Heterotopic ossication prophylaxis with indomethacin increases the risk of long-bone nonunion

T. A. Burd, M. S. Hughes, J. O. Anglen

From the University of Missouri, Columbia, USA

ndomethacin is commonly administered for the prophylaxis of heterotopic ossication (HO) after the surgical treatment of acetabular fractures. Nonsteroidal anti-inammatory drugs such as indomethacin, have been associated with delayed healing of fractures and mechanically weaker callus. Our aim was to determine if patients with an acetabular fracture, who received indomethacin for prophylaxis against HO, were at risk of delayed healing or nonunion of any associated fractures of long bones. We reviewed 282 patients who had had open reduction and internal xation of an acetabular fracture. Patients at risk of HO were randomised to receive either radiation therapy (XRT) or indomethacin. Of these patients, 112 had sustained at least one concomitant fracture of a long bone; 36 needed no prophylaxis, 38 received focal radiation and 38 received indomethacin. Fifteen patients developed 16 nonunions. When comparing patients who received indomethacin with those who did not, a signicant difference was noted in the rate of nonunion (26% v 7%; p = 0.004). Patients with concurrent fractures of the acetabulum and long bones who receive indomethacin have a signicantly greater risk of nonunion of the fractures of the long bones when compared with those who receive XRT or no prophylaxis.
J Bone Joint Surg [Br] 2003;85-B:700-5. Received 25 October 2002; Accepted after revision 14 February 2003

The most common orthopaedic injury which is associated with acetabular fractures, is fracture of a long bone.1 One of the more frequent late complications which can occur from surgery for an acetabular fracture is heterotopic ossication (HO).2-6 Most trauma surgeons agree that some form of prophylaxis against the formation of heterotopic bone should be used perioperatively for patients who undergo posterior or extensile surgical approaches to the acetabulum, which are associated with high rates of HO.2,7,8 Prophylaxis usually entails either systemic indomethacin for six weeks or a single dose of local radiation to the soft tissues around the acetabulum,9-12 both of which have been shown signicantly to decrease the incidence of the formation of heterotopic bone after open reduction and internal xation (ORIF) of acetabular fractures.7,9-12 Both treatments involve some small risks which have been well described. These include gastrointestinal symptoms and haemorrhage with indomethacin and wound-healing problems, malignant transformation and genetic alteration with radiation. In addition, in multiple animal models and clinical studies, indomethacin and other non-steroidal anti-inammatory drugs (NSAIDs) have been shown to impair healing of fractures with reduced callus and altered remodelling of the Haversian systems.13-27 Our aim was to determine if indomethacin, when used for prophylaxis for HO after surgery for acetabular fracture, increases the risk of nonunion in patients with associated fracture of a long bone.

Patients and Methods

Fractures of the acetabulum often result from high-energy trauma, most commonly a motor-vehicle collision. Additional musculoskeletal injuries often occur in these patients. Between July 1992 and January 2001, 282 patients underwent ORIF for an acetabular fracture by a single surgeon (JOA). During that period, patients at risk for HO because of the use of a posterior or extensile surgical approach were given prophylaxis under an investigational protocol, which randomised them to treatment with either indomethacin or radiation.12 Those who were randomised to indomethacin received 25 mg three times daily for six weeks and those randomised to radiation received a single dose of 800 centigray of external beam irradiation within 72 hours of surgery. There was no verication of indomethacin compliance and no specic instructions about the use of over-the-counter nonsteroidal medication were given or monitoring of such use performed.
THE JOURNAL OF BONE AND JOINT SURGERY

T. A. Burd, MD, Chief Resident in Orthopaedics M. S. Hughes, BS, Medical Student J. O. Anglen, MD, Clinical Professor, Orthopaedic Surgery Department of Orthopaedic Surgery The University of Missouri Hospital and Clinics, MC213 Columbia, Missouri 65212, USA. Correspondence should be sent to Professor J. O. Anglen. 2003 British Editorial Society of Bone and Joint Surgery doi:10.1302/0301-620X.85B5.13970 $2.00
700

HETEROTOPIC OSSIFICATION PROPHYLAXIS WITH INDOMETHACIN INCREASES THE RISK OF LONG-BONE NONUNION

701

Table I.

