Do Non-Steroidal Anti-Inflammatory Drugs (NSAIDs) Adversely Impact Fracture Healing? A Critical Review of the Literature

Abstract

Purpose of Review

The purpose of this review was to critically appraise the literature and establish an evidence-based clinical guideline for the use of non-steroidal anti-inflammatory drugs (NSAIDs) in a fracture setting.

Recent Findings

With few exceptions, studies in animals suggest that NSAIDs impair fracture healing. It is unclear if nonselective or cyclooxygenase-(COX)2-selective NSAIDs pose differing effects on fracture healing. Human studies show NSAID use to be a consistent risk factor for fracture non-union in skeletally mature populations across the literature and indicates that indomethacin in particular poses a significant risk for non-union of adult acetabular fractures.

Summary

Current evidence appears to suggest no harm in using ketorolac or ibuprofen in a pediatric fracture population, while indomethacin poses a significant risk for non-union in adult acetabular fracture patients when used for six weeks. Despite the majority of available clinical studies showing NSAID use as a recurring risk factor for fracture non-union in adult populations, a lack of standardization amongst studies makes it difficult to determine any clinical recommendations about timing, dosage, duration, or type of agent administered. More high-quality prospective studies are needed.

Introduction

Non-steroidal anti-inflammatory drugs (NSAIDs) are among the most commonly used medications in the world and the tenth most commonly prescribed drug in the United States [1] with increasing use. In 2014, nearly 20% of Americans over the age of 18 reported taking NSAIDs at least three times per week for three months, a 41% increase compared to 2005 [2]. NSAIDs have anti-inflammatory, analgesic, and anti-pyretic properties, and anti-platelet activity. Their uses are thus numerous, and indications range from headaches, menstrual cramps, and musculoskeletal pain to secondary prevention of myocardial infarctions.

The major action of NSAIDs is inhibition of two isoforms of the cyclooxygenase enzyme (COX-1 and COX-2), resulting in decreased synthesis of prostaglandins and thromboxane A2. This inhibition may be reversible or irreversible, depending on the drug, and may be non- selective or selective for COX-2. Constitutive expression of cyclooxygenases differs by tissue, with COX-1 suggested to be responsible for prostaglandin E (PGE) production for cytoprotection of the gastric mucosa and regulation of renal vascular tone [3].

Originally thought to be inducible only in inflamed tissues and pathologic conditions, COX-2 has since been discovered to be active at baseline in the endocrine system, uterus, kidneys, and other tissues [4–6]. Recent discovery of COX-2 expression in the epithelium of the stomach and colon may help to explain the adverse gastrointestinal effects elicited by selective COX-2 inhibitors developed in an effort to target inflammation without affecting the gastric mucosa or kidneys [3, 6]. Differential expression of the two isoforms in bone at baseline and during bony healing may also partially account for discrepancies in studies examining the effects of NSAIDs on fracture repair.

Fracture healing occurs by two mechanisms depending on the mechanical stability of the bone microenvironment. Primary or “direct” fracture healing occurs under conditions of absolute stability and fragment contact at the fracture site. Microscopically, in direct fracture healing, “cutting cones” of osteoclasts form from osteons near the ends of fracture fragments, resorbing bone longitudinally, only to be followed by trailing osteoblasts laying down new lamellar bone. Direct fracture healing is estimated to take between one and two months and can deposit new bone to close gaps up to one millimeter in length [7]. When gaps closer to one millimeter are present, woven bone is laid down first and then remodeled into mature, lamellar bone, which is more suitable biomechanically for load bearing.

