Prosthetic Joint Infection

Javier Cobo; Jose Luis Del Pozo


Expert Rev Anti Infect Ther. 2011;9(9):787-802. 

In This Article


General Aspects

Management of a patient with a PJI represents an extraordinary challenge that needs a multidisciplinary approach. Decision making in PJI is usually complex and difficult to delineate in algorithms that are either too simple (with a high number of patients not represented in them) or too complicated to understand and follow. This fact exemplifies the high number of variables that should be taken in account in the management of this complex disease.[47] The difficulty lies not only in the complexity of the disease, but also in the knowledge and correct analysis of the literature. Information appears scattered in a multitude of small series. Different authors have used different classifications and approaches. Some focus only on surgical aspects and others emphasize the antimicrobial therapy.[48,49] Furthermore, most investigators define success as infection eradication minimizing functional outcomes that are frequently sacrificed to the microbiological success.[49] Almost all the questions that should be addressed regarding to the antimicrobial therapy (i.e., choice, dosage, and extent, route of administration and convenience of combination) remain without definitive responses in the case of PJI. Proof of the absence of evidences is given in the absence of published guidelines.

However, in spite of complexity and uncertainties, some advances and consensus have been reached throughout the last few years. Some studies have demonstrated that following a few principles and algorithms of management improves outcomes.[50–52] Nevertheless, we should remain open to changes and new information. Meanwhile, decisions should be individualized, and infectious diseases specialists and orthopedic surgeons should work together in order to provide the best approach for each patient based on a critical review of the current information.[26]

The main objectives of PJI treatment are to alleviate pain, to restore the function and to eradicate (or control) the infection.[3,26] It is not always possible to achieve all three objectives and this reality explains why the decision is frequently so complex. Aggressive surgical approaches often lead to a higher probability of curing the infection, but they carry the risk of the worse functional results.[49]

Once PJI has been diagnosed, the first and more relevant question to address is whether it is necessary to remove the implant. In general, it appears that is easier to cure a foreign body infection by removing it. Nevertheless, some other issues should be also considered.

Are we Treating an Acute Infection? According to the current evidences, both early PJI and hematogenous PJI may be managed without removal of the implant.[3,23,26,47,48] Several case series have shown success rates higher than 60% with debridement and retention of the prosthesis in these types of infections.[25,53–57] However, literature is not homogeneous and some authors are not so confident.[58–60] Although it is well recognized that a short time (no more than 2–4 weeks) between PJI diagnosis and debridement is crucial to achieve a successful outcome, the time boundary to rule out the possibility of keeping the implant is not well defined.[3,23,61,62] The limit to consider a PJI as 'early' is not agreed in the literature. Some authors keep this term for infections appearing during the first month after surgery,[23,25,56] but others extend the limit until the third month.[48,54,63] It would be necessary to have information specifically of those cases managed without removal of the implant that presented with an 'early' PJI during the second or third month after surgery to answer this question. There are a scarce number of publications addressing the hematogenous PJI outcome since most of them are usually classified as 'acute' PJI together with early PJI. A recent study has shown that microbiological cure is more likely to be achieved when implant removal is performed, but that those caused by streptococci or Gram-negative bacilli showed an excellent cure percentage in spite of the retention.[64] In summary, even for a modest success rate, the opportunity of solving the infection with a much more simple surgical procedure made this approach a reasonable option. Furthermore, this treatment modality does not prevent a more aggressive step being taken in the future. By contrast, chronic late infections that appear months or years after surgery in an insidious way are almost never cured without implant removal.

Is There an Implant Loosening? If the prosthesis is loose, its function is usually compromised and patients suffer pain. This situation is frequent in chronic late infections. In the vast majority of prosthetic loosening, implant removal is the only solution, even in cases where infection could be eradicated.