Patients with concurrent fractures of the acetabulum and long bones Number of Number of long- Number of patients bone fractures nonunions 118 72 190 5 11 16 Patients with a nonunion (%) 7 29 13 Fractures with a nonunion (%) 4.2 15.3 8.4

Treatment group

No indomethacin 74 Indomethacin 38 Total 112

The initial review of the 282 patients revealed 112 (40%) with concomitant fractures of the femur, tibia, humerus, and/or forearm (radius and/or ulna). These patients formed the population of this study. All fractures of long bones were managed by the same surgeon who treated the acetabular fracture, using the same techniques, approaches and implants as any other patient in the practice, without reference to, or knowledge of the randomisation group. Approval was obtained from the Institutional Review Board to review the medical records and radiographs. Clinical data obtained included age, smoking habits, fracture status (open or closed), and the Injury Severity Score (ISS). Three groups were identied: (i) those who received no prophylactic treatment because of the use of a low risk2 anterior surgical approach (n = 36); (ii) those who received radiation treatment (n = 38), and (iii) those who received indomethacin (n = 38) (Table I). We dened a nonunion as any fracture which required further treatment for failed healing. The indication for further surgery was failure of clinical or radiographic healing after a minimum of three months, or obviously failed osteosynthesis. This decision was made by the treating surgeon without awareness of the method of prophylaxis administered, or that this would be the topic of a later study. Nonunion was classied as atrophic, hypertrophic or oligotrophic.28 The records and radiographs were reviewed by observers who were not involved in the care of the patients or in determining the method of prophylaxis or in the decision as to whether or not the fracture should be considered as a nonunion. Statistical analysis. This was performed using SAS Software (Cary, North Carolina) with the level of signicance set at p < 0.05. The condence interval (CI) and odds ratio were determined with the CI set at 95%. Chi-squared analysis was used to compare the incidence of nonunion between the groups. The nal step was to perform an analysis which took into account those patients with multiple fractured long bones and, consequently, not all observations on nonunion were independent. We therefore used the Genmod procedure in SAS for the analysis. The correlation structure was modelled using the compound symmetry model.

Results
A total of 190 long-bone fractures occurred in the 112 patients (mean 1.7 fractures/patient). Sixteen fractures which required additional treatment for nonunion were
VOL. 85-B, No. 5, JULY 2003

identied in 15 patients. The mean time at which a nonunion was diagnosed was 5.3 months (3 to 11) after injury. The patients with nonunion had 28 associated fractures of a long bone with a mean of 1.9 fractures/patient. This mean was not statistically signicant from the whole group (p = 0.656). Comparison between no prophylaxis and radiation. Patients who had had an ilioinguinal approach for acetabular ORIF did not receive any form of prophylaxis.2 In this group, two of 36 (6%) developed nonunion of a fracture of a long bone compared with 3 of 38 (8%) in the radiation group. No statistical difference was seen between these two groups (p = 0.708), which for simplication, were then combined and compared with the patients who received indomethacin. Comparison between the use and non-use of indomethacin. Patients who received indomethacin had an incidence of 29% (11 of 38 patients) of nonunion of a fracture of a long bone. This was signicantly different from the incidence of 7% (5 of 74 patients) of nonunion seen in the group which did not receive indomethacin (p = 0.004). There were no cases of acetabular nonunion. In terms of the total number of fractures of a long bone which were at risk of complications of healing, there were ve cases of nonunion out of a total of 118 fractures (4.2%) in patients who did not receive indomethacin and eleven of 72 fractures (15.3%) in patients who did receive indomethacin. This difference is also signicant (p = 0.029) Comparing those patients who received indomethacin with those who did not, the odds ratio for developing a nonunion of a fracture of a long bone was 5.32 to 1. Separate comparison of the indomethacin group with the no treatment group, and with the radiation group, yielded odds ratios of 6.9 to 1 and 4.3 to 1, respectively. Demographic multivariate analysis. There were no signicant differences between the three groups with regard to age, use of alcohol, smoking habits and fracture status. The only signicant difference was the mean ISS (p = 0.0056) (Table II). The patients in the indomethacin group had an ISS which was approximately four points higher than that for the other patients. The reason for this difference is unknown but the magnitude is low. Even adjusting for the ISS, there is still a signicant group effect (p = 0.042). No signicant relationship between the rate of nonunion and the ve covariates listed was identied. There was no significant difference between the treatment groups in the number or distribution of fractures of a long bone, or in the distribution of fractures which were either ipsilateral or contralat-

702

T. A. BURD, M. S. HUGHES, J. O. ANGLEN

Table II.

