Bacterial Biofilms and Periprosthetic Infections

The diagnosis and treatment of periprosthetic infection after joint arthroplasty is often frustrating for the orthopaedic surgeon. The application of certain diagnostic criteria and different treatment strategies can be better directed if these infections are placed in the context of microbial biofilms. An understanding of this biofilm mode of microbial infection can help to explain the phenomenon of culture-negative infection as well as provide an understanding of why certain treatment modalities often fail. Continued basic research into the role of biofilms in infection will likely provide improved strategies for the clinical diagnosis and treatment of periprosthetic infection.

Periprosthetic infection is a devastating complication of joint arthroplasty with substantial morbid sequelae. In even the best scenarios, patients routinely experience a prolonged period of lost function, lost employment, and lost time that often culminates in a compromised result. Contributing to these poor outcomes is the difficulty of making a proper diagnosis early and providing the proper treatment. A basic understanding of the pathology of these infections may help to direct both the diagnosis and treatment of periprosthetic infection or provide a framework from which current data may be understood.

In most medical settings, the current standard of diagnosing infection continues to rely on a variation of Koch’s postulates, which were formulated in 1884^1,2. In practice, this involves the isolation of an infecting organism from tissue or fluid that is cultured in a specialized broth or on solid growth media. Once isolated, the pathogen can be identified and the sensitivity of the pathogen to various antimicrobial agents can be determined. This simple paradigm has formed the basis of infectious disease treatment for decades, and its effectiveness in most cases is indisputable. The problem, however, is that most bacteria do not grow naturally as simple colonies on agar plates, as do their microbiological laboratory counterparts. Most bacteria in nature grow as biofilms.

The biofilm theory of microbiological growth has been thoroughly studied and has a firm scientific foundation^3,4. The biofilm theory has been well accepted in areas such as marine fouling, water treatment, and the food industry. By this theory, bacteria grow and exist utilizing two different modalities^5-7 (Fig. 1). In one form, these unicellular bacteria can survive and grow in a complex biofilm matrix⁸ that has a structure and function analogous to the extracellular matrix that is the hallmark of higher-order multicellular organisms. As with the extracellular matrix, the biofilm matrix is produced by cells but, in this case, bacterial cells. This biofilm matrix offers protection as well as provides an organizing scaffold, which can facilitate the metabolic activity and even communication between its members within. Bacteria can also exist in a planktonic form, in which they act as more traditionally viewed unicellular organisms. In the planktonic phenotype, there is no structural organization between the individual cells as well as no development of chemical gradients and accompanying microniches. The state of the bacteria as either the planktonic form or the biofilm form has major implications for the treatment of bacterial infections. Bacteria in the planktonic form allow for the spread of infection in that free bacteria can spread to or infect other sites (bacteremia and/or sepsis); but, at the same time, these bacteria are more susceptible to attack by the immune system and antimicrobial agents. On the other hand, bacteria in the biofilm form may not have the same freedom to roam, but they are nonetheless better protected from immune system attack and are less susceptible to antibiotics. It is important to realize that fungal infections, such as Candida, can also exist as biofilms.

Fig. 1.

Schematic drawing showing salient features of biofilm formation in a staphylococcal biofilm model based on studies by Boles and Horswill⁵, Otto⁶, and Resch et al.⁷. The blue boxes show the main processes in biofilm formation, the yellow boxes show the chemical environment, and the red boxes show the phenotype. In planktonic cells, expression of the accessory gene regulator (agr) system results in the production of secreted virulence factors, reduced attachment, and the suppression of biofilm accumulation by increasing dispersal. EPS = extracellular polymeric substances.

