A periprosthetic joint infection represents a devastating complication, may be difficult to eradicate, and may impair the quality of life of the patient. In the study by Zmistowski and colleagues, the authors asked whether the quantity of life, or life expectancy, was altered in patients surgically treated for a periprosthetic joint infection as compared with patients undergoing revision surgery for aseptic reasons. The authors also asked whether there were factors predictive of mortality that could be identified in the group of patients with periprosthetic joint infection. The answers to both questions are important and worthy of investigation.
From the authors’ institutional database from January 2000 to March 2010, 2955 patients were identified as having undergone either a revision hip or knee arthroplasty. Using a modification of the Musculoskeletal Infection Society’s definition of periprosthetic joint infection (the authors’ institution does not utilize histological analysis as a diagnostic criterion for infection), 436 patients were identified as having undergone revision surgery for a diagnosed infection. The remaining 2342 patients who underwent revision surgery for all other causes were used as a comparison group. The risk of mortality was assessed for both groups. For the group of patients with periprosthetic joint infection, a multivariate analysis was subsequently performed. The authors showed, in answer to their first question as to whether quantity of life is altered, that mortality was increased at one year in patients treated for infection compared with patients undergoing an aseptic revision. An increase of more than fivefold in the risk of death was found for patients treated surgically for a periprosthetic joint infection. Relative to the authors’ second question of factors predictive of mortality, the study showed that increasing age, a worse preoperative status (as evaluated by the Charlson Comorbidity Index), and a history of stroke or cardiovascular disease or being treated for a polymicrobial infection were predictive of an increased risk of mortality. This information is important for the clinician and makes rational sense; clearly, both the burden of infection and the treatment of an infection may influence mortality. This study raises two questions. Is the surgical intervention required to treat a periprosthetic joint infection comparable in morbidity with the surgical intervention required in an aseptic revision? Are all infections comparable in the health burden with the host?
Is the magnitude of the surgical procedure itself an explanation for the increased death rate identified in the population with periprosthetic joint infection? The study was not designed to compare the specific type of surgical treatment chosen for the individual patients in the periprosthetic joint infection group (i.e., irrigation, debridement, and liner exchange; single-stage component removal and reimplantation; or two-stage component removal and later reimplantation) with the revision surgery performed for the individual patients in the aseptic group (head and liner exchange, one-component revision, or two-component revision). Any surgical intervention inherently has a relative associated morbidity (up to and including mortality) and places a burden on the patient’s current health; this health burden may be difficult to quantify, yet one could rationally assume that for either group the morbidity would be greatest in a two-component exchange and the least in a liner exchange. Answering specifically whether the mortality rate identified in patients with periprosthetic joint infection is influenced by the surgical treatment chosen needs further investigation.
Any infection requires the interaction of the host and the infecting agent. This study shows that polymicrobial infections place a patient at a higher risk of death. Is a patient with a polymicrobial infection already of a different host (immunologic) status compared with a patient with a staphylococcal infection? The study was not designed to evaluate differences between those patients with an acute infection as compared with those with a chronic infection. Does a chronic infection place the patient at a higher risk of mortality due to an assumed health burden secondary to the length of the infection? If the patient is assumed to have been successfully treated for an infected joint, does having a periprosthetic joint infection predispose this patient to a subsequent infection (e.g., pneumonia) in the future? Could this influence the mortality rate that is seen? The study raises more questions that require scientific insight.
The treatment of periprosthetic joint infection remains difficult and is in evolution. The number of patients requiring operative intervention for infection is likely to increase. The authors cite the work of Bozic et al., who have reported that infection is the most common cause of failure in total knee arthroplasty and the third most common cause of failure in total hip arthroplasty1,2. The use of irrigation, debridement, and liner exchange for the treatment of early periprosthetic joint infection has come under scrutiny as failure of this surgery may influence the results of a subsequent two-stage exchange3. The gold standard for the treatment of chronic periprosthetic joint infection has been a two-stage exchange, with a reported success rate of 85% to 90%4. In a meta-analysis of the literature examining the reported results of two-stage exchange compared with single-stage exchange for the treatment of an infection after total hip arthroplasty, Wolf et al., using a Markov decision analysis model, concluded that although a two-stage procedure resulted in a greater likelihood of the eradication of infection, it also yielded a greater chance of death compared with a single-stage exchange5. What constitutes the optimal treatment for a periprosthetic joint infection requires the consideration of multiple variables and may be considered a moving target. As our understanding of those factors that may influence the postoperative outcome evolve, we should ask whether a sixty-year-old patient with diabetes, morbid obesity, and methicillin-resistant Staphylococcus aureus should be surgically treated in the same fashion as an eighty-year-old patient with chronic obstructive pulmonary disease and Staphylococcus epidermidis? Although two-stage surgery is felt to be optimal at times, the result may be a lessening of the patient’s longevity. Identifying those patients most at risk for a diminished longevity from our well-intentioned surgical interventions must be a goal of future research.
Zmistowski et al. are to be commended for questioning the influence of the burden of a periprosthetic joint infection on mortality. The take-home message is that a periprosthetic joint infection results in an unhealthy state. The tools to optimize the treatment, with two goals in mind, infection eradication as the optimal, combined with consideration of the impact of our recommended treatment(s) on patient longevity, will develop as further studies yield insight into this increasing problem.
Note: The author would like to thank William L. Bargar, MD, Jeffrey N. Katz, MD, MSc, and John P. Meehan, MD, for their insight and suggestions.