Details of the patients and multivariate analysis Number of Number of patients who Mean age patients who consumed Open in years smoked (%) alcohol (%) fractures (%) Mean ISS 18 (24) 12 (32) 30 (27) 0.25 16 (22) 5 (13) 21 (19) 0.44 16 (22) 13 (34) 29 (26) 0.23 14.5 19 16 0.0056*

Group

No indomethacin (n = 74) 38.1 Indomethacin (n = 38) 39.5 All patients (n = 112) 38.6 p value 0.21* *ANOVA and Kruskal-Wallis test Fishers exact test

Table III. Distribution of fractures of a long bone by treatment group. There was no difference between the groups in the number or location of associated long-bone fractures, nor in whether the associated lower extremity fractures were on the ipsilateral or contralateral side to the acetabular fracture No indomethacin No treatment Radiation Indomethacin Humerus Proximal 1 Middle 3 Distal 1 Total 5 Forearm Proximal 1 Middle 4 Distal 11 Total 16 Femur Proximal 5 Middle 13 Distal 5 Total 23 Tibia/bula Proximal 5 Middle 6 Distal 7 Total 18 Number of patients with associ- 29 ated fractures of a lower limb (%) Total Relationship to acetabular frac- 16/7/6 ture ipsilateral/contra-lateral/ both sides Total 1 5 2 8 2 2 5 9 7 6 0 13 5 6 15 26 27 56 (76) 17/5/5 2 2 2 6 1 7 5 15 5 8 7 20 12 9 10 31

Table IV. Severity of open fractures in each group according to Gustilo and Anderson33 Grade Closed 1 No indomethacin 93 Indomethacin 62 *Pearson chi-squared test 2 1 2 4 7 3 19 2 p value* 0.014

shoulder, and hemiarthroplasty of the hip, respectively, as treatment for their nonunion. The time to healing of the nonunion was therefore not recorded for these three patients. Two patients (cases 10 and 14) were lost to follow-up before healing of the nonunion.

Discussion
NSAIDs are the most commonly prescribed class of medication throughout the world.28 This group of drugs has been found to inhibit prostaglandin-mediated bone remodelling in vitro and in animal models in a dose-dependent fashion.19 Numerous reports have shown their interference with the healing of bone in both experimental and human fractures,13,15-18,22,23,25 which may be due to inhibition of cyclo-oxygenase 2 activity,29 and this may also be expected to occur with the newest anti-inammatory drugs, the COX2 inhibitors. Both indomethacin and radiation therapy administered perioperatively have been shown to be effective in reducing the formation of heterotopic bone after surgery for an acetabular fracture.7,10,12 Stannard and Alonso8 surveyed 14 senior trauma surgeons in order to identify the best method of prevention of HO and found mixed results, thus reecting the lack of consensus on prophylaxis. Most (50%) believed that indomethacin was the best treatment in posterior and extensile approaches to the acetabulum. Numerous studies have veried that indomethacin is effective in preventing HO after extensile approaches to repair acetabular fractures.4,10,30,31 In a recently published, prospective, randomised study of 166 patients,12 we showed equivalent results for the efcacy of indomethacin and radiation therapy in prophylaxis for HO after ORIF of acetabular fractures. Indomethacin was found to be well tolerated, approximately 200 times less expensive and as effective as radiation. Our study appeared to favour indomethacin, proTHE JOURNAL OF BONE AND JOINT SURGERY

31 (75)

33/12/11

17/8/6

eral to the acetabular fracture. Table III shows the distribution of fractures of a long bone in each treatment group. There was a higher proportion of open fractures in the group which did not receive indomethacin (25 of 118, 21%) than in the group which did (10 of 72, 14%), although this difference was not statistically signicant (Fishers exact test p = 0.25). There was a statistically signicant difference in the severity of the open fractures, with the indomethacin group having fewer severe fractures. The distribution is shown in Table IV. Table V gives details of the patients with a nonunion and the treatment required. Three of the patients (cases 2, 8 and 13) underwent total hip arthroplasty, hemiarthroplasty of the

