Much research has been conducted on the effectiveness of impregnated bone grafts as local antibiotic delivery vehicle in treating or preventing orthopaedic infections (Peeters et al. 2019). The most common antibiotics for impregnation are vancomycin and tobramycin. Although cefazolin is widely used systemically for infection prophylaxis, studies in which bone grafts are impregnated with cefazolin are rather limited. This work aimed to investigate several factors (i.e. concentration of impregnation solution, duration of impregnation, timing of impregnation throughout the production process, type of bone) that could influence the decisions to be made when establishing a cefazolin impregnation protocol for bone chips. First, an in-house impregnation protocol was set up in which three types of bone grafts, unprocessed fresh frozen bone (UFFB), processed fresh frozen bone (SFFB) and lyophilized bone (SLB), were impregnated with different concentrations of cefazolin. During these preliminary tests, we observed a fast release of cefazolin within the first 2 hours. The bone chips were rinsed with saline after impregnation to remove the remaining cefazolin on the surface of the bone chips. However, we noticed that the number of wash steps affected the amount of released cefazolin, particularly when bone chips were only impregnated for 10 min. Extensive washes of bone chips that were impregnated for a longer time did not result in less released cefazolin. This might suggest that a longer incubation time will enhance a deeper penetration of cefazolin into the bone chips. Additionally, Mathijssen et al. impregnated bone chips for only 10 min and noted a negative effect on cefazolin release when rinsing cefazolin impregnated bone chips twice with saline. This effect was not observed after vancomycin impregnation (Mathijssen et al. 2010). The quantitative and qualitative differences with respect to the nature of antibiotic binding sites could explain this observation. Both the cefazolin concentration of the impregnation fluid and the time used for impregnation positively influenced the subsequent release of cefazolin. When using our in-house impregnation protocol, we noticed a difference in cefazolin release between frozen and lyophilized bone grafts (Fig. 3b). Additionally, when impregnating bone grafts cleaned with supercritical CO2 there was a beneficial effect on the release of cefazolin when impregnation was carried out immediately after washing. In our hands, the impregnation efficiency of cefazolin is enhanced when impregnation takes place directly on wet bone grafts without a freeze drying step in advance. This is in sharp contrast with the finding of Winkler et al. that the loading capacity of bone grafts is increased when the tissue is in a completely dry state (Winkler et al. 2000). However, Winkler only examined bone chips impregnated with vancomycin and tobramycin and not with cefazolin. In contrast, Coraça-Huber et al. showed that the release of gentamicin is similar for lyophilized bone and fresh frozen bone throughout the experimental period (Coraça-Huber et al. 2016).
The release time of vancomycin from bone chips impregnated with our in-house impregnation protocol was much shorter than that from bone chips impregnated according to Winkler’s protocol. Winkler claimed that an appropriately processed allograft is an excellent carrier for antibiotics. Supercritical CO2 is an alternative method to clean and process bone, leaving a pure scaffold of bone matrix that can bind antibiotics throughout the whole graft (Winkler and Haiden 2017).
Due to the apparent difference between our in-house protocol and Winkler’s protocol, we decided to revise our in-house cefazolin impregnation process and rely on bone grafts purified with supercritical CO2 prior to solution impregnation and lyophilization as Winkler used to do with vancomycin and tobramycin. To investigate the impregnation efficiency of cefazolin different setups were considered such as a different bone size, gamma irradiation and the incorporation of the impregnation process at different time points during the entire bone production process. Our results showed that impregnation after freeze drying resulted in less cefazolin release compared to bone grafts impregnated before freeze drying. Finer morselized bone granules released more cefazolin than coarser bone chips, probably due to the increased surface area of the granules. Additionally, Witsø et al. studied several variables that may affect the adsorption and subsequent release of vancomycin and netilmicin and determined the concentration of the released antibiotics with a fluorescence polarization immune assay (Witsø et al. 2002). He noticed that the amount of eluted vancomycin was not influenced by the degree of bone morselizing; however, the amount of eluted netilmicin from fine morselized bone was higher than the amount eluted from coarse morselized bone. Like cefazolin, netilmicin is a small molecule, which can probably explain the similar result.
When looking at the elution kinetics, we observed a drop in the cefazolin concentration in the first 24 hours, which means that the degradation of cefazolin occurs faster than the accumulation of released cefazolin. As the cefazolin stability observed in aqueous solution did not fully explain the degradation observed in the bone chips elution experiments performed in newborn calf serum, we tested whether the elution fluid affects the stability of cefazolin. Cefazolin was dissolved in newborn calf serum or in aqueous solution and incubated at 37°C for 4 days (newborn calf serum condition) and 8 days (aqueous solution condition). Cefazolin dissolved in newborn calf serum is degraded much faster at 37°C compared to cefazolin in aqueous solution see supplementary Fig. 7. However, the decrease seen in newborn calf serum is still less than the decrease of 40–50% seen after 24 hours in the presence of the bone chips. Probably, the freeze drying process might have an additional impact on the stability of cefazolin. Figure 2b shows that temperature affects the stability of cefazolin, which is in accordance with the results of Mathijssen et al., who noticed a reduction in cefazolin activity when cefazolin was stored at 37°C for one month (Mathijssen et al. 2010). On the other hand, gamma irradiation does not affect cefazolin stability (data not shown). This could be of interest for the manufacturers of processed impregnated bone chips who perform impregnation after irradiation, necessitating an additional sterilization step at the end of the production process, as is the case for vancomycin and tobramycin impregnation following Winkler’s protocol.
Overall, the elution kinetic results are in line with the preliminary results of our in-house cefazolin impregnation protocol, indicating that supercritical CO2 cleaning and the additional freeze drying procedure after impregnation do not influence the effectiveness of cefazolin impregnation, in contrast to vancomycin impregnation. Thus, the elution profile of bone chips is different for different antibiotics, and using an existing impregnation protocol for a different antibiotic will not necessarily lead to similar results. The unexpected behavioural difference between antibiotics is actually something Witsø also observed when comparing the in vitro release of netilmicin and vancomycin (Witsø et al. 2002).
Despite its rapid degradation, the initial burst release of cefazolin was still a hundredfold above the minimal inhibitory concentration (MIC), potentially being locally toxic (Pilge et al. 2016). Its clinical implications, especially with regard to infection prophylaxis, remain unclear and require further clinical research.