Fractures can also heal via by secondary or “indirect” healing. This process occurs when mechanical strain at the fracture site is between two and ten degrees of motion and can have both intramembranous and endochondral components. Indirect fracture healing begins with hematoma formation, as fracture fragments expose bony vascular supply to the surrounding soft tissues, with a contribution from periosteal vasculature. The fracture hematoma contains growth factors such as vascular endothelial growth factor (VEGF), mesenchymal stem cells, fibroblasts, endothelial cells, and white blood cells [8]. The fracture hematoma is present through approximately the first two weeks in a stage known as the inflammatory phase. Granulation tissue is then established by fibroblasts and white blood cells within the fracture hematoma. The extracellular matrix within this granulation tissue guides osteogenic and chondrogenic precursor cells about the fracture site. Then, over several weeks in the “repair” phase, intramembranous ossification occurs in a predominantly subperiosteal location, while soft callus is established within the granulation tissue of the fracture hematoma by chondrocytes via an endochondral mechanism. This process gradually proceeds with bony cortex establishment across the fracture site, which will then calcify and eventually undergo bony deposition and eventual maturation and remodeling [8].

While commonly prescribed for musculoskeletal pain, numerous animal studies have demonstrated potential adverse effects of NSAIDs on bony healing including reduced mechanical strength and increased rates of non-union in the context of fractures and spinal fusion [2, 9–11]. Previous review articles have described a consensus in the animal literature regarding the association of NSAIDs with the risk of non-union but acknowledge the lack of human data to support or refute this finding [12–14]. Few human studies have been conducted, and those that exist have demonstrated conflicting results [15, 16]. One systematic review of human studies has argued that existing human studies are not generalizable beyond the elderly population [17]. Such reviews have also described the need for exploration of the underlying mechanisms by which this effect might occur [13]. Regardless, public opinion among orthopaedic surgeons is near-unanimous: a 2012 survey found that 86% of orthopaedic attending respondents believed NSAIDs to be a risk factor for non-unions [18].

This review therefore set out to re-examine NSAID pharmacology in the context of the physiology of fracture repair and to clarify the effects of these drugs on histologic, radiographic, and mechanical measures of bony healing.

Materials and Methods

Relevant articles were identified utilizing the PubMed/MEDLINE and Embase databases for relevant publications from January 1, 1990 to December 31, 2022. The following search terms were utilized: “NSAID”, “ibuprofen”, “celecoxib”, “cyclooxygenase”, “prostaglandin”, “fracture”, “nonunion”, “fracture healing”, “bone formation”, and “ossification”. Screening was performed at the title and abstract stages, and screening for inclusion was then performed via full-text review by two of the authors (N.V.S. and B.G.D.). A focus was placed on studies within the most recent 10 years, although commonly cited or classic studies from earlier were not excluded. Levels of evidence that pertained level V evidence were excluded. Original studies and systematic reviews were included if they focused on NSAID pharmacology, fracture healing, cyclooxygenase (COX) enzyme mechanisms in fracture healing, effects of NSAIDs on fracture healing in animals, and effects of NSAIDs on fracture healing in humans. Human studies investigating the effect of NSAIDs on fracture healing were appraised according to level of evidence, and the findings were summarized.

Role of COX Enzymes in Fracture Healing

Prostaglandins, specifically PGE, have been shown to play a role in bone formation in response to mechanical stimulation, in part via downstream activation of insulin-like growth factor 1 (IGF-1) transcription [19]. In vitro studies have shown mechanical loading of murine osteocytes causes up-regulation of COX-2 and increased production of PGE, suggesting COX-2 involvement in the remodeling phase of both direct and indirect fracture healing [20].

PGE2 has also been shown to influence bone resorption, osteoclast differentiation, and osteoblastogenesis, and thus may have the overall effect of balancing osteoclast and osteoblast activity in bone turnover [21–24].

While our understanding of PGE2 on bone metabolism deepens, the relative responsibilities of COX-1 and COX-2 in prostanoid production in bone remain uncertain. Previous understanding held that COX-1 is constitutively active in osteoblasts, osteoclasts, and osteocytes, and COX-2 expression is induced at the site of a fracture or inflammatory source by interleukins and other mediators [25]. A recent review aggregating studies on COX-2 knockout mice, however, suggests lower bone density in COX-2 knockouts as well as compensatory alterations in calcium and vitamin D metabolism due to the combined skeletal and renal effects COX-2 deficiency [21].The authors conclude “COX-2 is dominant in osteoblasts,” a departure from the earlier conception of this isoform as inducible only in settings of trauma or inflammation. The role of COX-1 is less certain.