The Biology of Infection

It is useful to dissect the course of infection to identify areas that may be exploited for its treatment and prevention. This will be reviewed as it applies to bacterial infections. To initiate infection, an inoculum of bacteria must first become established at a favorable site of infection. This may seem obvious, but it should be remembered that humans routinely coexist with the staphylococci that normally inhabit their skin. It is generally only when these bacteria breach the natural barriers of the dermis, such as in surgery, that they become pathogens. Bacteria that initially enter via the transcutaneous route to initiate localized infection are generally thought to be in the planktonic state. After entry, these bacteria must then adhere to the periprosthetic tissue or implant surface⁹. The molecular mechanisms of bacterial adhesion are becoming better elucidated. In the case of staphylococci¹⁰, for example, adhesins belonging to the MSCRAMM (microbial surface components recognizing adhesive matrix molecules) protein family facilitate the binding of these bacteria to a variety of extracellular matrix proteins. Staphylococcus aureus has more than twenty genes encoding such adhesins. Furthermore, the adhesins that bind to fibronectin in the extracellular matrix also seem to mediate the internalization of these S. aureus bacteria into human cells, where they can even replicate inside the host cell¹¹. Once the bacteria have adhered to a site, they will then go through a replication phase. It is during this planktonic initiation phase that the bacteria are perhaps most vulnerable. As with any infection in an immunocompetent host, the course of the infection is ultimately dependent upon the ability of the host’s immune system to clear the infection. Antimicrobial agents alone will not eliminate infection, but they can certainly tip the balance in favor of the host. Likewise, any mechanism that helps the infecting bacteria to evade or counteract the action of antimicrobial agents or the host’s immune system will provide an advantage to the pathogen. The establishment of a biofilm confers just such an advantage to the bacterium. Once adhered to a surface, S. aureus will enter a phase of growth and colonization. In the course of this process, S. aureus will also release virulence factors that are toxic to the host. The entire process is an extremely well-coordinated system in which specific bacteria genes are turned on and off in response to growth and environmental cues¹⁰. Furthermore, a tightly regulated system of communication named quorum sensing^12,13 has been discovered to occur between these bacteria. This quorum sensing, perhaps akin to the paracrine hormonal system of communication in multicellular organisms, may help to organize the overall growth of the colony as well as coordinate biofilm formation. The production of certain proteins, such as Aap in Staphylococcus epidermidis and SasG in S. aureus, appears to play a role in intercellular adhesion¹⁰. Ultimately, bacteria will become embedded in a matrix composed of polysaccharides, glycoproteins, and extracellular DNA (eDNA) that form the bacterial biofilm. Evidence suggests that S. aureus controls the release of eDNA into the biofilm matrix from the regulated lysis of cells controlled by the cidA gene¹⁴. Bacteria growing in biofilms are fundamentally different from their planktonic counterparts. The former can be thought of as a different phenotype of the latter, although they share the same genotype. In a biofilm, bacteria can be tolerant to antibiotics at concentrations that are several hundredfold greater than that needed to kill planktonic bacteria. Biofilm bacteria are also much more resistant to the toxic effects of human immunity, although leukocytes likely still penetrate biofilm layers. Biofilms also facilitate the exchange of nutrients and provide a continual source of bacteria that can detach from the biofilm, either as released planktonic cells or as biofilm fragments, which can then go on to infect other sites or cause acute systemic infection. In biofilms, bacteria seem to exist in a more quiescent, less virulent state. However, these biofilms can still elicit a host inflammatory response that contributes to continual adjacent tissue destruction that results ultimately in the clinical symptoms of pain and implant loosening seen in longstanding chronic periprosthetic infection. Bacterial biofilms have been directly identified on a variety of medical implant devices, including peritoneal catheters¹⁵, vascular catheters¹⁶, contact lenses¹⁷, orthopaedic devices¹⁸, and joint replacement implants¹⁹. Biofilms have also been identified in a variety of medically important non-implant-related chronic infections such as prostatitis²⁰, cystic fibrosis²¹, endocarditis²², otitis media²³, and osteomyelitis^24,25.

