Infection is a common, serious complication related to insertion of fracture-fixation devices. Although potentially beneficial, the administration of systemic antimicrobial prophylaxis and compliance with the usual measures to control infection have not completely eradicated infections associated with the use of these devices3,11. Infection is of particular concern after the fixation of open fractures18 and has been reported to develop in as many as 14 per cent of seventy-nine patients9,13. An infection associated with a fracture-fixation device is generally difficult to eradicate without removal of the device and is expensive to treat. For example, findings from the Dutch Trauma Trial indicated that it cost a mean of $22,000 (in 1991 dollars) to treat one patient who had a deep-wound infection after internal fixation of a closed fracture3. Because of the difficulty and expense associated with the treatment of orthopaedic device-related infections, there has been great interest in the role of various technologies in their prevention.
Although infection with low-virulence organisms, such as Staphylococcus epidermidis, may not manifest clinically until months or years after insertion of a fracture-fixation device, it is generally thought that most of the pathogens that cause infection are introduced at the time of the injury or the operative treatment or within a few weeks after it11. Colonization of the surface of the device is a prerequisite for the development of an infection related to an orthopaedic implant, as it is for infections related to other types of medical implants. Therefore, we hypothesized that coating the surface of fracture-fixation devices with effective antimicrobial agents would reduce the rate of bacterial colonization and provide significant protection against clinical infection.
The primary objective of the present study was to examine the efficacy of intramedullary nails coated with the antiseptic combination of chlorhexidine and chloroxylenol in reducing the rate of infection with Staphylococcus aureus after the fixation of open tibial fractures in rabbits.
*No benefits in any form have been received or will be received from a commercial party related directly or indirectly to the subject of this article. Funds were received in total or partial support of the research or clinical study presented in this article. The funding source was Zimmer Incorporated, Warsaw, Indiana.
†Read in part at the Annual Meeting of the Orthopaedic Research Society, San Francisco, California, February 13, 1997.
‡Department of Medicine, Infectious Disease Section, Room 4B-370 (R. O. D. and M. M.), Department of Orthopaedic Surgery (J. F., C. C., and G. C. L.), and Department of Pathology (G. S.), Baylor College of Medicine and Veterans Affairs Medical Center, 2002 Holcombe Boulevard, Houston, Texas 77030.
Orthopaedic Devices
Stainless-steel 2.8-by-100-millimeter Kirschner wires (Zimmer, Warsaw, Indiana) were used as intramedullary nails for the fixation of open tibial fractures in rabbits. Before the operation, the rabbits were randomized to receive either a regular uncoated intramedullary nail or an intramedullary nail that had been coated with an antiseptic combination of chlorhexidine and chloroxylenol. A proprietary dipping technique (intellectual property of Baylor College of Medicine, Houston, Texas) was used to bind the antiseptic agents to the surface of the metallic devices. The mean concentrations of chlorhexidine and chloroxylenol bound to the surface of the coated nails, as determined with high-performance liquid chromatography, were 3.10 and 3.15 milligrams per centimeter, respectively. All of the intramedullary nails were packaged and then were sterilized with ethylene oxide gas before they were inserted in the animals.
Rabbits
Forty-four New Zealand White, specific-pathogen-free rabbits (body mass, four to five kilograms) were anesthetized with an intramuscular injection (0.5 milliliter per kilogram of body weight) of ketamine (42.8 milligrams per milliliter) and xylazine (0.7 milligram per milliliter). In order to simulate perioperative antibiotic prophylaxis in humans, immediately after induction of the anesthesia each animal received an intramuscular injection of cefazolin (eighty milligrams per kilogram of body weight) to which the infecting strain of Staphylococcus aureus was demonstrated to be susceptible (minimum inhibitory concentration of cefazolin, one microgram per milliliter).
Tibial Fracture
We used a previously described rabbit model of open tibial fracture22, with some modifications. The right leg of the rabbit was shaved from the knee to the ankle and was draped and prepared in a sterile fashion with use of povidone-iodine skin cleanser. A two-centimeter-long incision was made on the medial side of the knee, and a paramedian arthrotomy was performed by incision of the joint capsule. After the fat pad had been mobilized and the medial tibial plateau had been exposed, the arthrotomy incision was covered with gauze. Another two-centimeter-long incision was then made at the level of the tibial tubercle and was extended distally along the lateral border of the tibia. After the overlying fascia had been incised and the tibialis anterior muscle had been retracted laterally, the periosteum was incised and was elevated approximately 0.5 centimeter distal to the tibial tubercle. The intramedullary canal was then entered at the anteromedial aspect of the tibial plateau and was reamed along its entire length with use of, in succession, 2.0-millimeter, 2.4-millimeter, and 2.8-millimeter stainless-steel Kirschner wires attached to a handset drill (Stryker Surgical, Kalamazoo, Michigan). After the canal had been reamed, the nail was placed in the canal and was advanced one centimeter distally. The tibia was fractured 0.5 centimeter distal to the tibial tubercle with a handset sagittal saw and a 0.5-centimeter saw blade (Stryker Surgical). Rotational stability was maintained by leaving the fibula intact and fracturing the tibia proximal to the tibiofibular syndesmosis, which occurs at the junction of the proximal and middle thirds of the tibia. A one-milliliter tuberculin syringe was placed through the incision adjacent to the tibial tubercle and into the fracture with the aperture pointing distally within the intramedullary canal. The fracture was then reduced by advancing the intramedullary nail toward the ankle joint.
