Effects of NSAID use on bone healing: A meta-analysis of retrospective case–control and cohort studies within clinical settings. (April 2020)

• Record Type: Journal Article
• Authors: Ali, Muhammad Usman; Usman, Mehvish; Patel, Kuntal
• Is Part Of: Trauma. Volume 22:Number 2(2020:Apr.)
• Page Start: 94
• Page End: 111
• Publication Date: 2020-04
• Subjects: Non-steroidal anti-inflammatory drugs -- cyclooxygenase inhibitors -- bone healing complications -- fractures -- spine; Traumatology -- Periodicals; Disaster medicine -- Periodicals; Wounds and injuries -- Periodicals; Electronic journals; 617.1
• DOI: 10.1177/1460408619886211
• Languages: English
• ISSNs: 1460-4086

Abstract

Introduction: This meta-analysis aims to determine whether non-steroidal anti-inflammatory drug (NSAID) use is significantly associated with adverse bone healing outcomes within clinical settings, including trauma and elective spine settings. It will also explore bone healing outcomes with the type, route, dosage and duration of NSAID exposure and aims to demonstrate the effects of various other confounding factors on bone healing outcomes. Methods: Electronic databases including MedLine, Embase and Cochrane were searched from January 1975 to December 2017. A distinct analysis of observational long bone and spine studies in adult populations was performed, assessing the effects of NSAID exposure on bone healing. Studies from paediatric population and randomised trials also underwent separate assessment in the synthesis. Meta-analysis was conducted in compliance with QUORUM and PRISMA guidelines. Results: Quantitative assessment of observational studies suggested a significant risk of bone healing complications with NSAID exposure across each clinical group: long bone (p = 0.0004) and spine (p = 0.02). Analysis of paediatric studies revealed a non-significant association of poor outcomes with NSAID use (p = 0.36), while assessment of randomised trials demonstrated a statistically significant risk of complications following NSAID administration (p = 0.04). Meta-regression further suggested smoking to be a substantial confounder associated with adverse bone healing outcomes (p < 0.00001). Conclusion: Inclusion of only fair and moderate-quality retrospective cohort and case–control studies in the current synthesis limits the clinical application of its findings and therefore warrants the need for further research. Thus, attempts to conduct high-quality prospective cohort and randomised trials to study the effects of NSAID use on bone healing would be very helpful and will provide a basis for more extensive research in future.