A quarter of patients treated for PJI dead within 5 years
Orthopedists are often chastised for a lack of consideration of the systemic health of their patients and a lack of understanding or care for general medicine. While this characterization is often unfair, the management of periprosthetic joint infection is an area in which the orthopedic surgeon needs to provide all-encompassing care.
Of course, while we discuss the many avoidable risk factors for periprosthetic joint infection (PJI), it is not an all-together preventable complication. PJI strikes even patients who have pristine management and no identifiable risk factors. While the impact that overall health has on the development of PJI is well known, what is less discussed is the effect of PJI on overall health.
The risk factors for PJI include, but are not limited to, obesity, increased age, male gender, uncontrolled diabetes or metabolic syndrome, and cardiac comorbidities. These patient characteristics are also predictive of decreased longevity. As has often been discussed in previous Infection Watch columns, the treatment of PJI is prolonged and often requires multiple surgical interventions with systemic antimicrobial therapy that may provide a risk of mortality. Lastly, undiagnosed or unsuccessfully managed PJI has the potential to progress to bacteremia and possibly sepsis providing another avenue for death. With this reality, it would not be unexpected to see an association between mortality and PJI.
Repeatedly, such a relationship has been reported anecdotally. In a recent report on two-stage exchange to treat periprosthetic hip infection, Berend and colleagues found a 7% mortality rate prior to reimplantation. They also reported a 90-day mortality incidence of 4%, with 48% of the cohort dying during the study period (average follow-up of 4.5 years; range: 2 years to 10 years). Toulson and colleagues also assessed the outcome of the management of periprosthetic hip infection, finding that more than 25% of patients were deceased at attempted follow-up.
Leung and colleagues studied the outcome of the treatment of periprosthetic hip infection caused by resistant organisms. They reported 24% mortality at an average of 4 years following treatment. In a similar manner, Zmistowski and colleagues reported that 26% of patients undergoing two-stage exchange for gram-negative PJI were not reimplanted because of death. With these reports, decreased longevity following PJI is strongly suggested. Yet, while convincing, this collection of evidence is not rigorous as they lacked control groups.
For this reason, we undertook a study to define the relationship between PJI and mortality, and establish predictors of death following treatment for PJI. Two cohorts were formed: 436 patients treated for PJI and 2,342 patients undergoing aseptic revision. The Social Security Death Index was used to identify deceased patients in both cohorts. The overall mortality in the combined cohorts was 3.4%, 5.4% and 15.2% at 1 year, 2 years and 5 years, respectively. The PJI cohort had statistically significantly greater incidence of mortality at 90 days (3.7% vs. 0.8%), 1-year (10.6% vs. 2%), 2-year (13.6% vs. 3.9%), and 5-year (25.9% vs. 12.9%) time points. However, the mortality incidence was only statistically significant for time periods between and 30 days and 90 days and 90 days to 1 year.
No difference was detected within 30 days of the initial surgery, suggesting the relationship between PJI and decreased longevity was not associated with an increased surgical insult. As mentioned, increased mortality in PJI patients may be explained by the typical patient characteristics predisposing them to PJI that would also predispose them to death. To assess this theory, an analysis was performed to see if PJI was a predictor of 1-year mortality, while controlling for age, gender, body mass index, number of procedures performed, overall patient health as measured by the Charlson Index, ethnicity and joint type. Even when adjusting for these factors, PJI provided a nearly six-fold increase in the odds of 1-year mortality. Within the PJI cohort, the predictors of 1-year mortality were found to be gram-negative organisms, increased age, increased Charlson Index and specific comorbidities, such as cardiac, gastrointestinal, and history of stroke. A trend of methicillin-resistant Staphylococcus aureus predicting mortality was observed.
These findings support previous reports of alarmingly high mortality in patients with PJI. Furthermore, analysis suggests that the increased rate of death is not due to an increase in potentially life-threatening surgical and medical therapy, but rather a combination of the disease process and poor overall health of patients with PJI. Furthermore, there are manageable patient characteristics that predict death in PJI patients.
What remains to be known is the effect of successful infection eradication on mortality. It is intuitive that an inability to clear an infection would have a negative effect on overall health and also lead to multiple surgical insults posing increased risk for death. Furthermore, patients with PJI have poor functioning joints, increased physical pain and are likely in a depressed state preventing adequate health maintenance. Yet, in a retrospective manner it was not possible to assess the infection status of patients at various time points prior to death.
It has been well established that PJI is a devastating complication that can destroy a previously functioning joint. Its prevention is difficult and it continues to pose a substantial burden as the rates of arthroplasty increase. However, with the current patient population presenting for arthroplasty, difficulty with PJI prevention is unlikely to change. Cram and colleagues demonstrated that during the past 2 decades, the number of patients undergoing primary total hip arthroplasty with obesity has tripled, the number of patients with diabetes has doubled and the number of average comorbidities has doubled while the age has remained unchanged.
When compared to aseptic revision, the 5-year relative survival (survival rate when isolating the impact of PJI) was 87.3%. It is important to note that while primary joint replacement may increase longevity, aseptic revision increases the risk of early death. Therefore, it is possible that the mortality attributable to PJI is being understated. To put this in perspective, PJI is more deadly than three of the five most common cancers (see chart). As the number of arthroplasties, and therefore prevalence of PJI continues to increase, an imperative task will be prevention of this disease. Despite the many advances in limiting PJI, the most essential element in the prevention of PJI is optimization of patients. Diseases, such as diabetes, obesity and malnutrition, are easily identifiable and require management prior to arthroplasty. It is essential that patients at high risk for PJI understand a real risk of death.
Quick and effective treatment
When PJI occurs, quick and effective treatment of the joint is a necessity to maintain the joint integrity and contain its progression. However, this new knowledge also requires surgeons to be aware of the patient’s overall health and, if necessary, involve a medical team in the management of chronic conditions that may predispose a patient to death.
In conclusion, more than 25% of patients treated for PJI are dead within 5 years. Fortunately, infection of the joint is not cancer. Many of its causes, ranging from poor host status to poor surgical technique, are preventable. While the scope of the orthopedist’s care has historically excluded management of chronic medical conditions, the prevention of PJI with its risk of death requires the orthopedist to ensure effective preoperative and long-term management of chronic diseases.
Berbari EF. Risk factors for prosthetic joint infection: Case-control study. Clin Infect Dis. 1998;27(5):1247-1254.
Berend KR. Clin Orthop Relat Res. 2012;doi:10.1007/s11999-012-2595-x.
Cram P. JAMA. 2011;doi:10.1001/jama.2011.478.
Leung F. Clin Orthop Relat Res. 2011;doi:10.1007/s11999-010-1725-6.
Peersman G. Infection in total knee replacement: a retrospective review of 6489 total knee replacements. Clin Orthop Relat Res. 2001;392:15-23.
Pulido L. Clin Orthop Relat Res. 2008;doi:10.1007/s11999-008-0209-4.
Toulson C. J Arthroplasty. 2009;doi:10.1016/j.arth.2008.07.004.
Zmistowski B. J Arthroplasty. 2011;doi:10.1016/j.arth.2011.03.044.
For more information:
Disclosures: Parvizi is a consultant to Zimmer, Smith and Nephew, 3M and Convatec. Zmistowski has no relevant financial disclosures.