To the Editor:

I have read with interest the article by Brown et al(1). They state that celecoxib did not delay fracture healing in their model. However,when looking at their data, one can come to an opposite conclusion.

It is well known that in young rats, the biomechanical strength of the femur recovers in 5 weeks after production of a fracture(2). Most authors studying factors that might delay fracture healing evaluate it at 4-5 weeks after fracture (3).

So, the evaluation of fracture healing at 12 weeks, as performed by Brown et al(1) in young adult rats is meaningless. At 12 weeks, nearly all fractures are healed even if we give substances that delay healing. The only exception could be the administration of substances that block fracture healing leading to non-union, which is not the case in this study.

If we look closer at the results at four and eight weeks of healing, one can see that the histologic grade found is significantly lower in the celecoxib and indomethacin-treated groups than in the controls. Moreover, the strength of the callus in the mechanical test (the “gold standard” in evaluating fracture healing) is significantly lower in the indomethacin- treated group at 4 weeks, and, as one can see in their figure 5, the values for celecoxib are also quite low, and near the significance limit.

When we take into account the results obtained at four and eight weeks, one can state that the more likely effect of treatment with celecoxib after fracture healing is to delay it. If more animals had been used, and more quantity of drug given, the results could probably have been more dramatic.

References:

1.- Brown KM, Saunders MM, Kirsch T, Donahue HJ, Reid JS. Effect of COX-2-specific inhibition on fracture-healing in the rat femur. J Bone Joint Surg Am 2004; 86-A:116-123.

2.- Ekeland A, Engesaeter LB, Langeland N. Influence of age on mechanical properties of healing fractures and intact bones in rats. Acta Orthop Scand 1982; 53:527-534.

3.- Bonnarens F, Einhorn TA. Production of a standard closed fracture in laboratory animal bone. J Orthop Res 1984; 2:97-101.

To the Editor:

I read with interest “Effect of COX-2 Specific Inhibition on Fracture -Healing in the Rat Femur” (2004, 86:116-123) by Brown, et.al. The authors concluded that celecoxib does not significantly delay healing following fracture of adult rat femora at 12 weeks. Their results and conclusions are quite contradictory to other research on the effects of COX-2 inhibitors on bone healing.1,2,3

Using the same rat fracture model,(4) we reported that rats with experimentally created non-displaced femoral fractures taking rofecoxib were significantly more likely to have clinical non-unions (p <_0.0001 reduced="reduced" radiographic="radiographic" healing="healing" maturity="maturity" p="p" increased="increased" average="average" fracture="fracture" angulation="angulation" greater="greater" mean="mean" callus="callus" width="width" and="and" a="a" lower="lower" histological="histological" grade="grade" compared="compared" to="to" control="control" rats.1="rats.1" goodman="goodman" et.="et." al.="al." _2="_2" simon="simon" _3="_3" noted="noted" similar="similar" detrimental="detrimental" effects="effects" of="of" cox-2="cox-2" inhibitors="inhibitors" on="on" bone="bone" formation.="formation." /> An explanation for these contradictory results and conclusions may be that the rats in the study by Brown et. al. did not receive a therapeutic dose of celecoxib. In the Materials and Methods section, the authors dosed the rats at 3 mg/kg/day as this dose was “in the range of the average recommended therapeutic dose in humans” but did not measure serum concentration of celecoxib. However, rats have a lower oral bioavailability of celecoxib compared to humans due to significant pre- systemic hepatic metabolism5 and it is inappropriate to perform a simple weight adjusted dosage of rats while assuming that their metabolism and absorption is similar to humans. The package insert for Celebrex6 states that the two-fold human exposure in male rats is 200 mg/kg/day based on AUC0-24 of serum concentrations in humans and rats. If we assume that the equivalent therapeutic human exposure for rats is ½ of the 2 fold exposure (the package insert states that “both peak and plasma levels…and the area under the curve are roughly dose proportional across the clinical dose range”) this suggests that a more appropriate dosing for the rats would have been 100mg/kg/day. This would suggest that 3 mg/kg/day of celecoxib in rats may be significantly subtherapeutic and might not demonstrate an observable effect in fracture healing.