Does the Patient Have the Option of Having a New Prosthesis Implanted? What would be the benefits of implant removal? What functional outcome could be anticipated with implant removal? Bone stock is reduced with every new surgery and this makes it difficult to implant a new prosthesis. Furthermore, success rates decrease with each revision and at the same time the risk of a new infection or persistence increases.[65,66]

What is the Surgical Risk? What is the patient's vital prognosis? We should not forget that we are discussing a surgery performed for quality of life and only in rare occasions whether or not to remove the implant is a vital consideration. Patients should be informed about advantages and disadvantages of the medical–surgical options and the probabilities of achieving the objectives of the treatment. The preferences and priorities (i.e., pain, functional state, avoiding surgery and hospitalization and so on) of the patient should be carefully considered. The patient's viewpoint does not always meet with the opinion of the surgeon.[67] It is surprising that patients' needs and expectatives are not mentioned in most reviews and recommendations on PJI management.

What Microorganisms are Involved & Which Antibiotics can be Used for the Patient? Some microorganisms are difficult to treat and this fact could be associated with worse outcomes. For example, several authors have shown a high failure rate when methicillin-resistant S. aureus (MRSA) is involved.[60,68] Nevertheless, this experience has not been met by others.[53,55] Other important issues are, for example, staphylococcus rifampin resistance, which would impact on outcomes,[69] or P. aeruginosa ciprofloxacin resistance precluding an oral prolonged therapy.

Therapeutic Options

Although Tsukayama's classification provides us with a management strategy in most cases, therapeutic management should be individualized for each patient. In the same way, it is critical to state that antimicrobial therapy must be consistent with the surgical approach. Main options to treat a PJI are represented in Figure 2 and are summarized in the following paragraphs.

Figure 2.

Prosthetic joint infection. (A) See text for discussion of alternatives. In general, patients with failure of conservative strategy (DAIR) and patients with chronic late infections should be treated with prosthesis removal, but in selected cases, especially if anticipated functional results are not optimal or the implant is stable, chronic suppressive therapy should be considered. Options should be discussed with the patient. (B) For patients treated with prosthesis exchange by one-stage procedure, 6 weeks of antibiotics after the procedure is reasonable. Most groups also treat for 6 weeks after the removal of the implant in two-stage procedures. Nevertheless some groups question this duration of therapy if antibiotics are added to the spacer. In any case antibiotic prophylaxis must be used. Antibiotics can be prolonged waiting for cultures obtained during the second stage, at least in selected cases. (C) For those patients who are not candidates for a new prosthesis, arthrodesis (knees) or Girdlestone procedure (hips) are selected.

Prosthesis Retention With Curative Intention With this option we assume both that a functional implant is salvagable and that infection eradication is possible. Several series have documented that cure of a substantial proportion of selected patients with acute infections can be achieved with debridement, irrigation of the joint and a definite course of antimicrobial therapy.[25,54–57,63] Technical surgical aspects are relevant. For example, several reports show that results are worst when debridement is performed by arthroscopy than when it is by open arthrotomy,[70–72] which allows exchange of some mobile components of the prosthesis, such as the polyethylene. Today, this option is kept for early or hematogenous PJI; assuming that the implant remains stable, soft tissues are preserved and debridement is performed promptly (see later discussion).[23,47]

Prosthesis Retention With Noncurative Intention In some situations it is assumed that infection cure is not possible without implant extraction, but removal is not feasible. This option implies indefinite antimicrobial therapy also known as chronic suppressive therapy. Although this strategy has been typically reserved for moribund patients or those with very high surgical risk,[48] it can be also considered for other patients that reject surgery or those in which the functional results of surgery are not expected to be acceptable.[56,72,73] Cessation of therapy in these patients is usually followed by a recurrence of the process during the first 4 months,[72] which supports the usefulness of the chronic suppressive therapy. The ideal situation for prescribing a chronic suppressive therapy is when the microorganism is susceptible to an oral and well-tolerated antibiotic, there is no pain and the implant remains stable.