Bacterial Inoculation
We used the same strain of Staphylococcus aureus Newman that we examined in another rabbit model of orthopaedic device-related infection8. After a pilot trial to compare the ability of different inocula of the strain of Staphylococcus aureus Newman to cause infection in this rabbit model, we concluded that a bacterial inoculum of 106 colony-forming units of Staphylococcus aureus was optimum because it produced infection in six of eight rabbits that had received an uncoated intramedullary nail for the fixation of an open tibial fracture.
After all instruments except those necessary to close the incisions had been cleared from the sterile field (to prevent bacterial contamination), 0.1 milliliter of 107 colony-forming units per milliliter of the strain of Staphylococcus aureus Newman was injected through the tuberculin syringe into the intramedullary canal adjacent to the fracture-fixation device. The tuberculin syringe was withdrawn, the incision at the site of the fracture was closed in a single layer with 3-0 Dexon (polyglycolic acid) suture, and the arthrotomy incision and joint capsule were closed in two layers with 3-0 Dexon suture.
Monitoring of the Rabbits
Anteroposterior and lateral radiographs of the right tibia were made postoperatively before the rabbits woke from the anesthesia. The rabbits were monitored daily for as long as six weeks, with particular attention given to the appearance of the wound, the mobility of the rabbit, signs of infection, and the rabbit's ability to thrive. The rabbits were killed by intracardiac injection of phenobarbital at six weeks. Nine rabbits that had received an uncoated nail and eight that had received an antiseptic-coated nail were killed earlier because of dehiscence of the wound, gross instability of the fracture, signs of infection, or failure of the animal to thrive.
Microbiological Evaluation
Specimens were collected with swabs post mortem from the incision at the site of the fracture, from any soft-tissue collections, from the tracks between the nail and the bone at the site of the tibial fracture and at the site of the bacterial injection, and from the nail. The specimens were collected in a sterile fashion and were cultured by inoculation onto chocolate blood agar plates (BBL Media; Becton Dickinson Microbiology Systems, Cockeysville, Maryland). In addition, the nail was cultured with the roll-plate technique, and quantitative cultures of ten-microliter aliquots of blood (obtained from the heart of the rabbit) and of original samples and serial dilutions of homogenized bone tissue were performed. The tibia was split longitudinally into two halves: one half was homogenized in phosphate-buffered saline solution at 4 degrees Celsius, and the other half was submitted for histopathological examination, as will be described.
Histopathological Examination
Coded bone specimens were fixed in 10 per cent formalin for at least two days and were then decalcified for two to three days. Sections parallel to the long axis of the tibia were processed in a standard fashion by dehydration into paraffin until the sections were thoroughly permeated. Specimens of bone embedded in paraffin blocks were cut into five-micrometer-thick sections, mounted on glass slides, and stained with hematoxylin and eosin. Sections of the coded bone specimens were examined by one pathologist (G. S.), who was blinded with regard to the type of intramedullary nail that had been used and the results of the bacterial cultures. The presence of inflammation, fibrosis, fat necrosis, necrotic debris, and woven bone was assessed with an ordinal scale of 0 to 3; 0 indicated absent, 1 indicated slight, 2 indicated moderate, and 3 indicated severe.
Radiographic Evaluation
Anteroposterior and lateral radiographs of the right tibia were made when the rabbits were killed. The radiographs were coded and evaluated in a blinded fashion by one orthopaedic surgeon (G. C. L.), who graded the extent of osseous union, the stability of the bone, and osteolysis. Minimum osseous union was given 1 point; moderate union, 2 points; and complete union, 3 points. Gross instability was assigned 1 point; some instability, 2 points; and stability, 3 points. Severe osteolysis, with or without fracture of the bone, was given 1 point; some osteolysis, 2 points; and no osteolysis, 3 points. An overall radiographic score (possible range, 3 to 9 points) was then calculated for each rabbit.