Our study in adult male rats used the availability of pharmacokinetic data on COX-2 inhibitors in rodents.6,7 Rats were given a dose of 8 mg/kg/day of rofecoxib to simulate two-fold human exposure based on AUC0- 24 of serum concentrations in humans and rats. At that dosage level, we found that rofecoxib produced a significant effect in delaying bone healing.

As COX-2 is induced in injury and inflammation, our results and those of Goodman et. al. and Simon et. al. suggest that COX-2 inhibitors should be used with caution in situations that require bone healing.

1 Leonelli, S, Goldberg B, Safanda, J, Bagwe M, Sethuratnam, S King, S. The Effect of Cyclooxygenase-2 (COX-2) Inhibitors on Bone Healing. 48th Annual Meeting of the Orthopaedic Research Society, Transaction #0065, Vol. 27, Dallas, TX, February 2002

2 Goodman SB, Ma T, Ikenoue T, Matsura I, Trindade M, Fox N, Wang N, Genovese M, Smith RL: COX-2-Selective NSAID Decreases Bone Ingrowth in Vivo. 48th Annual Meeting of the Orthopaedic Research Society, Transaction #0066, Vol. 27, Dallas, TX, February 2002.

3 Simon AM, Sabatino CT, O’Connor JP: Effects of Cyclooxygenase-2 Inhibitors on Fracture Healing. 47th Annual Meeting of the Orthopaedic Research Society, Transaction #0205, Vol. 26, San Francisco, CA, 2001.

4 Bonnarens F, Einhorn TA: Production of a standard closed fracture in laboratory animal bone. J Orthop Res, 2: 97-101, 1984.

5 Guirguis MS, Sattari S, Jamali F: Pharmacokinetics of celecoxib in the presence and absence of interferon-induced acute inflammation in the rat: application of a novel HPLC Assay. J Pharm Pharmaceut Sci 4:1-6, 2001.

6 G.D. Searle & Co., Celebrex Package Insert. 1998.

7 MERCK & Co., Vioxx Package Insert. 1998.

We thank Drs. Ambrose and Aspenberg for their observations regarding our article. To address Dr. Aspenberg’s concerns, we chose our animal doses of celecoxib based on human dosing to minimize the additional confounding factors of COX-1 cross-reactivity and celecoxib side effects in our study. Previous studies have used widely varying doses of COX-2 inhibitor with widely varying results(1-6). Subsequent extrapolation of data from these studies has been difficult. We hoped with our study to be able to understand more of the effect of celecoxib on fracture healing in humans based on the results we found in animals given proportionate amounts of the drug.

In reviewing the article by Paulson et al.(7) regarding the pharmacokinetics of celecoxib in rats, we noted that, contrary to Dr. Aspenberg’s concern that celecoxib is eliminated too quickly in the rat to exert a response, celecoxib is indeed demonstrated in many tissues including musculoskeletal tissue within a few hours of oral administration. In addition, the oral administration required to demonstrate this tissue distribution was 2mg/kg in their study. This dose is even lower than the human-based dose of 3mg/kg used in our study. Though the male rats exhibited excretion of celecoxib within a few hours after administration, Paulson demonstrated discrete evidence of distribution of celecoxib to musculoskeletal tissue. We agree with Dr. Aspenberg that more studies are necessary regarding animal dosing, plasma concentration, and COX-1 cross-reactivity of celecoxib before more definitive conclusions for human use can be drawn.

To address Dr. Ambrose’s letter, we consulted with our statistician regarding the power of our studies. 19 animals were sacrificed at each time point, yielding an average sample size of 4-5 femora for mechanical analysis and 2-3 femora for histological specimens. Though our original power analysis predetermined this to be an adequate sample size for our study, reanalysis demonstrated our powers to be lower (20-90) for most of our results, likely due to larger coefficients of variation than anticipated. Thus, future studies in this area will indeed need larger sample sizes for more statistically accurate findings. However, the trends we found in healing rates of fractured femora in control- and indomethacin -treated animals were consistent with results found in historical studies(8-10), demonstrating that our positive and negative control groups were consistent with previous bone-healing studies. Thus, future studies should include a more specific analysis of dose- and time-dependent effects of celecoxib on bone healing as well as a larger study group.

Regarding our statistical analysis of results, we compared the results of each group against each other with three separate unpaired two- tailed t-tests and adjusted for multiple comparisons. This gave us the same results as an Analysis of Variance with post-hoc comparison for the three treatment groups, although it was more labor-intensive. In retrospect, the ANOVA with post-hoc comparison would have been easier and more concise.

