Prevalence, Prevention, and Economic Implications of Infection After Total Knee Arthroplasty
Stephen J. Nelson, MD
Patrick K. Strotman, MD
James A. Browne, MD
INTRODUCTION
Periprosthetic joint infection (PJI) is one of the most dreaded complications following total knee arthroplasty (TKA). The combination of increasingly drug-resistant pathogens with a large foreign body in a capacious potential space makes these infections particularly difficult to fully eradicate, requiring intense intervention—both surgical and antibiotic. Even when treated appropriately, PJI carries a significant risk factor for mortality.1 Such outcomes are distressing to both the patient and surgeon. With such weighty consequences, the old adage rings true, “An ounce of prevention is worth a pound of cure.” This chapter will comment on the prevalence of infection following TKA and then discuss risk factors, steps for prevention, and the economic implications associated with PJI.
PREVALENCE
While PJI is a relatively uncommon complication, it has supplanted polyethylene wear and aseptic loosening as the most common reason for revision TKA.2,3,4 Bozik et al evaluated the national inpatient sample and found that 25.2% of revision procedures were performed for infection. Its incidence has been reported to be in a range from 0.7% to 2.2%.5,6,7 Kurtz evaluated the Medicare 5% national sample administrative data and noted that the incidence of infection within 2 years of surgery was 1.55%.7 Crowe examined 3419 primary TKA procedures performed at NYU Lagone Medical Center between 2009 and 2011 and noted an infection rate of 0.76%.5 Pulido evaluated their rates of PJI at Thomas Jefferson University Hospital in a series of 4185 patients prospectively gathered between 2001 and 2006. They also reported an infection rate of 1.1%.6 Despite increasing attempts at prevention, there is some evidence that the incidence of PJI may be slowly increasing given the increase in life expectancies and higher volumes of joint arthroplasty procedures.8,9
Though acute postoperative infection is of primary concern for the orthopedic surgeon, Kurtz and colleagues found that the risk of infection extends beyond the 2-year mark postoperatively, as 0.46% of patients developed an infection between 2 and 10 years.7 In an evaluation of the Finnish health registry, Huotari noted that the incidence rate of late PJI (after 2 years) after hip or knee arthroplasty was approximately 0.07% per prosthesis year.10
Infection rates are exponentially higher following total knee revision procedures, with Frank et al reporting a rate of 7%.11 If the reason for revision TKA is infection, Cochran et al report risk of reinfection being 24.6% for 1-stage and 19% for 2-stage revision patients, respectively, at 1 year postoperatively. In their study, the cumulative risk of revision at 6 years was 38.3% and 29.1% for 1- and 2-stage revision patients, respectively.12 It is unclear whether the higher risk of infection after revision procedures is due to the increased length and complexity of surgery, the incomplete eradication of the infection which indicated the revision procedure in the first place, or the failure to diagnose implant failure as an infectious process. It is likely all of these factors drive the higher infection rate in patients who have undergone multiple revisions.
PREVENTION
The prevention of infection in TKA can be broken down into three phases: preoperative prevention, intraoperative prevention, and postoperative prevention. Efforts during each of these periods are aimed at either increasing the immunocompetence of the patient or decreasing the bioburden of surgery.
Preoperative Prevention
Appropriate patient selection and optimization cannot be overemphasized in the battle to decrease the incidence of infection. There have been several patient-specific risk factors for infection identified in TKA, including inflammatory arthritis, male gender, age, ASA >2, lower socioeconomic status, and hepatitis C.7,13,14,15,16 Season of the year has also been implicated in the likelihood of surgical site infection (SSI).17 Such risk factors may or may not be modifiable. Modifiable risk factors include malnutrition, diabetes, obesity, depression, smoking, opiate dependence, methicillin-resistant Staphylococcus aureus
(MRSA) colonization, and recent arthroscopic surgery or steroid injection. Each of these modifiable risk factors will be addressed individually. It should be noted that there is little prospective data supporting optimization of patient risk factors to decrease PJI rates available at this point in time.
(MRSA) colonization, and recent arthroscopic surgery or steroid injection. Each of these modifiable risk factors will be addressed individually. It should be noted that there is little prospective data supporting optimization of patient risk factors to decrease PJI rates available at this point in time.