A single study assessed the relative timelines of COX-1 and-2 mRNA production in femoral fracture calluses of rats. COX-1 and − 2 mRNA levels were nearly equal at baseline, and COX-1 mRNA did not rise in the first 21 days post-fracture. In contrast, COX-2 mRNA was elevated over baseline by a factor of ten during days three to fourteen post-fracture, returning to baseline by day 21 [4]. Additional human and animal studies have assessed the levels of COX-2 in fracture callus, with one human study finding that the calluses surrounding non-unions had significantly lower COX-2 concentrations than those of healed fractures [26]. Another study performed on mice found high co-expression of COX-2 and bone morphogenic protein (BMP)-7, a stimulator of bone formation and heterotopic ossification, in fracture callus [27]. The same study found NSAIDs to decrease the mass of bone contained in heterotopic nodules whose formation was stimulated by BMP-7, suggesting the anabolic effects of BMP-7 on bone may in part be mediated through COX-2 expression. These studies posit an understanding of COX-2 as constitutively active in bone, with a fracture-inducible elevation to support healing. COX-2 knockout mice have been found to have fewer osteoblasts and chondrocytes, with abnormal persistence of mesenchymal stem cells in callus [28]. COX-2 also appears to interact with BMP-2 in the regulation of chondrocyte differentiation during endochondral ossification [3, 29]. A separate investigation in an osteosarcoma cell line found reversible suppression of osteoblastic differentiation and bone nodule formation in the presence of a COX-2 inhibitor [30]. A 2017 study demonstrated inhibition of COX-2 increased degradation of a transcription factor in the Wnt-beta catenin pathway, which may prevent osteoblast maturation [31]. These findings suggest two additional mechanisms by which NSAIDs may affect indirect fracture healing in the repair phase: reduced chondrocyte differentiation and reduced osteoblast maturation. As such, both intramembranous and endochondral ossification mechanisms can be affected by COX-2 inhibition either by inhibition of osteoblast differentiation and maturation, or by the enzyme’s effects on chondrocytes.

Another function of the PGE and prostacyclin (PGI) produced by COX-1 and − 2 is vasodilation, and their expression can thus increase blood flow to the site of a fracture. Vasodilation enhances hematoma formation in the inflammatory stage of indirect fracture healing allowing white blood cells to direct deposition of the granulation tissue scaffold for the repair phase. This provides yet another mechanism by which NSAID-induced COX inhibition may affect indirect healing in its initial stage.

While the interplay of COX enzymes and their products on fracture healing has not yet been entirely elucidated, multiple mechanisms by which COX inhibition may modulate the indirect and direct pathways of healing have been identified and are summarized in Table 1. These mechanisms affect all three stages of the indirect pathway as well as direct fracture healing, underscoring the need for rigidly standardized studies with respect to both timing and fracture stabilization methods. Additionally, as COX-1 appears to be expressed at the same levels in non-traumatized and healing bone, further inquiry into the biochemical effects of isolated COX-1 inhibition is necessary to construct a more complete comprehension of each isoform’s role in healing.

Table 1.

Effects of Inhibition of cyclooxygenase on fracture healing by healing stage

Stages of Fracture Healing	Mechanisms
Direct pathway	Reduced osteoclast differentiation and osteoblastogenesis Inhibition of PGE-induced IGF-1 transcription and PGE response to mechanical loading
Indirect pathway
Inflammatory phase	Reduced recruitment of white blood cells and reduced vasodilation
Repair phase	Reduced chondrocyte differentiation and osteoblastogenesis Abnormal persistence of MSCs in callus Inhibition of PGE-induced IGF-1 transcription
Remodeling phase	Reduced osteoclast differentiation and osteoblastogenesis Inhibition of PGE2 response to mechanical loading

Effects of NSAIDs on Fracture Healing in Animals

Multiple animal studies have been undertaken over the past two decades to elucidate the effects of NSAIDs on fracture healing.