To clarify our specimen preparation, femora for histological analysis were fixed immediately after harvest in paraformaldehyde. Femora for mechanical studies were preserved at -20°C due to testing machine availability at our facility. As noted in our study, femora were not preserved for longer than 10 days prior to mechanical testing.

Much interest is currently being directed to the effect of COX-2 inhibition on fracture healing. Of special notice is the work on the COX-2 knockout mouse, which demonstrates profound abnormalities in bone healing after fracture(6). Though this demonstrates that the presence of COX-2 is necessary for bone healing to occur, it does not answer the question as to whether COX-2 inhibition also has a similar effect on fracture healing. One possibility is that there exists a baseline level of COX-2 necessary and sufficient for fracture healing to occur and that the different dosages of COX-2 inhibitor may influence the different results seen in various studies. However, little is known of the exact role COX-2 plays in the inflammatory and bone regenerating phases of fracture repair. Our study was an attempt to begin to raise and answer some of these issues as well as to stimulate an interest in further research into this important and as yet poorly understood topic.

1. Gerstenfeld L, Thiede M et al.: Differential inhibition of fracture healing by non-specific and cyclooxygenase-2 (COX-2) inhibitors on bone healing. Trans Orthop Res Soc. 2002; 27-65. 2. Mullis B, Copeland S et al.: Effect of COX-2 inhibitors and NSAIDs on fracture healing in a mouse model. Trans Orthop Res Soc. 2002; 27: 712. 3. Long J, Lewis S et al.: The effect of cyclooxygenase-2 inhibitors on spinal fusion. J Bone Joint Surg Am. 2002; 84: 1763-8. 4. Leonelli S, Goldberg B et al.: The effect of cyclooxygenase 2 (COX-2) inhibitors on bone healing. Trans Orthop Res Soc. 2002; 27: 65. 5. Zhang X, Xing L et al.: A critical role of COX-2 in intramembranous and endochondral bone repair. Trans Orthop Res Soc. 2002; 27: 202. 6. Simon A, Manigrasso M, O’Connor J.: Cyclo-oxygenase 2 function is essential for bone fracture healing. J Bone Miner Res. 2002; 17: 963-76. 7. Paulson S, Zhang J et al.: Pharmacokinetics, tissue distribution, metabolism, and excretion of celecoxib in rats. Drug Metabolism and Disposition, 28(5): 514-21, 2000. 8. Sudmann E, Dregelid E et al.: Inhibition of fracture healing by indomethacin in rats. Eur J Clin Invest. 1979; 9: 333-9. 9. Elves M, Bayley I, Roylance P.: The effect of indomethacin upon experimental fractures in the rat. Acta Orthop Scand. 1982; 53: 35-41. 10. Allen H, Wase A, Bear W.: Indomethacin and aspirin: effect of nonsteroidal anti-inflammatory agents on the rate of fracture repair in the rat. Acta Orthop Scand. 1980; 51: 595-600.

To the Editor:

The paper entitled "Effect of Cox-2-Specific Inhibition on Fracture- Healing in the Rat Femur" deals with an important but controversial area of orthopaedic research. However, the paper as published does not give the reader enough information to evaluate the results.

Our main concern deals with the lack of information within the Materials and Methods section. The sample size of the study groups is never stated. It was stated that each treatment group had 19 animals, but the paper did not specify how many animals were sacrificed at each timepoint, how many specimens were used for histological analysis, and how many for mechanical testing. As both of these measures are destructive tests, it is presumed that the same specimen cannot be used for both. Given this lack of information, the reader can only assume that the average sample size is somewhere around 3. Given these small numbers, and the high standard deviations, it is not surprising that few significant differences could be found amongst the treatment groups. One can only presume that the power was very low for these statistics. Therefore, it seems that the conclusions of the paper were too strong – the fact that differences were not detected does not mean that they did not exist.

The authors are to be commended for performing an a priori power analysis to determine suggested group size. However, the reader can clearly see that the assumed coefficients of variation were grossly underestimated, and therefore the sample size suggested by the power analysis would also be underestimated.

In addition, the authors state that “data were grouped and were analyzed with unpaired two-tailed t-tests and were adjusted for multiple comparisons where indicated.” How were the data grouped, and would not an Analysis of Variance with post-hoc comparison have been a more reasonable test for three treatment groups?