In Vitro Antimicrobial Activity
The in vitro antimicrobial activity of the antiseptic-coated intramedullary nails was assessed by determination of the zone of inhibition of Staphylococcus aureus, with use of a modified Kirby-Bauer technique15, both before insertion and after removal of the coated nails from the rabbits. The strain of Staphylococcus aureus Newman was grown at 37 degrees Celsius for eighteen hours in trypticase soy broth (Becton Dickinson Microbiology Systems) to a concentration of 0.5 McFarland units (108 colony-forming units per milliliter). A cotton swab was dipped in the bacterial suspension and was then rubbed in a streaking fashion across the whole surface of a Mueller-Hinton agar plate (Becton Dickinson Microbiology Systems). Two-centimeter segments of the antiseptic-coated intramedullary nails were tightly embedded in a groove created by the removal of agar from the center of the agar plate that had been freshly inoculated with Staphylococcus aureus. After the agar plate had been incubated at 37 degrees Celsius for twenty-four hours, the zone of inhibition was determined by measuring the diameter of the clear zone perpendicular to the long axis of the intramedullary nail.
Analysis of the Data
Isolation of the same infecting strain of Staphylococcus aureus, as confirmed with antibiotic susceptibility assay, from cultures of both the nail and the bone indicated the presence of orthopaedic device-related osteomyelitis, irrespective of the concentration of the isolated organisms. A two-tailed Fisher exact test was used to compare the rates of infection between the two groups of rabbits. The Wilcoxon rank-sum test was used to compare the radiographic scores of the rabbits that had received an antiseptic-coated nail with those of the rabbits that had received an uncoated nail.
Rates of Infection
Of the forty-four rabbits in which internal fixation of the tibial fracture was successful, one died within a day after the operation. Cultures were performed on specimens obtained post mortem from the remaining forty-three rabbits (twenty-one that had received an uncoated nail and twenty-two that had received an antiseptic-coated nail), and the results were included in the analysis.
The two groups of rabbits were well matched with regard to the duration for which the nail had been implanted (mean, 29.2 days; median, thirty-nine days; and range, eight to forty-three days for the uncoated nails and mean, 32.5 days; median, forty-one days; and range, five to forty-two days for the antiseptic-coated nails).
Device-related osteomyelitis was diagnosed, on the basis of growth of the infecting strain of Staphylococcus aureus in both the roll-plate cultures of the nails and the quantitative cultures of the homogenized bone tissue, in thirteen (62 per cent) of the twenty-one rabbits that had received an uncoated nail compared with two (9 per cent) of the twenty-two rabbits that had received an antiseptic-coated nail. The use of antiseptic-coated intramedullary nails was associated with a significant reduction in the rate of device-related osteomyelitis (relative risk, 6.81; p = 0.0003; 95 per cent confidence limits, 1.74 to 26.62). Staphylococcus aureus also grew on culture of specimens obtained with swabs from the track between the nail and the bone or from the site of the incision and the soft tissues, or both, in all fifteen rabbits in which osteomyelitis was documented.
Staphylococcus aureus grew on culture of specimens from the incision at the site of the fracture, from the soft tissue, and from the track between the nail and the bone in one rabbit that had been killed, because of dehiscence of the wound, eight days after implantation of an uncoated nail. However, this rabbit was not considered to have had device-related osteomyelitis because cultures of the nail and the bone tissue were sterile.
Cultures of the blood specimens from all forty-three rabbits were sterile.
Histopathological Examination
The histopathological findings were generally similar in the two groups of rabbits. For example, eight rabbits in each group had grade-1 or 2 inflammatory changes. There were no obvious allergic or toxic reactions to the coating materials. Foreign-body giant cells or granulomas, or both, occurred with the same frequency in the two groups (in six of the twenty-one rabbits that had received an uncoated nail and in seven of the twenty-two rabbits that had received a coated nail). Not unexpectedly, staining of the bone samples with hematoxylin and eosin demonstrated bacteria in only one of the fifteen rabbits that had device-related osteomyelitis. This rabbit had received an uncoated nail.
Radiographic Evaluation
The mean radiographic score was 6.9 points (range, 4 to 9 points) for the rabbits that had received an uncoated nail and 7.2 points (range, 4 to 9 points) for the rabbits that had received an antiseptic-coated nail. The median score was 7 points for both groups.
In Vitro Antimicrobial Activity
As expected, there was an inverse relationship between the mean zone of inhibition produced by the antiseptic-coated nails against the strain of Staphylococcus aureus Newman and the duration that the nail had been implanted. The mean zone of inhibition was fifteen millimeters before the intramedullary nail was inserted, thirteen millimeters three weeks after insertion, and twelve millimeters six weeks after insertion.
Antiseptic Levels in Serum
No chlorhexidine or chloroxylenol was detectable in serum with the use of high-performance liquid chromatography (detectability limit, one microgram per milliliter).