Prosthesis Exchange This is the most frequently chosen therapy. Removal and new prosthesis implantation can be performed during a single surgical procedure or, more commonly, in two separate stages. In this case, an antibiotic-loaded spacer is implanted in order to preserve the anatomy, facilitate the second-step stage and achieve high local levels of antibiotics. Most groups also administer systemic antibiotics directed against isolated microorganisms. Finally, several weeks later, a new prosthesis is placed after removing the spacer.[74] Benefits of a one-stage step prosthesis exchange are a shorter hospitalization and less mechanical complications;[49,75] however, there is less published experience, and rates of success appear lower than with the conventional two-step stage approach. The actual trend is to keep the one-stage procedure for selected cases with less aggressive infections.[76,77]

Resection Arthroplasty In some scenarios, either a new prosthesis cannot be replaced or maintaining its function is not possible. The objectives are to control the infection and avoid pain. Girdlestone procedures in prosthetic hip infections usually control the infection but produce shortening of the limb and strong motor difficulties. For prosthetic knee infections, an instrumental arthrodesis is the most commonly performed procedure when the patient is not a candidate for a new prosthetic joint. Infection and pain are usually controlled but the limb will remain rigid and shortened.[78] In rare instances, amputation can be necessary to control infection.

Antimicrobial Therapy Depending on the Surgical Strategy

For each surgical strategy, duration of antimicrobial therapy, dosage, route of administration and convenience of combined therapy need to be defined. The level of evidence is low and almost all the recommendations are based on short literature series, experimental data and, finally, on expert experience and opinion.

Prosthesis Retention With Curative Intention The most important issues to consider are duration of therapy, route of administration, usefulness of rifampin-based combinations and the role of other antistaphylococcal agents, including some of those that have been recently commercialized.

Duration & Route of Administration By definition, in this scenario, antimicrobial therapy should finish in a defined time. There is consensus in administering a 'prolonged' therapy, but the length of therapy has varied enormously: from 4 weeks used in the Tsukayama et al. series[25] to 12 months reported by Aboltins.[55] Based on the only clinical trial published regarding antimicrobial therapy in orthopedic implant related infections (with approximately a dozen of PJI included), 3 and 6 months of therapy are recommended for prosthetic hip and knee infections, respectively.[79] However, recent series have communicated good results in spite of shorter antimicrobial courses (i.e., less than 3 months) for both locations.[56,63] Furthermore, in a recent nonrandomized study comparing 6 against 12 weeks of therapy in different PJI types, no benefits were observed in the 12-week arm.[80] C-reactive protein level normalization is used by some groups to make the decision of stopping therapy,[81] nevertheless, data from a short series are against normalization of C-reactive protein as a requirement to stop antimicrobial therapy in PJI.[82] Thus, it is not currently possible to state the optimal duration of antimicrobial therapy in PJI managed without implant removal with curative intention. Series are not comparable, and small sample sizes do not allow rejection of the idea that longer therapy could cure a higher proportion of patients. Such uncertainties invite an individualization of the antimicrobial therapy taking into account adverse events and clinical response. Another issue is the optimal length of intravenous therapy. The aforementioned study[80] and other similar experiences suggest that a long duration of intravenous therapy is not always necessary.[53,54,80]

Rifampin Usefulness Several studies carried out in the 1990s showed that traditional in vitro studies failed to predict success of antimicrobial therapy in biofilm-related infections.[83] Biofilm bacteria are in a stationary mode of growth and, under these circumstances, most antibiotics show almost no effect. Quinolones and specially rifampin are among the antimicrobials that exhibit better activity in this complex scenario.[84,85] Today, it is well accepted that rifampin is an essential drug for PJI therapy, at least if the implant is retained.[79,86,87] Rifampin usefulnesss in staphylococal implant-associated infections is based on three points: the aforementioned activity against nongrowing bacteria; the reduction in the risk of development of resistance to ciprofloxacin[88] and perhaps other antistaphylococcal agents; and the clinical experience briefly summarized in the following section. In 1992, Widmer et al. published a small case series including patients whose implants could not be removed. Nine out of 11 were successfully treated with antimicrobial combination regimens containing rifampin.[86] Subsequently, the same group published results from a clinical trial, again with few patients, comparing ciprofloxacin monotherapy (15 cases) with combination therapy consisting of ciprofloxacin with rifampin (18 cases). In this study, differences favoring combination therapy were demonstrated.[79] Interestingly, most failures were due to the development of ciprofloxacin resistance. This study had an enormous impact because of the high success rate (89%) in infections treated without implant removal. However, more than a half of the cases involved osteosynthesis fixation devices for fractures, which can be managed with antimicrobials waiting for fracture healing. Also, most published PJI case series show better results when rifampin is used in combination with other antistaphylococcal agents,[53–55,72,81] in contrast with others showing bad results that do not mention its use.[58,59] Finally, a recent retrospective study has demonstrated better results with rifampin-based regimens in patients managed with implant retention.[69] In spite of preliminary studies that used 450 mg/every 12 h[53,81] there are data supporting a 600 mg dosage per day. Rifampin should never be used in monotherapy owing to the high risk of resistance development. This risk is substantially reduced when it is used in combination with other antistaphylococcal agents.