Table V. Details of patients with a nonunion and required further treatments

Case 1 N N N N 1 N 1 N N 1 2 L-femur 31B2.1+32C2.1 31-A+32A3.2 31-A3.2 31AC A2.3+shaft ext 12-C3.2 C C R L C L C R Oligo 1 R-tibia 43-C3.3 C L Atrophic 2 L-tibia 43-C3.3 C L Hyper 4 L-tibia 41-C2.3 C L Oligo ORIF, BS ORIF, ICBG, IBS XN,ICBG, IBS ORIF, OS, IBS Oligo HWR, ICBG, ORIF, Hemi Atrophic Oligo HWR, ORIF, ICBG, IBS OS, ORIF, IBS L Atrophic HWR, ORIF, ICBG, IBS 1 R-humerus 11-A3.2 C R Atrophic HWR, ICBG, Hemi 3 R-tibia 42-C3.1 3B R Atrophic XN, IBS, ICBG, BS 3 11 3 6 3 8 2 L-femur, R- L-32C3.2; C femur R-32B2.2 L Oligo, oligo RORIF, BS & L XN, BS 4.4 3 L-ulna 22-A1.2 1 L Oligo ORIF, ICBG, IBS 8 30 6.6 6 N/A 3 6 15 10 2 L-femur 32-C1.3 31-B2.1 C R Oligo THA 3 N/A 1 L-ulna 21-B1.1 C L Oligo XP 3 UNK 2 L-femur 32-C3.2 C R Atrophic XN, ICBG 7 6 13 2 44 39 14 9 24 24 60 18 18

Age (yrs)

ISS

Prophylaxis for HO* Nonunion location

Smoking habit (PPD)

Number of long-bone fractures

AO/OTA fracture class

Side of Open acetabular Type of (grade) fracture nonunion

Treatment of nonunion

Time from injury to diagnosis of nonunion (mths) Nonunion healing Follow-up time (mths) (mths)

VOL. 85-B, No. 5, JULY 2003

20

10

24

18

59

22

48

10

45

37

27

53

18

34

66

10

18

10

11

38

Rad

12

65

Rad

L-femur

N/A

15

13 N 3 R-femur

38

10

Rad

R-femur

4 6

UNK 15

10 21

14

39

14

No TX

15 2

50

10

No TX

L-humerus

6 5

8 10

14 22

HETEROTOPIC OSSIFICATION PROPHYLAXIS WITH INDOMETHACIN INCREASES THE RISK OF LONG-BONE NONUNION

Mean

42

13

*HO, heterotopic ossication; I, indomethacin; Rad, radiation; No TX, no treatment Gustilo and Anderson classication33 BS, external bone simulator; Hemi, hemiarthroplasty; HWR, hardware removal; IBS, implantable bone stimulator; ICBG, autogenous bone grafting; ORIF, open reduction and internal xation; OS, osteotomy; THA, total hip arthroplasty; XN, exchange nailing; XP, exchange plating Abbreviations: AO/OTA, Arbeitgemeinshaft fr Osteosynthesefragen/Orthopaedic Trauma Association;34 ISS, injury severity score; Oligo, oligotrophic; n/a, not available; PPD, packs per day; R/L, right/left; UNK, unknown

703

704

T. A. BURD, M. S. HUGHES, J. O. ANGLEN

vided that there were no contraindications from gastrointestinal or renal comorbidities. We did not assess the potentially deleterious effects that this potent NSAID may have on healing of bone when taken for prolonged periods, which was the purpose of our re-evaluation of these patients. There are limitations to our study. The analysis was performed retrospectively, although the study data were collected prospectively as part of a randomised investigation. Nonunion was determined at the time of clinical follow-up and not retrospectively. At that time, the senior author (JOA) was unaware of the patients treatment group or that the issue would subsequently be the topic of a study. We did not collect details about the initial treatment of the fractures of a long bone, a feature which could affect rates of nonunion, but they were all treated by a single surgeon who used consistent methods and philosophies. We did not control for, or collect information about, the use of NSAIDs for control of pain or other purposes. The denition of a nonunion can be controversial and arbitrary. We chose to include only those fractures which, in the judgement of the senior author, required surgical treatment for complications of healing. This determination was made before it became a topic of a study. It constitutes an easily identied event in the medical records and is a denition which has a signicant impact on a patients care. Although the groups were randomised, there is the possibility that they differed in some way. The patients who received indomethacin had a slightly higher ISS and a less severe spectrum of open fractures than those who did not, neither of which would account for the increase in the rate of nonunion which was observed. We found no signicant differences between the groups in age, smoking, alcohol use, associated number or location of fractures, or the rate of open fracture. An intriguing possible future solution to this problem was recently reported by Martin, Boden and Titus,32 who used recombinant human bone morphogenic protein-2 (rhBMP-2) to compensate for the detrimental effects of ketorolac on the formation of bone in a rabbit spine fusion model. Further research may indicate that if indomethacin is necessary for prophylaxis for HO after ORIF for acetabular fracture in patients with additional fractures of a long bone, fracture site rescue with rhBMP-2 may lessen the risk of nonunion. While our study included only patients with an acetabular fracture with associated fractures who were also treated with indomethacin for six weeks, we believe, based on the literature,13,15-19,22,23,25 that the results can be generalised to other patients with fractures who may use other NSAIDs for other reasons. Unfortunately, the newer COX2 inhibitors would seem to suffer from the same drawback.29 We recommend the avoidance of NSAIDs for analgesia or anti-inammatory purposes during healing of the fracture. HO can be a signicant complication after extensile and posterior approaches to the acetabulum. Most surgeons

