1.3.1 Periprosthetic Joint Infection

In 2012, Rodriguez-Merchan reported the risk factors for infection following TKA [11]. They were obesity, diabetes, a history of open reduction and internal fixation, male sex, remnants of previous internal fixation material, body mass index (BMI), congestive heart failure, chronic pulmonary disease, preoperative anemia, depression, renal disease, pulmonary circulation disorders, rheumatologic disease, psychoses, metastatic tumor, peripheral vascular disease, and valvular disease.

At 30 days, the overall percentage of surgical site infection is 1.1%, whereas the published rate of deep infection is 0.1%. The lifetime frequency of PJI after TKA ranges from 0.7% to 4.6% [18].

Evangelopoulos et al. have reported that PJI is the predominant cause of early failure of primary and revision TKA, followed by aseptic loosening, instability, pain, malalignment, and inlay wear [6]. Reinfection percentage of the septic primary TKAs was 5%. Infection was the major cause of a second revision, reaching as high as 50% in all cases. The outcomes of this study suggested that septic failure of a primary TKA is likely to occur within the first 2 years after implantation. Septic failure of primary TKA did not influence survival of the revision prosthesis.

Rhee et al. studied the risk factors for PJI, revision, death, blood transfusion, and longer hospital stay 3 months and 1 year after primary total hip arthroplasty (THA) and primary TKA [19]. They analyzed all primary THA and TKA cases between 2000 and 2014. A total of 10,123 primary THA and 17,243 primary TKA procedures were performed. With THA, the risk of PJI was higher in patients with heart failure and those with diabetes. For TKA, liver disease and blood transfusion were associated with a higher risk of PJI. Revision rates were higher among patients with hypertension and those with paraparesis/hemiparesis for THA and higher among patients with metastatic disease for TKA. Important risk factors for death included metastatic disease, older age, heart failure, myocardial infarction, dementia, rheumatologic disease, renal disease, blood transfusion, and cancer. Multiple medical comorbidities and older age were associated with higher rates of blood transfusion and longer hospital stay.

Matar et al. reported a higher failure frequency of two-stage revision for infected TKAs in significantly compromised (host-C) patients [20]. They performed a prospective consecutive series (level IV of evidence) of two-stage revisions of infected TKAs in host-C-type patients with a minimum 2-year follow-up using objective and patient-reported outcome measures. Thirteen patients were included, with a median 5-year follow-up (range 2–10). Median age was 68 years (range 59–73) at time of initial presentation. All patients were a type-C host, using the McPherson classification system. All patients had primary TKAs in situ, with proven chronic PJI; the infecting bacteria were Staphylococcus aureus in 5 of 13 patients, coagulase-negative Staphylococci in 5 of 13, and the remaining three patients had mixed growth. All patients underwent a two-stage revision protocol. At the final follow-up, 9 of 13 patients were infection-free, achieving satisfactory results. Two patients had recurrent infections with different bacteria and were treated with suppressive antibiotics and salvage knee fusion, respectively. Moreover, two patients had chronic pain and poor functional results with insufficient extensor mechanism and significant bone loss; they later underwent salvage knee fusion. This study highlighted the challenge of treating infected TKA in physiologically compromised patients, with 9 of 13 (69%) achieving satisfying clinical results [20].

Fu et al. analyzed the correct timing of second-stage revision in managing PJI, as well as investigating dependable indicators and risk factors [21]. They reviewed and followed 81 TKA patients with infection who underwent two-stage revision in a 5-year period (2010–2014). The study included 56 men and 25 women; all patients were verified as PJI with the same phenotypic cultures. The average age was 64.8 (range 36–78) years, and the mean follow-up time was 46.5 (range 12–72) months after the second-stage surgeries. Serum C-reactive protein (CRP), erythrocyte sedimentation rate (ESR), and intraoperative frozen section (FS) at the time of reimplantation were analyzed. The spacer detention time and antibiotic treatment time were compared. Ten patients underwent failed first- or second-stage surgical procedures. The overall success frequency was 87.7%. The intraoperative FS proved to be good indicator at the time of reimplantation; the sensitivity and specificity were 90% and 83.1%. Serum CRP and ESR showed a poor diagnostic value at the time of reimplantation. A typical bacterial infection, positive FS, and prior sinus were high-risk factors for failure of two-stage revision. Spacer detention time between 12 and 16 weeks had a higher success percentage than over 16 weeks. The main conclusion was that the proper timing of reimplantation should be linked with dissipation of clinical symptoms and negative intraoperative FS with spacer detention time at 12–16 weeks [21].

