Authors: Lewis G, Nichols S, Le N, Hester D, Levin LS, Schoenfisch M, Klitzman B, Duke University Medical, CenterDurham, NC
Title: Effect of nitric oxide release from compression plates: anti-infective efficacy
Purpose: he purposes of this study were to quantify the anti-infective effect of implants coated with NO-eluting xerogel and to assess factors influencing rates of implant-related infection.
Methods: The anti-infective efficacy of the nitric oxide (NO)-releasing xerogel coating was quantified using a rabbit contaminated internal fixation model. Stainless steel 2mm compression plates (Synthes; West Chester, PA) were coated with a sol-gel coating and allowed to dry. Half were then charged with diazeniumdiolate NO donors and then sterilized and stored at -80 C prior to use. Staphylococcus aureus cultures were grown from a -80 C 15% glycerol stock overnight (strain provided by Dr. Mark Shirtlif, Univ. Maryland-Baltimore Dental College, Baltimore, MD). The resultant suspension was then transferred into fresh TBS and grown further to achieve a log phase culture with a standard concentration of 10^8 CFU/ml. Twelve adult white New Zealand rabbits were anesthetized initially with ketamine/xylazine and maintained with 2% isoflurane. The lateral thigh skin was incised bilaterally, the femurs were exposed and a single screw hole was drilled. The wounds were then inoculated with 30 ul S. aureus culture (3x10^6 CFU) and one hour of in situ incubation elapsed. Each compression plate was affixed with one screw. One plate/screw had xerogel coating only and the contralateral plate/screw had NO-releasing xerogel. Four rabbits had added monitoring of wound surface temperature and laser Doppler blood flow. Rabbits were observed daily for local or systemic infection as well as general health. On day seven, rabbits were anesthetized and the plates, bone and muscle were explanted for qualitative microbiological analysis. A standardized muscle biopsy from the deep head of the vastus intermedius muscle was obtained from each implant site and was then homogenized in saline and subjected to serial dilution and culture to quantify the CFU/mg dry muscle.
Results: S. aureus were successfully cultured from all subjects at the time of explantation. Positive cultures were obtained for the wound swab, the implant, and the bone sample. Contamination by other bacterial species was observed bilaterally in only one rabbit. For the inert plate sites infection levels ranged from 2 to 2.5x10^5 CFU/mg with a mean of 2.9x10^4. For the reactive plate sites infection levels ranged from 6 to 8.8x10^4 CFU/mg with a mean of 1.8x10^4. There was no statistically significant difference in levels of infection between the two groups (p = 0.5443) overall. Intra-operative data were analyzed to determine which parameters measured correlated with bacterial count. There was a statistically significant inverse correlation between the temperature of the muscle surface at the time of inoculation and the final bacterial load (r = -0.9094; p = 0.0119). Intra-operative core temperature was also weakly correlated to infection level (r = -0.4162; p = 0.0! 431). Laser Doppler blood flow at the time of implantation did not correlate with infection levels.
Discuassion and Conclusion: Prior studies by our group and others evaluating NO-releasing xerogel coatings on other implantable devices have shown a marked decrease in infection levels. Thus the lack of observed difference in infection levels was surprising. There are many reasons why the anti-infective efficacy of supplemental NO may have been reduced. Perhaps the establishment of infection at the surface of the bone follows a different time-course than that of subcutaneous tissue and, therefore, the level and time course of NO delivery from this formulation may not be optimal. Other NO delivery systems may have greater anti-infective properties than the donor used in this study. There was an enormous amount of variability in the infection levels observed for both the reactive and control groups (standard deviation = 24289 and 70419 CFU/mg, respectively). Other variables had undue influence on infection. Indeed, in those rabbits that had a lower core body temperature!, the reactive plate sites yielded a lower calculated bacterial load than the inert plate site. At higher temperatures ( & 39C) the results were more mixed. Further investigation may better elucidate the interplay between temperature and NO-releasing coatings.