonselective NSAIDs. In 339,864 long-bone fracture episodes from insurance claims, COX-2 inhibitor fills had an adjusted OR of 1.84 for nonunion and nonselective NSAIDs had 1.07 (not significant). Opioid fills were also linked (OR 1.69), which may reflect more severe fractures (George et al, J Bone Joint Surg Am 2020;102:1230-1238, PMID 32675672).
- Ibuprofen is first-line fracture analgesia in children. In a randomized trial, oral ibuprofen controlled outpatient fracture pain as well as oral morphine with fewer adverse effects (30.9% vs 56.1%). In 808 children with higher-risk fractures, ibuprofen exposure did not raise bone healing complications (OR 0.8, 95% CI 0.4-1.8) (Poonai et al, CMAJ 2014;186:1358-1363, PMID 25349008; DePeter et al, J Emerg Med 2017;52:426-432, PMID 27751698).
- Short NSAID courses show no measurable harm to adult fracture healing. In a placebo-controlled trial of nonoperative Colles fractures, 3 or 7 days of ibuprofen did not change fracture position, bone density or function and lowered pain over the first 3 days. A 2026 umbrella review found no rise in nonunion with use of 14 days or less (Aliuskevicius et al, Orthopedics 2021;44:105-110, PMID 33373464; Komargodski et al, J Am Acad Orthop Surg 2026, PMID 42489592).