In 2018, Rajgopal et al. analyzed whether previous failed debridement, antibiotics, irrigation, and implant retention (DAIR) affect the outcome of subsequent two-stage revision performed for PJI after TKA [22]. They performed a retrospective study of 184 knees with completed two-stage RTKA for PJI, operated by a single surgeon in a 12-year period (2000–2011). The series was divided into two groups: those with prior failed DAIR (F-DAIR) (88 knees) and direct two-stage revision (96 knees). At an average follow-up of 5.3 years, the failure frequency was 23.86% (21/88 knees) in the F-DAIR group and 15.62% (15/96) in the direct two-stage revision group. A previous F-DAIR procedure was associated with approximately twice the risk of failure compared with direct two-stage surgery. Excluding PJIs caused by methicillin-resistant Staphylococcus aureus, methicillin-resistant Staphylococcus epidermidis, and Pseudomonas from analysis showed similar failure percentages between the two groups. The frequency of culture negativity and PJI with resistant organisms was higher in the F-DAIR group. The percentages of eradication of methicillin-resistant Staphylococcus aureus and Pseudomonas infection were much lower in the F-DAIR group. The main conclusion was that a failed previous DAIR led to higher failure percentages, lower functional results, and an increased risk of wound-related complications [22].

1.3.2 Obesity

In 2014, Rodriguez-Merchan reported that although some articles (with low grade of evidence) did not find that obesity adversely affected TKA outcomes, most found that obesity adversely affected TKA results [23]. Regarding complication rates and survival rates, obesity was demonstrated to have a negative influence on outcome after TKA. The improvements in patient-reported result measures, however, were similar irrespective of BMI. Regarding the impact of TKA on obese patients, an extra cost of $3050 has been reported per patient. Considering that 50% of the US population is obese and that 600,000 TKAs are implanted per year, the impact for the US health system could be as much as $915 million (300,000 × 3050). TKA in obese patients could be justifiable because the functional improvements appear to be equivalent to those of patients with a lower BMI. However, in obese patients, the risk of complications is higher, and the prosthetic survival is lower. Moreover, TKA in obese patients has a significant impact on the health-care system, which should be considered [23].

Tohidi et al. analyzed 10-year mortality and revision after TKA in patients with morbid obesity [24]. A total of 9817 patients were analyzed, aged 18–60 years, treated with primary TKA in a 5-year period (2002–2007). Patients were followed up for 10 years after TKA. Risk ratios of mortality and TKA revision surgery in patients with BMI > 45 (morbidly obese) compared with BMI ≤ 45 (nonmorbidly obese) were determined, making an adjustment for age, sex, socioeconomic status, and comorbidities. Approximately 10.2% (1001) of the group was morbidly obese. Patients with morbid obesity were more likely to be female than the nonmorbidly obese (82.5% vs. 63.7%) and showed higher 10-year risk of death but were otherwise analogous in characteristics. Approximately 8.5% (832) of the patients had at least one revision surgical procedure in the 10 years following TKA; the revision percentage did not vary by obesity. The main conclusion was that patients with morbid obesity ≤60 years had a 50% higher 10-year risk of death but had no difference in the risk of revision surgery [24].

1.3.3 Diabetes Mellitus

Being younger and male, having various comorbid conditions or greater diabetic severity, getting care at regional or public hospitals, and not having a diagnosis of degenerative or rheumatoid arthritis have been recognized by Tsai et al. as risk factors postoperative PJI after TKA for patients with diabetes. As for the risk of RTKA, postoperative PJI and being younger were significant risk [25]. This study examined the 2002–2012 data from Taiwan’s National Health Insurance Research Database to conduct a retrospective cohort analysis of patients with diabetes older than 50 years of age who underwent TKA.