One early animal study found no difference in ulnar fracture healing between rabbits on low-dose ketorolac, a non-specific NSAID, and no NSAID, but found high ketorolac doses to slow mineralization between weeks 2 and 4 of healing [32]. A more recent rat study found ketorolac had no effect on callus quality or mechanical strength of healing tibia fractures during the first 21 days after injury [33]. A rat study comparing ketorolac to the COX-2-selective NSAID parecoxib, however, found ketorolac to reduce mechanical strength of the fracture site relative to controls. Union rates in both experimental groups were lower than in the control group at 21 days post-fracture, but all groups experienced union of all fractures at 21 days [4]. A follow-up study by the same investigators reported no significant differences in healing at 35 days post-injury between controls, ketorolac, and COX-2-selective NSAID groups, but significantly more non-unions in the COX-2-selective group at day 21 [5]. The authors found that fracture calluses from rats in the COX-2-selective group contained more PGE than those in the ketorolac group but exhibited delayed cartilage remodeling. Similar histological findings have been replicated with the non-selective NSAID carprofen in beagles and ibuprofen in rabbits, demonstrating the importance of COX enzyme activity during the repair phase of indirect healing and endochondral ossification [34, 35].

Indomethacin, an NSAID with relative COX-1 selectivity, was found to decrease force at failure in mechanical testing of healing murine diaphyseal femur fractures, but no difference in metaphyseal screw pullout was demonstrated [36]. The authors hypothesized the difference in fracture locations might explain these contradictory mechanical testing findings, and metaphyseal bone was better able to recruit mesenchymal stem cells than diaphyseal bone to counteract any deficit in differentiation due to COX inhibition. An inhibitory effect of indomethacin on fracture repair was replicated, with persistent defects in healing at 28 days post-fracture when compared to control mice [37]. However, a prior study in male rats found no difference in radiographic evidence of healing or mechanical testing between rats given indomethacin or no NSAID [38]. The same authors found histologic evidence of altered bone formation in rats given a COX-2-selective NSAID.

A randomized controlled trial of 240 mice receiving either a COX-2-selective NSAID or no drug showed effects in the later stages of fracture healing on radiographs and mechanical testing. The NSAID group demonstrated lower blood flow to the fracture site, but NSAIDs were found to be an independent predictor of poor healing beyond low blood flow [39], underscoring the importance of the vasodilatory properties of COX activity for fracture repair and that multiple mechanisms are likely responsible for the enzyme’s effects on bone healing.

A 2013 investigation found no effect of COX-2 inhibition on endochondral and intramembranous bone grafts in rabbits [40]. The rabbits, however, were examined at day 60 post-graft, after which an early, reversible delay in healing may have been missed. The absence of a long-term effect of COX-2 inhibition has been corroborated by additional studies which found no difference in the mechanical testing results of healed fractures between COX-2- specific NSAID and control groups at four to six weeks post-fracture [41, 42].However, deleterious effects of COX-2 inhibition on fracture repair have been found in rabbits and rats at earlier time points in the healing process, and an additional rat study found higher non-union rates in rats treated with celecoxib versus acetaminophen at eight weeks post-fracture [3, 43, 44]In one of the latter studies, the authors also found ibuprofen to be associated with decreased bone remodeling at 6 weeks after a stress fracture, while the COX-2-selective agent had no such long-term effect, raising questions about dosing timelines and isoform selectivity [43]. This finding was replicated in rabbits, with reduced peak torque at six weeks post-fracture in the ibuprofen group compared to controls [35]. The same study included a COX-2-specific group, which exhibited inferior mechanical testing results at fracture sites as well as a higher. NSAIDs were administered for 4 weeks post-fracture.