Diagnosis of Periprosthetic Infection in the Context of Biofilms

The diagnosis of chronic periprosthetic infection can be difficult and frustrating for both the patient and the practitioner. Often, chronic periprosthetic infection has to be diagnosed using indirect methods such as erythrocyte sedimentation rate levels, C-reactive protein levels, synovial fluid cell count, and synovial fluid neutrophil percentage²⁶. Newer studies have also used the analysis of leukocyte esterase levels in aspirated fluid²⁷. All of these diagnostic criteria are basically measures of the native immune response to periprosthetic infection rather than a direct identification of an infecting organism. With this difficulty in mind, both the Musculoskeletal Infection Society as well as the American Academy of Orthopaedic Surgeons have published criteria and pathways for diagnosing periprosthetic infection^28,29. The failure to isolate an infecting organism from the aspiration of suspicious joints is not unusual in cases of chronic periprosthetic infection. These difficulties can be better understood when the infection is placed in the context of biofilms. Planktonic bacteria can be isolated and grown with traditional hospital microbiological culture techniques. Biofilm bacteria generally do not culture well using these techniques³⁰. Therefore, chronic infections, which are present mostly in a biofilm state, will be difficult to positively identify using these traditional techniques. Some newer methods have used molecular biological techniques to diagnose periprosthetic infection. This includes the use of polymerase chain reaction methodologies. One recent use of such techniques was able to positively identify the presence of bacteria by ultimately detecting the presence of bacteria-specific ribosomal RNA³¹. A positive test in this case confirms the presence of bacteria, but further tests are required to identify the specific bacterial strain. This requires the design of polymerase chain reaction primers unique to these specific strains. A recent promising technology, called the Ibis technology³⁰, which combines the techniques of polymerase chain reaction and mass spectrometry, has been introduced for the identification of infecting organisms. The Ibis technology was recently used to diagnose the presence of Bacillus cereus in two orthopaedic implant-related infections³². Ibis technology was also recently successful in identifying organisms in culture-negative periprosthetic infection as well as in cases of revision arthroplasty thought to be due to aseptic loosening³³. The latter is especially concerning as fifty of fifty-seven revisions that were not thought to be related to infection may have involved an unrecognized, subclinical infection. This adds to the suspicion, which has been previously suggested³⁴, that many aseptic failures of arthroplasty components may actually involve low-grade chronic infections. Such new molecular techniques may provide value both in the initial diagnosis of periprosthetic infection as well as helping to confirm the clearance of infection when planning a reimplantation. The utility of using molecular techniques for the diagnosis of biofilm-dominated infections has already been proven. In chronic otitis media, molecular techniques have been used to identify a pathogen in 80% to 100% of cases, in which traditional culture provided only 20% to 30% positive results²³. A similar improvement in diagnosis using molecular techniques has been demonstrated in a pilot study for cases of infective endocarditis³⁵.

Treatment of Periprosthetic Infection in the Context of Biofilms

Great effort has been made to define treatment paradigms for periprosthetic infection. As part of these paradigms, periprosthetic infections are often classified as acute or chronic. By one such classification scheme³⁶, early postoperative periprosthetic infection occurs within four weeks after the index procedure and late chronic infection occurs more than four weeks postoperatively. Early infection is usually accompanied by pain, poor healing, erythema, and prolonged wound drainage. Another category of this type of periprosthetic infection includes the acute infection at the site of a previously well-functioning joint replacement, often years after surgery. These infections are often thought to result secondarily from the hematogenous spread of an infection somewhere else in the body and are accompanied by an acute onset of pain accompanied by erythema and joint effusion, like septic arthritis in a native joint. Chronic periprosthetic infections are often more indolent and may simply present as chronic pain, which may be the only clinical symptom. The type of infection, acute or chronic, has implications for its treatment. Acute infections have been treated surgically with a variety of techniques, including irrigation and debridement with component retention; irrigation and debridement with removal of the prosthetic components and the immediate placement of new components (a single-stage exchange); and irrigation and debridement with removal of the prosthetic components, placement of an interim antibiotic cement spacer, and placement of new prosthetic components usually weeks to months later (a two-stage exchange). Chronic infections are treated surgically with either a single-stage exchange or a two-stage exchange. Simple irrigation and debridement with component retention as the treatment for a chronic infection has an unacceptably high failure rate³⁶.

There are conflicting data as to the merits of a simple irrigation and debridement with component retention as the treatment of choice in patients with acute periprosthetic infection. Success rates of this treatment vary considerably. In a recent review of the literature for periprosthetic infection after total knee arthroplasty³⁷, the average failure rate of irrigation and debridement with component retention to clear the infection was 68% (range, 61% to 82%). One explanation for these varying results is that the success of the procedure may be related to the virulence of the infecting organism. For example, irrigation and debridement with component retention as a treatment for infections involving methicillin-resistant bacteria has had very poor results, with a failure rate reported to be as high as 84% at a minimum two-year follow-up period³⁸. However, another recent study has suggested that irrigation and debridement with component retention has a high failure rate regardless of the infecting organism³⁹.

For the treatment of chronic infection, two-stage exchange procedures have generally been considered the gold standard, with a success rate reported to be between 80% and 100%³⁷ when used to treat periprosthetic infection after knee arthroplasty. Similar excellent results have been reported for treating periprosthetic infection after hip arthroplasty^40,41. Comparable success rates have also been noted for single-stage exchange procedures for the treatment of periprosthetic infection after both knee^42,43 and hip arthroplasty⁴⁴.