Optimal nutritional status to facilitate appropriate wound healing is a requisite for PJI prevention. Saleh et al18 observed that patients who have 5 days or more of postoperative wound drainage were 12.5 times more likely to develop infection than those without drainage. Malnutrition impedes synthesis of collagen-derived peptides and causes immune system dysfunction. Greene noted that preoperative nutritional status was strongly correlated with the likelihood of a postoperative wound complication.19 A preoperative lymphocyte count of less than 1500 cells/mm was associated with a fivefold greater frequency of developing a major wound complication, and an albumin level of less than 3.5 g/dL had a seven times greater frequency.19 Jaberi and colleagues also examined patients who experienced prolonged wound drainage and found that malnourished patients (serum transferrin <200 mg/dL, serum albumin <3.5 g/dL, or total lymphocyte count <1500/mm3) undergoing total joint arthroplasty are more likely to develop deep infection.20 Yi et al found that having one or more of the three laboratory parameters outlined above is independently associated with both chronic septic failure and acute postoperative infection complicating an aseptic revision arthroplasty.21
Diabetes has been associated with mixed outcomes following TKA. Marchant et al evaluated the Nationwide Inpatient Sample and found that patients with uncontrolled diabetes exhibited significantly increased odds of surgical and systemic complications, higher mortality, and increased length of stay during index hospitalization following lower extremity total joint arthroplasty.22 The authors also found an increased risk for wound infection (OR = 2.28; 95% CI = 1.36 to 3.81; P = .002). Conversely, Adams et al23 evaluated the Kaiser Permanente Total Joint Replacement Registry and were unable to identify an increased risk of revision arthroplasty or deep infection in patients with diabetes. Despite this apparent discrepancy, Cancienne was able to identify an association between perioperative HbA1c and the rates of deep postoperative infection requiring surgical intervention after TKA.15,16 The analysis indicated that the inflection point where the risk of infection increased was an HbA1c of 8.0 mg/dL (Fig. 72-1). The authors caution that the A1c should be interpreted in the context of the patient’s other risk factors for PJI.15,16 An attempt at reducing the A1c prior to surgery seems warranted but may not serve as a good cutoff to recommend against surgery.
Obese patients have been found to have an increased risk of infection following TKA. The exact pathophysiology is unknown but may be related to increased surface tension at the incision site or the need to perform a more involved dissection for exposure that results in hematoma formation, seroma collection, or prolonged drainage. Jämsen performed a single center series study of 7181 patients in Finland undergoing lower extremity arthroplasty and demonstrated that the infection rate increased from 0.37% in patients with a normal body mass index (BMI) to 4.66% in the morbidly obese group.24 Nunez et al performed a case control study of 60 patients with
an average BMI of 39.9 who underwent primary TKA. The authors demonstrated nearly twice the risk (11.6% vs. 6.6%) of in-hospital wound problems and three times the rate of PJI compared to controls (5.0% vs. 1.6%).25 A recent study from the Mayo Clinic found a striking association between BMI and risk for deep wound infection (Fig. 72-2). The smoothing spline analysis demonstrated that beginning at a BMI threshold of 35 kg/m2, there was a 7% increased risk of superficial or deep infection per unit increase in BMI above 35 kg/m2 (HR, 1.07; P < .001).26
an average BMI of 39.9 who underwent primary TKA. The authors demonstrated nearly twice the risk (11.6% vs. 6.6%) of in-hospital wound problems and three times the rate of PJI compared to controls (5.0% vs. 1.6%).25 A recent study from the Mayo Clinic found a striking association between BMI and risk for deep wound infection (Fig. 72-2). The smoothing spline analysis demonstrated that beginning at a BMI threshold of 35 kg/m2, there was a 7% increased risk of superficial or deep infection per unit increase in BMI above 35 kg/m2 (HR, 1.07; P < .001).26
Whether or not weight loss prior to TKA reduces the risks of infection remains unclear. A number of authors have examined whether or not bariatric surgery prior to arthroplasty improves clinical outcomes in obese patients; Smith et al performed a meta-analysis of five studies to this end.27 The authors noted that the incidence of wound infection and overall medical complications was lower in the bariatric surgery group.27
Depression has also been associated with poor outcomes following TJA. Gold et al performed a retrospective study utilizing the all-payer California Healthcare Cost and Utilization Project database and found depression was more prevalent in those readmitted at the 90-day time point in their multivariable analysis.28 Ricciardi et al demonstrated a similar finding at the 30-day time point after reviewing 10,759 total knee replacements performed over a 5-year period between 2010 and 2014. In a separate multivariate analysis of the National Inpatient Sample, the rate of diagnosis of depression was 10%, and among those patients, there was a higher risk of infection (OR = 1.33).29 While identifying and treating patients with depression and those with other psychiatric diagnoses which result in poor coping skills may result in improved rates of satisfaction, it is unclear if this will impact rates of PJI.