Of lesser concern is the delay between harvest of the femora and fixation of the tissue for histological analysis. The authors state that femora were wrapped in saline-soaked cloth and frozen for a “period not exceeding 10 days”. After frozen storage the femora were fixed in a 4% paraformaldehyde solution. In order to ensure preservation of the tissue, the femora for histological analysis should have been fixed immediately upon retrieval. The reason for this delay was never stated.

To the Editor:

In the article ”COX-2-specific inhibition on fracture-healing in the rat femur”, Brown and co-workers conclude that the inhibitory effect on fracture repair may be smaller with selective Cox-2 inhibitors than with unspecific NSAIDs (1). This conclusion may not be valid, because the daily dose of the chosen drug (celecoxib) was the same per kg body wheight as in humans, which is usually insufficient in small animals. Further, in male rats, celecoxib is metabolised and eliminated within a few hours after administration (2). Thus, most of the time, the rats were untreated with celecoxib.This can be expected to inhibit fracture repair (3).

I congratulate the authors on the smart idea of training the rats to like chocolate for oral drug delivery. Perhaps this method could be modified to maintain a reasonable plasma concentration.

Per Aspenberg

1. Brown, K. M.; Saunders, M. M.; Kirsch, T.; Donahue, H. J.; and Reid, J. S.: Effect of COX-2-specific inhibition on fracture-healing in the rat femur. J Bone Joint Surg Am, 86-A(1): 116-23, 2004. 2. Paulson, S. K. et al.: Pharmacokinetics, tissue distribution, metabolism, and excretion of celecoxib in rats. Drug Metab Dispos, 28(5): 514-21, 2000. 3. Simon, A. M.; Manigrasso, M. B.; and O'Connor, J. P.: Cyclo-oxygenase 2 function is essential for bone fracture healing. J Bone Miner Res, 17(6): 963-76, 2002.

To the Editor:

I read with interest the article by Brown et al.(1) regarding the effect of COX-2 specific inhibition on fracture healing in the rat femur. The effect of COX-2 inhibitors on bone healing remains a controversial subject (2-4). Although there is evidence from animal studies to suggest that COX-2 inhibitors inhibit bone fracture healing (5), caution must be exercised when extrapolating these data to clinical practice. In animal studies, the NSAIDs and COX-2 inhibitors were given for several weeks rather than for short-term pain relief.

Additionally, there are no prospective human data indicating a negative effect of COX-2 inhibitors on bone healing. Reuben et al. performed a retrospective analysis of patients undergoing spinal fusion who received either celecoxib, rofecoxib, ketorolac or placebo (6). There was no significant difference in non-union rates between celecoxib, rofecoxib and placebo, while there was a significant difference between all three groups and ketorolac. We await randomized, controlled, prospective trials in humans.

Any adverse effects of COX-2 inhibitors must be weighed against possible benefits, such as improved quality of analgesia, earlier mobilization, and earlier weight bearing. The influence of dose, timing, and duration of administration also requires further study

Noor M.Gajraj Department of Anesthesiology and Pain Management U.T. Southwestern Medical Center Dallas, Texas 75390-9068

References

1. Brown KM, Saunders MM, Kirsch T, Donahue HJ, Reid JS. Effect of COX-2-Specific Inhibition on Fracture-Healing in the Rat Femur. J Bone Joint Surg Am. 2004; 86: 116-123 2. Wedel D, Berry D. "He said, she said, NSAIDs". Reg Anesth Pain Med. 2003; 28: 372-375 3. Hochberg M, Melin J, Reicin A. Cox-2 inhibitors and fracture healing: an argument against such an effect. J Bone Miner Res 2003; 18: 583 4. Gajraj NM.The effect of cyclooxygenase-2 inhibitors on bone healing. Reg Anesth and Pain Med. 2003; 28: 456-465 5. Simon A, Manigrasso M, O'Connor J.Cyclo-oxygenase 2 function is essential for bone fracture healing. J Bone Miner Res. 2002; 17: 963-976 6. Reuben S, Rizvi A, Steinberg R, Maciolek H. The effect of NSAIDs on spinal fusion. American Society of Regional Anesthesia Meeting. 2002. PD- 16.