Some of the controversy in these treatment procedures can be reconciled by placing them in the framework of biofilm eradication. Simply stated, any surgical treatment will ultimately fail if that treatment does not adequately remove the biofilm at the infection site. With this in mind, the varying success of these surgical treatments can be clarified. A simple irrigation and debridement with component retention can be successful if the infection is treated quickly enough so that a biofilm either has not been established on the prosthesis or the biofilm is adequately removed from the prosthesis. This also assumes that other infected tissue has been adequately debrided. Likewise with a chronic infection, a single-stage exchange can be successful if it adequately removes the biofilm at the infection site. This may require a more thorough debridement in that a chronic well-established infection may be established as a biofilm in the osseous tissue adjacent to the prosthesis. While the bacteria in a quickly treated acute infection may not have established a biofilm, it must be assumed that all chronic infections involve biofilms and so must be treated with this in mind. While a two-stage exchange may seem more reliable to many, it will also ultimately fail if a biofilm is left behind at the treatment site. The single-stage and two-stage procedures, by virtue of the removal of the arthroplasty components (and any biofilms contained on their surfaces) and allowing access to infected tissue around the removed components, simply allow for a better opportunity for removing the infecting biofilm at the site of the infection.

Future Directions

Future goals involve exploiting the fundamental processes that contribute to bacterial infection and biofilm formation. An obvious first step is to prevent bacterial adhesion. One strategy in this regard has been to design implant surfaces that are unattractive to bacteria. A specific example of this would be the use of antiadhesive silicone rubber surfaces in voice prostheses⁴⁵. While an obvious area for future research, this is also perhaps one of the most challenging as well, since any surface modification of cementless arthroplasty components would have to also allow for osseointegration. Another approach has involved the application of antibiotic coatings to implant surfaces. This has been studied for hydroxyapatite materials with some promising in vitro results⁴⁶. Other studies have involved vancomycin that is covalently linked to the implant surface⁴⁷. Recent studies in a sheep model have shown such an implant to be effective in inhibiting S. aureus infection while still promoting bone-healing⁴⁸. While certainly promising, these approaches utilizing antibiotic coatings are still susceptible to the remarkable capacity of bacteria to develop antimicrobial resistance. Another strategy to inhibit bacterial adhesion involves the use of biosurfactants^49,50. These are amphiphilic compounds that can be adsorbed onto surfaces. Some of these occur naturally and are actually produced by certain bacteria such as lactobacilli. These compounds may certainly be applicable for coating silicone surfaces. It is unknown if such compounds will find an application in orthopaedic devices. Yet other materials may prove to be more useful. Recent studies have used farnesol, a citrus-derived alcohol that inhibits quorum sensing, and have shown that it can inhibit S. aureus biofilm formation on titanium alloy discs⁵¹. While some materials may not be appropriate for coating implant surfaces, they may still find an application in the removal or disruption of biofilms. Such effective materials could be used to complement the usual debridement performed in surgical treatments of periprosthetic infection.

Interfering with the phenomenon of quorum sensing presents another area for possible therapeutic treatment. As the molecular details of quorum sensing become better understood, strategies for disrupting this bacterial crosstalk become more realistic. For example, researchers have found that they can prevent graft-associated infections by a variety of Staphylococcus species, including methicillin-resistant species, by using the quorum-sensing inhibitor ribonucleic-acid-III-inhibiting peptide⁵². The ultimate goal is to find such inhibitors that may be active across several bacterial species and are not unique to preventing only very specific infections.

Other research has focused on the idea of developing a vaccination against common pathogens such as S. aureus. In a rabbit model of chronic S. aureus osteomyelitis, researchers were able to develop a vaccination against specific biofilm antigens. This vaccination in combination with vancomycin treatment significantly reduced the infection in this model and even showed efficacy against methicillin-resistant S. aureus biofilm infections⁵³. This model also highlighted the important phenotypic differences between planktonic and biofilm bacteria in that vancomycin treatment was necessary to eliminate the former and vaccination treatment was necessary to eliminate the latter. Treatment with vancomycin or vaccination alone was not as effective as dual treatment in eliminating these infections. The isolation of specific antigens and the subsequent production of antibodies may find a valuable application in the diagnosis and treatment of biofilm infections. Antibody and/or antigen-derived molecular technologies have been used to great effect in cancer diagnosis and treatment. It is certainly reasonable to conjecture that such technologies could be applied to the treatment of periprosthetic infection.

In conclusion, many of the frustrations encountered in the diagnosis and treatment of chronic periprosthetic infection can be understood if placed in the context of biofilms. Basic research will continue to provide insights into the biology of biofilms that should provide opportunities for the improved diagnosis and treatment of these chronic infections. One specific challenge will be to develop strategies that are applicable for the treatment of a broad spectrum of pathogens and are not simply unique to a single bacterial or fungal strain.