believe that radiation therapy and/or the use of indomethacin is an important prophylactic measure against HO.8 Our earlier study12 showed indomethacin to be safe, as effective, and dramatically cheaper than radiation therapy. However, the data from this current study strongly suggest that in patients with acetabular fractures and concomitant fractures of a long bone, those who receive indomethacin for heterotopic ossication are at greater risk for nonunion of a fracture of a long bone. In this situation, the surgeon may wish to use radiation.
We wish to thank Dr Richard Madsen, Professor of Biostatistics, and James L. Cook, PhD, DVM, for their contributions to the statistical analysis of this project. No benets in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article.

References
1. DiPasquale TG, Nowinkski RJ. The acute care and evaluation of acetabular fractures. In: Kellam JF, Fischer TJ, Tornetta III P, Bosse MJ, Harris MB, eds. Orthopaedic knowledge update, Trauma 2. Rosemont, Illinois: American Academy of Orthopaedic Surgeons, 2000, p. 239-253. 2. Letournel E, Judet R. Fractures of the acetabulum. 2nd ed. New York: Springer-Verlag, 1993. 3. Matta JM. Fractures of the acetabulum: accuracy of reduction and clinical results in patients managed operatively within three weeks after the injury. J Bone Joint Surg [Am] 1996;78-A:1632-45. 4. McLaren AC. Prophylaxis with indomethacin for heterotopic bone after open reduction of fractures of the acetabulum. J Bone Joint Surg [Am] 1990;72-A:245-7. 5. Bosse MJ, Poka A, Reinert CM, et al. Heterotopic ossication as a complication of acetabular fracture: prophylaxis with low-dose irradiation. J Bone Joint Surg [Am] 1988;70-A:1231-7. 6. Reinert CM, Bosse MJ, Poka A, et al. A modied extensile exposure for the treatment of complex or malunited acetabular fractures. J Bone Joint Surg [Am] 1988;70-A:329-37. 7. Johnson EE, Kay RM, Dovey FJ. Heterotopic ossication prophylaxis following operative treatment of acetabular fractures. Clin Orthop 1994;305:88-95. 8. Stannard JP, Alonso JE. Controversies in acetabular fractures. Clin Orthop 1998;353:74-80. 9. Moore KD, Goss K, Anglen JO. Indomethacin versus radiation therapy for prophylaxis against heterotopic ossication in acetabular fractures: a randomized prospective study. J Bone Joint Surg [Br] 1998;80-B:259-63. 10. Moed BR, Maxey JW. The effect of indomethacin on heterotopic ossication following acetabular fracture surgery. J Orthop Trauma 1993;7:33-8. 11. Slawson RG, Poka A, Bathon H, et al. The role of postoperative irradiation in the prevention of heterotopic ossication in patients with post-traumatic acetabular fracture. Int J Radiat Oncol Biol Phys 1980;17:669-72. 12. Burd TA, Lowry KJ, Anglen JO. Indomethacin compared with localized irradiation for the prevention of heterotopic ossication following surgical treatment of acetabular fractures. J Bone Joint Surg [Am] 2001;83-A:1783-8. 13. Ro J, Sudmann E, Marton PF. Effect of indomethacin on fracture healing in rats. Acta Orthop Scand 1976;47:588-99. 14. Dimar JR, Ante WA, Zhang YP, Glassman SD. The effects of nonsteroidal anti-inammatory drugs on posterior spinal fusions in the rat. Spine 1996;21:1870-6. 15. Engesaeter LB, Sudmann B, Sudmann E. Fracture healing in rats inhibited by locally administered indomethacin. Acta Orthop Scand 1992;63:330-3. 16. Elves MW, Bayley I, Roylance PJ. The effect of indomethacin upon experimental fractures in the rat. Acta Orthop Scand 1982;53:35-41. 17. Allen HL, Wase A, Bear WT. Indomethacin and aspirin: effect of nonsteroidal anti-inammatory agents on the rate of fracture repair in the rat. Acta Orthop Scand 1980;51:595-600. 18. Altman RD, Latta LL, Keer R, et al. Effect of nonsteroidal antiinammatory drugs on fracture healing: a laboratory study in rats. J Orthop Trauma 1995;9:392-400.
THE JOURNAL OF BONE AND JOINT SURGERY