Role of Other Antistaphylococcal Agents It is difficult to discern whether the role of other antistaphylococcal agents is limited to 'accompanying' rifampin, preventing the development of resistance, or whether they are able to cure PJIs themselves. Ciprofloxacin has been the most frequent quinolone prescribed for PJI antimicrobial treatment; however, higher antistaphylococcal intrinsic activity and pharmacodynamic reasons have determined levofloxacin as the standard in combination with rifampin. Initially, a standard 500 mg/day dosage was used,[53,81] but experimental studies suggested that 750 mg/day could be more convenient.[89] In fact, experts currently recommend this dosage.[3,23] However, it should be remembered that there are scarce data regarding the safety of levofloxacin at high doses in prolonged therapy and some concern arises from at least one study.[90]

Linezolid is an attractive agent for bone and joint infections owing to its excellent oral absorption, its bone tissue penetration, its activity against multidrug-resistant Gram-positive bacteria and the good results in experimental models when combined with rifampin.[91–93] There are several studies reporting favorable clinical experience with linezolid in bone and joint infections.[94–98] Some of them include PJI, but numbers and details of the cases are scarce. Linezolid-related hematological toxicity appears variable in the different series. Although it is not prevented by pyridoxine administration,[99] it can be managed by careful monitorization. By contrast, neurotoxicity, although infrequent, may be nonreversible.[100] Both issues limit its use in prolonged therapies.

Daptomycin activity against multidrug-resistant Gram-positive bacteria and against nongrowing bacteria have stimulated the interest in this new antibiotic for the treatment of PJI, especially those in which MRSA is involved.[101] A recent review has summarized the published experience with daptomycin in bone and joint infections.[102] In spite of a mean length of therapy of 37 days, no relevant toxicity was detected. Since resistance development may be a limitation, higher dosage regimens and/or combination therapy studies are warranted.

Two other antistaphylococcal drugs, cotrimoxazole and fusidic acid, remain active against most staphylococci strains, and have been studied in the treatment of PJI. Cotrimoxazole is not well tolerated at high doses,[103] but at lower doses used in combination with rifampin, it has shown efficacy in treating bone and joint infections after debridement.[104] The optimal dose of cotrimoxazole is not well established when the prosthesis is retained. Fusidic acid, also in combination with rifampin, has shown to be useful in some series. Since there is some risk of resistance development, optimized dosage is recommended.[55,105]

Prosthesis Retention With Noncurative Intention The experience of some series of 'chronic supressive therapy' includes several dozens of patients treated with monotherapy or a minocicline–rifampin combination. The objectives are to control local symptoms and to reduce progression and the need for surgery. Success rates varied between 23 and 86%.[73,106–108] Although the studies of Marculescu[109] and Byren[72] were not properly published as chronic suppressive therapy, many patients included in these long series were probably chronic infections managed with implant retention and very long courses of antimicrobials. Analysis of both studies is difficult because of the heterogeneous nature of the population. However, two facts can be observed: first, success rate decreases with time (i.e., therapy delays a needed surgery); and second, stopping the therapy is associated with a more rapid treatment failure. These two findings confirm the utility of chronic suppressive therapy. In order to appropriately select the candidates for chronic suppressive therapy it is important to understand the aim of this strategy. Neither the dosage or the utility of rifampin-based combinations are established so far. Long-term toxicity to antimicrobials should be considered, especially in the elderly.[110] Cotrimoxazole appears to be an interesting option because of the low percentage of staphylococci strains resistant to it and the wide experience with prolonged therapy.