Another comparative study in rats utilizing the COX-2-specific drug parecoxib and the non-selective drug diclofenac for the first seven days post-fracture found no difference in healing and mechanical testing at 30 days [45]. These results were challenged by a newer study demonstrating diclofenac significantly reducing callus stiffness and volume compared to controls at 21 days post-fracture in rats [46]. However, rats in the more recent study appear to have been given diclofenac daily from the time of surgery until the day of sacrifice (day 21), resulting in a period of use three times longer than the prior study. This difference in dosing regimen and timing with respect to the stages of fracture healing may help explain the discrepancy. Another study in male rats administered dexketoprofen (non-selective), meloxicam (COX-2-selective), or diclofenac (non-selective) for 10 days post-fracture, and found only dexketoprofen and meloxicam to significantly impair healing [47]. Diclofenac treatment has also been found to result in fewer osteoblasts, osteoclasts, and macrophages at the site of bony injury in mice [48]. These findings suggest that short-term, early use of the non-selective agent diclofenac may not permanently impair fracture healing, but longer use may have adverse effects. Less is certain regarding dexketoprofen, which has previously been found not to affect fracture healing in rats at 2, 4, and 6 weeks post-injury, and meloxicam, which has been noted to reduce bone nodule formation in vitro [30, 49].

One 2005 study sought to elucidate the effect of NSAID timing. Rats with femoral shaft fractures were given etodolac, a COX-2-selective NSAID, for either the first week post-fracture, the third and final week before sacrifice, or for three continuous weeks. All three test groups exhibited poorer radiographic evidence of healing and mechanical testing scores compared to a non-NSAID control group [50]. However, no comparison with a non-selective agent was employed, and no long-term testing was performed to assess reversibility.

The effects of NSAIDs on fracture healing are likely agent-, timing-, and dose-dependent and may be reversible if treatment is short-term and/or in the early stages after a fracture. It remains difficult to draw more detailed conclusions regarding the timing of doses or the relative COX-2 selectivity of agents given the lack of standardization of these parameters across studies.

Effects of NSAIDs on Fracture Healing in Humans

Impact of NSAIDs on Fracture Healing and Bony Fusion

Despite widespread interest and a pervasive belief among surgeons that NSAIDs impair fracture healing, few human studies have been conducted to validate this suspicion. Beyond the fracture literature, a Level II systematic review determined NSAIDs may have dose dependent and duration-dependent effects on spinal fusion outcomes [2]. One 2014 Level III retrospective review of 1901 patients with long bone fractures found postoperative NSAID use to carry an odds ratio of 2.17 for fracture complications (including malunion, non-union, and infection) [51]. Another Level III retrospective review of nearly 10,000 patients with humeral shaft fractures found those exposed to NSAIDs within 90 days post-fracture had a relative risk for non-union of 3.7 [52]. An additional Level III case-control series in patients with femur fracture non-unions found post-fracture NSAID use to be a risk factor when compared to patients with healed fractures [53]. A recent review, however, suggests the association found in these retrospective studies may reflect the need for analgesia in cases of painful non-unions rather than a causal relationship between NSAIDs and poor fracture healing [22].

Role of NSAID Dosage, Route, Timing, Delivery, and Duration

Given variations in clinical care and methodology, timing, dosing, and duration of NSAID exposure or intervention remain unstandardized between studies. One Level III retrospective investigation of 80 patients found no difference in time to radiographic union in patients who received a single post-operative dose of ketorolac within 24 h compared to patients who did not [54]. Only one large Level III retrospective database analysis attempted to clarify the effects of timing, in which 309,330 fractures were reviewed to elucidate the effects of multiple medications on fracture healing. Both acute and chronic use of NSAIDs were associated with increased risk of non-unions [55]. Dose timing is especially relevant given that chronic use prior to injury has been suggested to affect bone metabolism due to impaired PGE production. A large United Kingdom Level III retrospective cohort study of over 200,000 regular NSAID users found an increased rate of fractures compared to matched controls [56].

Other Factors to Consider in NSAID Use in the Setting of Fracture Healing and Bony Fusion

As appropriate pain control is of high clinical importance, lack of adequate placebo control complicates prospective human study design and is not consistently accounted for in retrospective analysis. While a recent multicenter Level III retrospective cohort study found no difference in time to radiographic healing after tibial intramedullary nailing between patients who received postoperative NSAIDs or postoperative opioids for pain control [57], no non-NSAID control group was included in this study. As opioids themselves have recently been associated with fracture non-union and impaired spinal fusion [55, 58], it is possible the authors’ findings represent delayed healing in both experimental groups compared to non-treatment.

