Effect of NSAIDs on Bone Healing Rates: A Meta-analysis

Benjamin M. Wheatley, MD; Kyle E. Nappo, MD; Daniel L. Christensen, MD; Ann M. Holman, MLS; Daniel I. Brooks, PhD; Benjamin K. Potter, MD


J Am Acad Orthop Surg. 2019;27(7):e330-e336. 

In This Article

Abstract and Introduction


Introduction: NSAIDs inhibit osteogenesis and may result in delayed union or nonunion. The purpose of this meta-analysis was to determine whether their use leads to delayed union or nonunion.

Methods: We systematically reviewed the literature reporting the effect of NSAIDs on bone healing. We included studies of pediatric and adult patients NSAID exposure and healing bone. The outcomes of interest were delayed union, nonunion, or pseudarthrosis with at least six months of follow-up. A maximum likelihood random-effects model was used to conduct meta-analysis and meta-regression.

Results: NSAID exposure increased delayed union or nonunion (odds ratio [OR], 2.07; confidence interval [CI], 1.19 to 3.61). No effect was noted in pediatrics (OR, 0.58; CI, 0.27 to 1.21) or low dose/short duration of exposure (OR, 1.68; CI, 0.63 to 4.46).

Conclusion: Analysis of the literature indicates a negative effect of NSAIDs on bone healing. In pediatric patients, NSAIDs did not have a significant effect. The effect may be dose or time dependent because low-dose/short-duration exposure did not affect union rates.


Most fractures heal without complications within a relatively predictable time frame. Despite this phenomenon, up to 100,000 fractures undergo nonunion each year in the United States alone.[1] Nonunion, delayed union, or symptomatic pseudarthrosis may result in additional surgical procedures, require the use of fusion adjuncts such as a bone stimulator, or lead to prolonged immobilization, pain, and increased physician visits, all of which are time consuming and expensive for patients and healthcare systems alike. Consequently, surgeons continually seek methods to improve healing rates and, unfortunately, may be forced to make decisions with inadequate evidence.

A number of risk factors have been identified as potentially affecting the rate of bone healing. These include nonmodifiable risk factors such as age, sex, fracture characteristics, fracture location, and preexisting patient comorbidities.[2–4] There are also many modifiable risk factors to include alcohol and tobacco consumption, nutritional status, and medications such as diphosphonates and NSAIDs.[2–6] The magnitude of the effect of NSAIDs on bone healing is controversial, but many surgeons avoid these medications because of concern that they will delay or inhibit healing.

The mechanism of action and the potential for NSAIDs to prevent or delay bone healing have been convincingly demonstrated both in vitro and in animal studies.[7–9] Prostaglandins are thought to play a critical role in bone metabolism and healing.[10] The concentration of Prostaglandin E2 (PGE2) may control osteoblast behavior through the relative expression of the receptor activator of nuclear factor kappa-B ligand and osteoprotegerin, which is regulated through the enzymes cyclooxygenase (COX)-1 and COX-2.[10,11] The inhibition of COX isozymes and the subsequent decrease in PGE2 are thought to be the mechanisms through which NSAIDs may delay bone healing.[11]

In 1976, Rø et al[12] reported the effect of indomethacin on a closed femoral fracture in a rat model. They concluded that indomethacin delays the fracture healing process and creates a qualitatively inferior callus, which may predispose to nonunion and pseudarthrosis formation. These findings brought the safety of the use of NSAIDs in orthopaedic patients into question, and similar findings have been found in numerous subsequent animal studies.[7–9]

The evidence in humans is less clear. A previous meta-analysis demonstrated an increased risk of nonunion in the pooled effect of all studies; however, this effect was lost when considering only high-quality studies.[13] Because of the continued lack of consensus, many subsequent studies have added to the body of available evidence regarding NSAID exposure and bone healing.

The purpose of this meta-analysis was to determine whether the use of NSAIDs increases the risk of delayed union or nonunion in the setting of fracture, osteotomy, or fusion surgery. Secondary outcome measures included whether study quality, dose, or duration of NSAID exposure affected the reported rate of bone healing complication. In addition, we explored whether NSAID exposure demonstrated a differential effect in pediatric versus adult patients or in long bones versus the spine.