HETEROTOPIC OSSIFICATION PROPHYLAXIS WITH INDOMETHACIN INCREASES THE RISK OF LONG-BONE NONUNION

705

19. Glassman SD, Rose SM, Dimar JR, et al. The effect of postoperative nonsteroidal anti-inammatory drug administration on spinal fusion. Spine 1998;23:834-8. 20. Wood MK, Li G, Simpson AHRW. The effects of nicotine and NSAIDs on new bone formation during leg lengthening. J Bone Joint Surg [Br] 1998;80-B, Suppl III:264. 21. Giannoudis PV, MacDonald DA, Matthews SJ, et al. Nonunion of the femoral diaphysis: the inuence of reaming and non-steroidal antiinammatory drugs. J Bone Joint Surg [Br] 2000;82-B:655-8. 22. Hogevold HE, Grogaard B, Reikeras O. Effects of short-term treatment with corticosteroids and indomethacin on bone healing: a mechanical study of osteotomies in rats. Acta Orthop Scand 1992;63:607-11. 23. Dickson G, Connolly C, Marsh D. NSAIDsprostaglandin ratios and fracture repair. J Orthop Trauma 1999;13:322-3. 24. Sodemann B, Persson PE, Nilsson OS. Prevention of periarticular heterotopic ossication following total hip arthroplasty: clinical experience with indomethacin and ibuprofen. Arch Orthop Trauma Surg 1988;107:329-33. 25. Keller J Bunger C, Andreassen TT, Bak B, Lucht U. Bone repair inhibited by indomethacin-effects on bone metabolism and strength of rabbit osteotomies. Acta Orthop Scand 1987;58:379-83. 26. Rosen H. Nonunion and malunion. In: Skeletal trauma, fractures, dislocations, ligamentous injuries. 2nd ed. WB Saunders, Philadelphia, PA, 1998:619-660.

27. Weber BG, Brunner C. The treatment of nonunions without electrical stimulation. Clin Orthop 1981;161:24-32. 28. Baum C, Kennedy DL, Forbes MB. Utilization of nonsteroidal antiinammatory drugs. Arthritis Rheum 1985;28:686-92. 29. Simon AM, Manigrasso MB, OConnor JP. Cyclo-oxygenase 2 function is essential for bone fracture healing. J Bone Miner Res 2002;17:963-76. 30. Daum WJ, Scarborough MT, Gordon W Jr, Uchida T. Heterotopic ossication and other perioperative complications of acetabular fractures. J Orthop Trauma 1992;6:427-32. 31. Ghalambor N, Matta JM, Bernstein L. Heterotopic ossication following operative treatment of acetabular fractures: an analysis of risk factors. Clin Orthop 1994;305:96-105. 32. Martin GJ, Boden SD, Titus L. Recombinant human bone morphogenic protein-2 overcomes the inhibitory effect of ketorolac, a nonsteroidal anti-inammatory drug (NSAID), on posterolateral lumbar intertransverse process spine fusion. Spine 1999;24:2188-94. 33. Gustilo RB, Anderson JT. Prevention of infection in the treatment of one thousand and twenty ve open fractures of the long bones: retrospective and prospective analysis. J Bone Joint Surg [Am] 1976;58A:453-8. 34. AO/ATA. OTA Committee for ceding and classication: fracture and dislocation compendium. J Orthop Trauma 1996;10:Suppl I.

VOL. 85-B, No. 5, JULY 2003