Of the three available prospective randomized controlled trials (RCT) that have been undertaken in adult populations two have demonstrated a correlation between NSAIDs and non-union. A Level II RCT found a significantly higher rate of non-union in 38 patients with acetabular fractures who received indomethacin for prevention of heterotopic ossification when compared to patients who received radiation [10]. A Level I RCT studying 98 patients similarly found an increased incidence of non-union in acetabular fracture in patients who receive 6 weeks of indomethacin, but no increase in patients receiving 1 week of indomethacin, suggesting a duration-dependent effect [59]. These findings alone cannot be generalized across agents or fracture types, however, and must be replicated before treatment recommendations can be formulated. The third study, a Level I RCT, found no difference in the rate of healing of non-operatively managed Colles’ fractures in a COX-2-selective NSAID group versus a placebo group, and a non-significant decrease in the treatment group’s forearm bone mineral density [9]. This study contained only 42 subjects and may have been underpowered.

Finally, patient age must be accounted for. Adverse effects of NSAIDs appear more relevant to adult trauma patients, as two Level III retrospective reviews involving 221 and 808 pediatric patients found no association between ketorolac or ibuprofen (respectively) and complications in bony healing [60, 61]. A prospective Level I RCT examining post-fracture pain control in children with ibuprofen versus acetaminophen reported no non-unions to permit analysis [62]. A summary of all findings is listed in Table 2.

Table 2.

Summary of NSAID effects on fracture healing

Statement
NSAID use may be a risk factor for non-union in long bone fractures [51–53, 56, 57]
There may be dose-dependent effects of NSAIDs on non-union in long bone fractures, though evidence is limited [2, 54, 56]
NSAIDs may have duration-dependent effects on non-union in long bone fractures, though evidence quality is limited [2, 52, 54–57]
NSAID (specifically indomethacin) use increases the risk of non-union in acetabular fracture patients when used for six weeks [10, 59]
Chronic NSAID use is a predisposing risk factor for bone fracture, though recent clinical supporting evidence is limited [55]
NSAIDs (specifically ibuprofen and ketorolac) may pose minimal risk for non-union in a pediatric fracture population [60–62]

Discussion

Cyclooxygenase 1 (COX-1) is likely constitutively active in bone without significant post-fracture increase, while COX-2 may exhibit constitutive expression of a similar magnitude but with a fracture-inducible elevation. Multiple mechanisms by which both non-specific and COX-2 specific inhibition may impede fracture healing have been identified. These include effects on the indirect and direct pathways, as well as intramembranous and endochondral ossification.

Though results of animal and biochemical studies vary by drug, the consistency of histologic callus findings across agents and species suggest that NSAIDs likely have detrimental effects on fracture repair, but these effects may be reversible if NSAID use occurs early after an injury and for a short period. The timeline of effects may vary between COX-2-selective and non-selective agents, with non-selective agents having longer-lasting or later effects [63]. This class difference, along with timing of administration, may account for much of the disagreement seen in the animal literature. As noted previously, dosing in animal studies is generally higher than in human clinical dosing, which may account for discrepancies between findings in human and animal literature [8]. This observation, in conjunction with the finding that a single dose of post-operative ketorolac was not associated with an increased non-union rate, may ease the surgeons considering this post-operative regimen [54]. More recently, a recent meta-analysis of six randomized controlled trials, once stratified by treatment duration, demonstrated no nonunion risk for use < 2 weeks when compared to use > 4 weeks [64].

Given the limited literature available examining NSAIDs’ effects after fracture in humans, few conclusions can be drawn. Retrospective studies suggest an association with fracture non-union in adult and elderly populations, but do not reliably account for variations in agents, dosing, timing, or fracture site. In prospective studies assessed in skeletally mature populations, indomethacin was shown by two high quality RCTs to increase the risk of non-union in acetabular fractures, with a potential duration-dependent effect, but RCTs lack power and generalizability to other agents, doses, and fracture sites. In pediatric populations, retrospective cohort studies and a well-powered RCT suggest no effect of ketorolac or ibuprofen on bony healing. Overall, given the lack of standardization across human studies, more high-quality cohort and prospective studies are needed to formulate a comprehensive set of guidelines for NSAID administration in the setting of fracture.

