CORR Insights®: Meropenem-loaded Cement Is Effective in Preventing Gram-negative Osteomyelitis in an Animal Model

Where Are We Now?

Currently, the management of orthopaedic infection is left to the expertise of the attending physician. The practitioner must determine which microbe is causing the infection, whether an implant can be retained, and how to best treat the infection to restore patient health and function. One of the techniques used by surgeons to manage infection is the implantation of antibiotic-loaded spacers as part of a two-stage revision. Many different formulations exist. In general, antimicrobials are not soluble in polymethyl methacrylate, and the poragen fraction, or amount of water-soluble drug mixed into the cement, is directly responsible for how much drug is eluted from the finished spacer. The more antimicrobial that is added, however, the weaker the resulting cement becomes, and the more likely it is to fail under mechanical load or torque. My previous work [6] and that of other authors [1, 2, 4] have documented hand-mixing of cement for surgeon-directed use that can be used, where appropriate, to combat infection in the absence of standardized products.

The work presented by Wei et al. [7] is a good study in a string of studies [3] on the hand-mixing of relatively small doses (about 2 grams per 40 grams of acrylic cement) of antibiotic into bone cement to create an antibiotic-loaded cement for fixation in primary arthroplasty. The authors studied meropenem, ceftazidime, piperacillin, ciprofloxacin, aztreonam, and tobramycin for their efficacy against a gram-negative infection in the knee of a rat infected with P. aeruginosa. Similar to previous work [5], they noted reasonable strength of the resulting cement for implant fixation and antimicrobial release that was consistent with that of low-dose cement. Antimicrobial activity was maintained against the gram-negative bacteria tested. Although the observed release was low compared with that of high-dose cement, these levels are sufficient for treating infection based on the surgical model they created. This surgical model did not involve the mature biofilms that may be found on an infected implant. The work by Wei et al. [7], however, provides clear and useful data that meropenem, ceftazidime, piperacillin, ciprofloxacin, aztreonam, and tobramycin can be used in surgeon-mixed spacers and can be expected to elute into the wound space. These results should give surgeons confidence that any of the antimicrobials studied by Wei et al. can be incorporated into hand-mixed spacers and can be expected to elute in accordance with their poragen fraction to combat active infection.

Where Do We Need To Go?

Although this study [7] provided good benchtop data for recipes of particular doses, there is still no accepted or accurate method to determine how much antimicrobial is needed to successfully treat an infection, what concentrations are best, or how those amounts might vary if the wound is inadequately debrided. Future work also needs to evaluate the tradeoffs associated with higher concentrations of antibiotics in the cement; namely, how much mechanical strength is enough for applications in which antibiotic-loaded cement is used? Additionally, how much time is needed for the antimicrobial to release from a spacer to sterilize an infected wound? To get good, consistent results to manage infections, it is critical that we continue to fill these knowledge gaps.

How Do We Get There?

The answer to the first two questions is likely agent-specific, and it might be wound-specific. That is why work like that of Wei et al. [7] represents a good incremental step toward understanding orthopaedic infection. Questions about dose can only be studied in animal models in order to evaluate human outcomes. Infection surgeons I have worked with have argued that high-dose spacers should be used for two-stage revision, and these spacers should have at least 10 grams of poragen hand-mixed into the cement to provide increased elution, but that is based mostly on expert opinion and case series. Working with those surgeons, I proposed gadolinium-based models that sought to image wound concentrations near the surface of an implant with some success. Similarly, studying the interaction with debridement and locally dosed antimicrobials is tough, even in animal models, because the volume of tissue in animal appendages is much less than that found in humans, and the degree of granulation tissue and reaction varies from wound to wound. Presumably, a surgeon will not intentionally debride a wound poorly, but the surgeon will not always know in every procedure that all of the foreign or infected material has been removed. Thus, the strategy will likely be to limit the patient’s physical activity during the first stage of a two-stage revision in a lower extremity and increase the poragen fraction so the wound is fully bathed. Anecdotally, I know of successful procedures in which 10 grams of poragen was used, but I am unaware of any standard for sufficient strength in a two-stage revision, and I know such cement has strength that falls under the 70 MPa guideline for permanent fixation. Ideally, as the field matures, we need to come together to create standard recipes and practices with well-understood parameters to manage implant infections and ensure better patient outcomes after orthopaedic infection.

Assessing the effectiveness of debridement in humans is challenging. No investigational review board is likely to approve a study that intentionally underdebrides orthopaedic wounds in patients to study such relationships. By leveraging our previous work and other similar small-animal models, it might be possible to approximate the wound volume for major orthopaedic procedures and understand how the amount of granulation tissue and degree of debridement interacts with the local delivery of antibiotics. I suspect that these studies will find that retained fibrous tissue greatly reduces the permeability of antimicrobials, and such tissue must be well-debrided to achieve a reliably successful outcome. I am not sure what retained granulation tissue may do to local delivery in such wounds. In general, we know that the bactericidal effect of many antimicrobials such as gentamicin might be only a few days, but the actual time the material needs to be in place likely varies depending on the state of the wound, whether the microbe is in biofilm, and whether the microbe has genes that make it more resistant to a given antimicrobial. The establishment of a standardized large-animal model with clear links to clinical procedures (such as two-stage THA) would help to standardize and clarify the management of infected orthopaedic wounds and help bridge the knowledge gaps that still exist, despite continuing benchtop and small-animal work.