Age related effects of selective and non-selective COX-2 inhibitors on bone healing

Abstract

Introduction

Fractures are increasing worldwide and with an aging population, are frequent in the elderly. The healing of fractures progresses through various phases including the inflammatory stage. Aging is associated with slower healing and the use of non steroidal anti-inflammatory drugs (NSAIDs) may interrupt bone healing processes. We designed a study to compare the effect of diclofenac and celecoxib on fracture callus histomorphometry in a rat model of different age groups.

Methods

Using 5 and 15 month old rats, fractures were induced on the left tibia and the animals allocated to receive one of the drugs. Animals were sacrificed at day 21 and 42 and the fracture callus harvested for processing and histological evaluation. Tissue proportions and histological grades were determined and compared across the groups.

Results

Across all groups, the histological grade increased with time and animals in the young diclofenac group had the highest grade at day 42 (p = 0.004). The proportion of bone increased in all groups and was highest in the young diclofenac group at day 21 and day 42 (p = 0.003). Post hoc analysis showed that the young celecoxib and old celecoxib groups had the least proportion of bone (p = 0.032 and p = 0.003). The proportion of cartilage reduced in all groups at both time points.

Conclusion

Celecoxib was associated with lower histological grade and lower proportion of bone in older animals. We urge for caution regarding the use of celecoxib in older people for the management of pain associated with fractures. Diclofenac may be a better option in this group.

1. Introduction

Injuries including fractures are a leading cause of morbidity and mortality globally with low income countries bearing the brunt of this.¹ This has been linked to increased urbanization and mechanization of transport. In Africa, fractures are among the most common injuries with the limbs being the most afflicted sites.² The lower limb including the tibia is commonly affected. While the population of the African continent has been relatively young, it is gradually aging.³ The high proportion of injuries and the aging population is likely to result in an increase in fractures among the elderly. The physiology of the geriatric population is different from that of young adults. In particular, the aging process has been associated with a decline in the processes of bone healing.^4,5 This is associated with a longer time to union and an increase in the risk of non union. This inevitably increases the morbidity and the cost of managing the injuries.

The process of fracture healing is known to go through several stages including hematoma formation, inflammation, fibrovascular stage, bone formation and remodeling.⁶ The inflammatory stage of healing relies on the cyclo-oygenase-2 enzyme (COX-2) for the production of inflammatory markers that ultimately enable the recruitment of the necessary cells for healing.⁷ Deficiency in this enzyme has been shown to result in poor bone healing.⁸

Fractures are painful conditions requiring the use of analgesics. Non steroidal anti-inflammatory drugs (NSAIDs) are commonly used in the management of limb fractures.⁹ NSAIDs are also an integral component of multimodal analgesia which is important in reducing the quantity and duration of opioid use.¹⁰ While NSAIDs are beneficial in reducing the pain and inflammation associated with a fracture, they may be detrimental to the healing of the fracture itself. The use of NSAIDs has been reported to be associated with an increased risk of non union and delayed union.¹¹ This effect has been thought to be related to the inhibition of the inflammatory pathway.¹²

NSAIDs can be classified according to their inhibition of the two isoenzymes of COX with the COX-2 isoenzyme being the one responsible for the inflammatory pathway.¹³ Non selective or traditional COX inhibitors including diclofenac have an almost equal effect on both isoenzymes while COX-2 selective like celecoxib are more selective for the COX-2 isoenzyme.¹⁴ There are suggestions that selective inhibitors of COX-2 may have a more deleterious effect when compared to non selective inhibitors.¹⁵ Comparing a non selective with a selective COX-2 inhibitor may help to further evaluate any differences. As this effect may be even more pronounced in aged individuals, comparing young and older individuals may aid in discerning any differences.¹⁶

Analgesics are the most common over the counter (OTC) medication consumed by both adults and adolescents.^17,18 Both diclofenac and celecoxib are freely used and available as OTC medications in many parts of the world.^19,20 We therefore designed a study to compare the effects of diclofenac and celecoxib on fracture callus histomorphometry in the young and old using a rat model.

2. Materials and methods

2.1. Study design and population: animals

This quasi randomized study included fourty male Norwegian rats (Rattus norvegicus). Half of the animals were five months old and the rest aged 15 months. The ages of the animals were chosen to correspond to 18 and 50 human years respectively.²¹ Rats were chosen as they are cheap and easy to handle. The use of the rat model has also been well established in the study of the effects of aging.²² The sample size was calculated to detect a difference in the histological score of 0.3 at a sigma of 0.5 (α = 0.05 and β = 80%). In the laboratory, the animals were kept in pens floored with saw dust, at room temperature and humidity, and a 12-h light/dark cycle. They were fed on commercial pellets and given a steady supply of fresh water. Three animals in each group were used to provide the baseline data. The rest of the animals were divided into two groups of twenty animals each.