Tobacco smoking is a modifiable risk factor and has been associated to several adverse postoperative outcomes, including infection and need for revision. Nicotine alters perfusion within the microvasculature system, decreases wound bed oxygen tension, and impairs epithelialization.30 In an evaluation of the Mayo database, tobacco users undergoing total hip arthroplasty (THA) or TKA were found to have higher hazard ratios for deep infection (2.37; 95% CI, 1.19 to 4.72; P = .01) and implant revision (1.78; 95% CI 1.01, 3.13; P = .04).31 Sahota et al looked at complications in smokers undergoing total hip or knee replacement in the acute postoperative period. After utilizing propensity matching to limit confounding variables often observed in this population, they noted that a smoker is twice as likely to experience a surgical complication which was primarily driven by an SSI diagnosis.32 In 2002, Møller et al conducted a randomized controlled trial to demonstrate that smoking cessation 6 to 8 weeks before total hip or knee arthroplasty led to a significant decrease in wound complications and need for secondary surgery.33 A meta-analysis by Bedard et al demonstrated a significantly increased risk of any wound complication (OR, 1.36 [1.16 to 1.60]) and PJI (OR, 1.52 [1.07 to 2.14]) for current tobacco users relative to former tobacco users showing that some smoking-mediated changes may be reversible.34
Opiates have immunosuppressive properties, and preoperative opioid consumption has been found to increase the risk of PJI (adjusted OR, 1.53 [95% CI, 1.14 to 2.05], P = .005).35 A large database study by Cancienne et al examined patients who underwent primary TKA between 2007 and 2015. Their results demonstrated that the most significant risk factor for prolonged postoperative opiate use was filling a preoperative narcotic prescription (OR 5.74). Infection was associated independently with preoperative narcotic use (OR = 1.16) and postoperative narcotic use (OR = 1.33).36 It is important to utilize caution when prescribing opioids in patients with degenerative joint disease who may be candidates for arthroplasty and identify patients via state-wide prescription monitoring programs filling narcotic prescriptions in the preoperative and postoperative periods.
In a recent study, previous steroid injection into a native knee has also been identified as a possible risk factor for infection. Papavasiliou et al37 evaluated three PJIs in a series of 144 patients who underwent TKA. The authors noted that the patients who developed PJI all received an intra-articular steroid in the months before surgery. This finding was confirmed in the evaluation of a medical record database where the incidence of infection was found to be significantly higher in those patients who had a steroid injection within 3 months of surgery.38 There was no significant difference in patients who underwent TKA more than 3 months after injection, suggesting that a window of at least 3 months be maintained between corticosteroid injection and arthroplasty.
Knee arthroscopy within 6 months of arthroplasty has also been identified as a risk factor for postoperative infection. Compared to an age-matched group of control patients also undergoing TKA, there were higher incidences of infection (OR 2.0, P = .004), stiffness (OR 2.0, P = .001), and venous thromboembolism (VTE) (OR 1.6, P = .047) in patients who underwent TKA within 6 months after knee arthroscopy.39,40 There were no increased risks of complications when a waiting period was maintained beyond 6 months. If a patient has recently undergone an invasive procedure about the knee joint, it may be prudent to delay elective TKA to mitigate risk of PJI.
Each of the previously mentioned modifiable risk factors has focused on attempting to increase the immunocompetence of the patient; however, there is also a place for decreasing the bioburden of the host. Asymptomatic colonization with MRSA has been described as a risk factor for subsequent MRSA infection. In a series of 758 patients, MRSA colonization at admission has been found to increase the risk of subsequent MRSA infection, compared with methicillin-sensitive S. aureus (MSSA) colonization (RR, 13; 95% CI, 2.7 to 64) or no staphylococcal colonization (RR, 9.5; 95% CI, 3.6 to 25) at admission suggesting that asymptomatic MRSA hosts should be identified for preoperative MRSA intervention. Rao et al screened patients for nasal carriage of S. aureus and decolonized those patients with mupirocin ointment to the nares twice daily and chlorhexidine bath once daily for 5 days before surgery.41 The authors noted that of all 164 of 636 participants who tested positive for Staphylococcus and completed the decolonization protocol, none had a postoperative infection at 1-year follow-up.41 Kim also evaluated the implementation of a prescreening program for MSSA and MRSA and found that treatment of carriers with an eradication protocol consisting of intranasal mupirocin and chlorhexidine showers decreased infection rates to 0.19%, which was significantly lower than the infection rate observed prior to the decolonization protocol (0.45%, P = .0093).42
Preoperative bathing with chlorhexidine has also been found to decrease the incidence of wound infection following knee arthroplasty. In a meta-analysis of four clinical trials, use of preoperative chlorhexidine was associated with a reduced total incidence of infection (RR = 0.22; 95% CI = 0.12 to 0.40; P = .000).43 Banerjee et al noted that preoperative chlorhexidine multiple showers or topical applications of chlorhexidine may lead to more substantial reduction in colony counts of pathogenic organisms.44 The authors argue that whole-body cleaning rather than site-specific application may confer additional advantages.44