Prosthesis Exchange Antibiotic-loaded cement is usually placed during the surgery. It has been demonstrated to be safe and provide high local levels of antibiotics (gentamicyn, vancomycin and clindamycin are the most widely used). In some countries, antibiotic-loaded cements are commonly used as prophylaxis for primary arthroplasties. Their use in infected arthroplasties seems to be useful,[49,77] but is not supported by clinical trials. When the implant is removed and local antibiotics are added to the cement, the role of the systemic antimicrobial therapy could be considered less relevant. Main questions to be addressed are: the optimal length of therapy, what antimicrobials should be chosen and the convenience of combinations containing rifampin. Furthermore, infectious disease specialists are frequently queried regarding the optimal time to perform the second surgical stage and the need and interpretation of cultures obtained on (or before) it.

Duration In one-stage procedures, 6 weeks of antibiotic therapy after the surgical procedure is the most commonly used duration of therapy.[51,76] Classically, 6 weeks has been the extent of antimicrobial therapy administered between the two stages in a two-stage prosthesis exchange. It should be followed by a 'prophylaxis' before the second stage (in our opinion it should be wider in spectrum than the usual one), and stopping antibiotics once the cultures obtained during the second procedure are informed as sterile.[51,76] Nevertheless, this length of therapy in a two-stage procedure is not supported by well-designed, randomized clinical trials. In the last 2 years, several published reports propose substantially shortening the length of antimicrobial therapy, suggesting a nonessential role when a two-stage surgery and antibiotic-loaded cement spacers are used.[111–114] However, more information or comparative noninferiority studies are needed to confirm this hypothesis.

Role of Rifampin In classical series of PJI managed with prosthesis exchange, rifampin has not been usually included in the antimicrobial regimen. Since the implant is removed, its role may not be as important, but we do not have comparative sudies to rule it out.

Procedures for the Second Stage Usefulness of cultures before the second stage is not well established. In Mont's opinion it could be of benefit because it allows identification of a 9–14% of cases that otherwise would relapse.[115] Although other studies have not found any benefit from performing such cultures,[116,117] similar percentages of positive cultures (multidrug-resistant coagulase-negative staphylococci in most cases) obtained during the second stage are reported.[66,116] There are different (nonexclusive) explanations for this finding. It could be due to a microbiological failure produced by resistance development to the antimicrobials used, or a selection of pre-existing resistant species or clones not cultured (but present) at the first stage. In this case, use of glycopeptides, lypopeptides or linezolid (agents with a very low prevalence of resistance among Gram-positive bacteria) would be reasonable whatever the susceptibility profile of the isolated bacteria during the first stage. Another option is reinfection or spacer colonization occurring during the first stage, even in the presence of preloaded antibiotics in the cement.[118–120] Finally, in some instances, certainly, they only represent a contamination. Overall, it appears that the presence of positive cultures in the second stage is not necessarily associated with poor outcomes, perhaps because most patients are treated with antibiotics.[66] However, more studies are needed to address this question.

Optimal time and criteria to decide when to perform the second stage are also not well defined. The most common strategy consists of completing the 6 weeks of antibiotics after the first stage and verifying normal C-reactive protein levels, a sort of 'green-light', to perform the second stage. However, several recent reports have questioned this strategy.[82,116] Others suggest waiting 2 weeks after stopping antibiotics in order to confirm that C-reactive protein levels do not rise and to increase the yield of cultures obtained during the second stage; however, this strategy has not demonstrated better results.[66,117]

Resection Arthroplasty After a definitive prosthesis removal of the hip (Girdlestone procedure) the most logical approach is to administer 4–6 weeks of antibiotics, as in other types of osteomyelitis. In the case of instrumental arthrodesis of the knee, the procedures would be identical to those of the two-stage prosthesis exchange.[78]