2.2. Fracture model

To simulate a common method of long bone fracture management and to standardize the fractures, a closed fracture model stabilized with an intramedullary device was chosen. At the commencement of the study, the animals were anaesthetized using ketamine 100 mg/kg and xylazine 10 mg/kg IP prior to fracture induction. The surgical area was shaved and cleaned with a preparation of iodine in alcohol. An intramedullary 1.25 mm Kirshner wire inserted from the knee joint into the tibia under sterile conditions.²³ A closed fracture of the left tibia was induced by dropping a 460 g weight from a distance of 20 cm.²⁴ Half the animals received 5 mg/kg of celecoxib (Pfeizer Pharmaceuticals LLC, Illertissen, Germany) while the rest received a similar dosage of diclofenac (Novartis Pharma AG, Basel Switzerland) in two divided doses, 12 h apart by oral gavage for twenty eight days.

2.3. Tissue harvesting and processing

At day 21 and day 42, five animals from each group were euthanized using 3% halothane in an airtight glass chamber until they had no pupillary reflexes and little response to pain. The animals were then perfused with formal saline using the transcardiac method after clearance of the blood using normal saline. The healing tibia was subsequently harvested and volume in cm3 (mls) estimated using Scherle's method.²⁵ The specimen was subsequently fixed in formal saline for 48 h before being subjected to decalcification in ethylenediaminetetraacetic acid (EDTA) for fourteen days with the end point determined by both the flexibility and calcium oxalate methods.²⁶ This was followed by dehydration in increasing concentrations of alcohol and embedding in paraffin wax before sectioning to produce ten 7 μm slides. The slides were stained using heamatoxylin and eosin (H&E) and every second slide selected for photography. Ten photomicrographs were taken for each slide using a Motic BTU8 digital camera (Motic, Kowloon, Hong Kong) mounted on a Ritchter Optica UT1 microscope (Ritchter Optica, Kowloon, Hong Kong). Every second image was used for histological analysis and stereology using and Image J software (NIH, Massachusetts, USA).

2.4. Statistical methods

The histological grade was determined using the ten point scale by Huo et al.²⁷ in which 1 represents mesenchymal fibrous tissue and 10 is mature bone. The volume density of bone, cartilage and fibrous tissue was determined by counting the points falling on a specific tissue and dividing by the number of points falling on the fracture callus. The volume density was determined using Vv_tis = ∑Pt_tis/∑Pt_tot∗100, where Pt_tis is the number of points falling on the tissue; Pt_tot is the total number of points falling on the fracture callus and Vv_tis is the volume density of the tissue (Fig. 1). Bone was described as vascularized tissue with cells in an osteoid matrix, cartilage as rounded cells in lacuna in a hyaline matrix while fibrous tissue as spindle shaped cells in a fibrous matrix.

Fig. 1.

Histological micrographs showing a section of fracture callus with a superimposed set of points for estimating volume densities generated using the Image J software.

The data from the Image J was extracted into Ms Excel (Microsoft Corp, Washington, USA) before exporting to SPSS v17.0 (SPSS Inc, Chicago, USA) for analysis. After determining normality of the data by histograms and box plots, the means, medians, standard deviations and interquartile ranges were calculated as appropriate. The one way ANOVA and Kruskal Wallis tests was used to compare the means and medians across the four groups and the p was set at 0.05. Post hoc analysis was performed using Tukey HSD and Dunn tests respectively.

2.5. Ethical considerations

Ethical approval was obtained from the animal ethics committee of the University of Nairobi and the animals were handled in accordance with the Home Office Guidance on the operation of the Animals (Scientific Procedures®) Act 1986, published by HMSO, London (or the comparable guidelines for the state of Israel). All procedures were performed according to the Guide for Care and Use of Laboratory Animals (NIH publication No. 85–23, revised 1985).

3. Results

A total of 40 male rats were included. At day 21 the diclofenac group (A) had more bone as compared to the celecoxib group which was mainly cartilage (B). At day 42, the diclofenac group (C) had a higher proportion of bone while the celecoxib group had a mixture of bone and cartilage (D) (Fig. 2).

Fig. 2.

(A–D). Photomicrographs of the fracture callus showing the fractured end of bone (asterix), the area of new bone (arrow) and the areas with cartilage (stars); Hematoxylin and Eosin stain X40. Figures A and C are the young diclofenac groups at day 21 and 42. Figures B and D are the old celecoxib groups at day 21 and 42.