Conclusions

With few exceptions, studies in animals suggest that NSAIDs impair fracture healing, an effect that is likely reversible if use is short-term in the early period after injury. It is unclear if non-selective or COX-2-selective NSAIDs pose differing effects on fracture healing. While some available clinical studies have shown NSAID use as a recurring risk factor for fracture non-union in adult populations, a lack of standardization amongst studies makes it difficult to determine any clinical recommendations about timing, dosage, duration, or type of agent administered. Despite this, current evidence appears to suggest no harm in using ketorolac or ibuprofen in a pediatric fracture population, and indomethacin poses as a significant risk for non-union in acetabular fracture patients.

Key References

• Sagi HC, Jordan CJ, Barei DP, Serrano-Riera R, Steverson B. Indomethacin Prophylaxis for Heterotopic Ossification after Acetabular Fracture Surgery Increases the Risk for Nonunion of the Posterior Wall. J Orthop Trauma. 2014;28:377–83.

• • This paper highlights, in a robust clinical trial/setting, that indomethacin, which had long been recommended for use for heterotopic ossification prophylaxis, actually conferred risk for posterior wall nonunion following acetabular fixation.

• Donohue D, Sanders D, Serrano-Riera R, Jordan C, Gaskins R, Sanders R, et al. Ketorolac Administered in the Recovery Room for Acute Pain Management Does Not Affect Healing Rates of Femoral and Tibial Fractures. J Orthop Trauma. 2016;30:479–82.

• • This study highlights an important clinical tool in perioperative pain management, demonstrating safety of use of ketorolac in the acute pain management phase immediately postoperatively and the lack of association with healing rates in long-bone fracture fixation.

• George MD, Baker JF, Leonard CE, Mehta S, Miamo TA, Hennessy S. Risk of Nonunion with Nonselective NSAIDs, COX-2 Inhibitors, and Opioids. J Bone Joint Surg Am. 2020;102(14):1230–8.

• • This large-sample study utilizing a robust insurance claims database demonstrated that COX-2 inhibitors, and not nonselective inhibitors, were associated with greater risk of nonunion, and that while opioids were found to be associated with nonunion risk, these patients utilizing them may have had more complex fracture patterns.

• Farii HA, Farahdel L, Frazer A, Salimi A, Bernstein M. The effect of NSAIDs on postfracture bone healing: a meta-analysis of randomized controlled trials. OTA Int. 2021;4(2):e092.

• • This meta-analysis of six randomized controlled trials provides strong evidence demonstrating that while nonunion risk was higher in patients utilizing NSAIDs, this risk decreased to nonsignificance once indomethacin was excluded. Moreover, this also demonstrated no significant nonunion risk in patients receiving NSAIDs for < 2 weeks compared to those utilizing them for > 4 weeks.

Author Contributions

Authors SGS, LP, and JSA wrote the main manuscript text and interpreted the data/literatureAuthors FAS, RB contributed to literature review, tables and references preparation, and general editingAuthors SGS, NVS, JBM, CBP, and BGD were involved in data/literature interpretation, critical revision and partial drafting the manuscript text, were involved in the guidance and oversight of the manuscript.All authors approved the version to be published, and all authors agree to be accountable for all aspects of the work and agree that there are no questions related to the accuracy or integrity of any part of the work.

Funding

This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Human and Animal Rights and Informed Consent

This article does not contain any studies with human or animal subjects performed by any of the authors.

Competing Interests

Sarah G. Stroud, Lara Passfall, Juhayer S. Alam, Frank (A) Segreto, Rachel Baum, Neil V. Shah, Jad Bou Monsef, and Carl (B) Paulino declare that they have no conflict of interest. Bassel G. Diebo reports he has received speaker/consulting honorarium from Clariance, Spineart, and Spinevision, and research support from Alphatec Spine and Medtronic.