The histological grade of all groups increased with time but was higher in the diclofenac group in both age groups (Fig. 3). While at day 21 there was no statistically significant difference in the histological grade (p = 0.299), by day 42 the difference across the groups was statistically significant (p = 0.004) and post-hoc analysis revealed that the difference was between the young animals that received diclofenac and the old animals that received celecoxib at day 42 (p = 0.002).

Fig. 3.

Histological grade.

The proportion of bone in the fracture callus increased in all groups over time and the differences were statistically significant at day 21 (p = 0.003) and day 42 (p = 0.003). Post hoc analysis revealed that at day 21, the proportion of bone was higher in the young diclofenac group when compared to the young celecoxib and old celecoxib groups (p = 0.032 and p = 0.003 respectively). At day 42 the young diclofenac group had a higher proportion of bone than the young celecoxib and old celecoxib groups (p = 0.045 and p = 0.003 respectively).

The proportion of cartilage in the fracture callus reduced in all groups over time (Fig. 4). Animals that received celecoxib had higher proportions of cartilage that those that received diclofenac but this difference was not statistically significant at day 21 and 42 (p = 0.222 and p = 0.056).

Fig. 4.

A and B. Proportion of bone and cartilage in the fracture callus.

4. Discussion

The present study shows that the proportion of bone in the fracture callus increased over time in all the studied groups. This was more pronounced in younger animals and those that received diclofenac. The proportion of cartilage was higher in the animals that received celecoxib. The process of fracture healing is known to go through several stages including hematoma formation, inflammation, fibrovascular stage, bone formation and remodeling.⁶ The bone is formed mainly by endochondral ossification in which cartilage formation precedes bone. It is thus expected that the proportion of cartilage should reduce while that of bone increases as the fractures heal and the findings of this study are similar to others in literature.²⁸

A new finding in this study is the demonstration that increasing age and use of COX-2 selective inhibitors resulted in persistence of cartilage and lower proportions of bone in the fracture callus. While celecoxib has been shown to inhibit the chondrogenic phase of fracture healing,²⁸ this study reveals that this effect is worse in older age groups. Older animals have been shown to have reduced fracture healing potential which may be related to senescence of the osteoprogenitor cells.^29,30 The differences in proportion of bone between the young and old animals that received a particular NSAID were not statistically significantly different. Conversely, statistically significant differences were seen when we compared the young diclofenac group with the old celecoxib group. This implies that the differences seen cannot be attributed to age alone and the drug used was a factor as well. This study therefore demonstrates that older animals may suffer the deleterious effects of COX-2 inhibitors more that younger ones. The findings in this study could be due to celecoxib inhibiting the COX-2 prostaglandin 2 (PGE2) pathway that mesenchymal stem cells (MSCs) rely on to promote osteogenesis in the initial stages of fracture healing.³¹ As the selective COX-2 inhibitors preferentially inhibit this enzyme, it is likely that their effects are more profound when compared to the non-selective NSIADs.

At a molecular level, this could be due to celecoxib suppressing Wnt target genes resulting in reduced expression of runt-related transcription factor 2 (RUNX2) and ALP consequently inhibiting osteoblast-mediated mineralization.³² Celecoxib has also been implicated to reduce the expression of Receptor activator of nuclear factor-κB ligand (RANKL) providing another probable mechanism of action.³³The consequence of this is reduction in the mechanical properties of the fracture callus, delayed union and non union. While NSAIDs are important agents in the management of pain after fractures, this study highlights the need to avoid the use of selective COX-2 inhibitors in the elderly. Although some non selective NSAIDs may have lower COX 2 selectivity, the choice of diclofenac in this study was informed by the most common molecules in clinical use.

To manage pain associated with fractures or their surgical correction, several options are available to the surgeon including paracetamol (acetaminophen), NSAIDs and opiods. This study demonstrates that selective COX-2 inhibitors may be unsuitable in the elderly to manage this pain as this may lead to delayed and non union. The surgeon should use other molecules available in his armamentum.

4.1. Limitations

While this study demonstrates differences in the histological structure of the fracture callus, evaluation of the healing callus mechanically and radiologically may provide greater insights into the potential clinical implications of our findings. The choice of drugs in the study was informed by the most common drugs in each of the classes; evaluation of more molecules may help elucidate if the findings seen are a class effect or are specific to individual molecules.

5. Conclusion

This study demonstrates poor fracture healing in older age groups with the use of celecoxib manifest by lower histological grade and a smaller proportion of bone. This could result in a mechanically weaker union and delayed union. Use of selective COX-2 inhibitors in the elderly in the setting of fractures should be discouraged. Further studies to quantify the inhibition pathways for bone healing between the young and old rats are warranted.

Funding

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

Declaration of competing interest

The authors have no conflicts to declare.