1.3.4 Pulmonary Disease

Gu et al. published the influence of chronic obstructive pulmonary disease (COPD) on postoperative results in patients undergoing RTKA [26]. A retrospective cohort study was performed using data collected from the American College of Surgeons National Quality Improvement Program database. All patients who underwent RTKA between 2007 and 2014 were identified and stratified into groups based on COPD status. The percentage of complications after surgery was assessed with univariate and multivariate analyses where appropriate. Patients with COPD developed more postoperative adverse events, including deep wound infection, organ infection, wound dehiscence, pneumonia, reintubation, renal insufficiency, urinary tract infection, myocardial infarction, sepsis, and death. Patients with COPD also returned to the operating room and had extended hospital stays. COPD was demonstrated to be an independent risk factor for development of wound dehiscence, pneumonia, reintubation, renal insufficiency, and renal failure. COPD was also recognized as an independent risk factor for unplanned returns to the operating room. The main conclusion was that patients with COPD are at greater risk for wound dehiscence, pneumonia, reintubation, renal insufficiency, and renal failure complications in the postoperative period than those without COPD. Although risks for independent adverse events remain relatively low, consideration of COPD status is an important factor to consider when selecting surgical candidates and evaluating preoperative risk [26].

1.3.5 Drug Abuse

Roche et al. have published that patients who abuse drugs are at increased risk for RTKA [27]. The Medicare database within the PearlDiver Supercomputer (Warsaw, IN, USA) was queried to identify 2,159,221 TKAs performed during an 8-year period (2005–2012). Drug abuse was subdivided into cocaine, cannabis, opioids, sedatives/hypnotics/anxiolytics (SHAs), amphetamines, and alcohol abusers. There was a significant increase in the number of primary TKAs in users of cocaine, cannabis, opioids, SHAs, amphetamines, and alcohol. Amphetamine users had the fastest mean time to revision (691 days). At 30 days, 90 days, and 6 months postoperatively, cocaine users had the highest proportion of patients requiring RTKA (7%, 12%, and 20%, respectively); and at 1 year postoperatively, it was abusers of alcohol (38%). PJI was the most common cause of RTKA in all drug abuse/drug-dependent groups. Based on these outcomes, patients who abuse drugs are at increased risk for RTKA [27].

1.3.6 Opioid Use

Bedard et al. have found preoperative opioid use to be independently associated with a greater risk for early RTKA. Younger age, obesity, and smoking were also associated with increased risk. These findings support efforts to reduce inadequate opioid prescribing [28]. The Humana administrative claims database was queried to identify patients who underwent unilateral TKA during a 9-year period (2007–2015). Patients were followed for the occurrence of an ipsilateral revision procedure within 2 years. Preoperative opioid use was defined as having an opioid prescription filled within the 3 months before TKA. Age, sex, diabetes, obesity, chronic kidney disease, and anxiety/depression were also analyzed. A total of 35,894 primary TKA patients were identified, and 1.2% had had an RTKA procedure within 2 years. Some 29.2% of the patients filled an opioid prescription within the 3 months before TKA. Preoperative opioid users were significantly more likely to undergo early RTKA (1.6% vs. 1.0%); preoperative opioid use, younger age, obesity, and smoking were associated with early RTKA [28].

Weick et al. found that preoperative opioid use was associated with higher readmission and revision rates in TKA [29]. This prognostic study (level IV of evidence) showed that preoperative opioid use was associated with significantly increased risk of early revision and significantly increased risk of 30-day readmission after TKA. This study illustrated the increased risk of poor results and augmented postoperative health-care utilization for patients with long-term opioid use prior to TKA.

Law et al. have reported that cannabis use increases risk for RTKA [30]. A retrospective review of the Medicare database for TKA, RTKA, and causes was performed using Current Procedural Terminology and International Classification of Diseases ninth revision codes (ICD-9). Patients who underwent TKA were cross-referenced for a history of cannabis use by querying ICD-9 codes 304.30-32 and 305.20-22. Cannabis use was prevalent in 18,875 (0.7%) TKA patients, with 2419 (12.8%) revisions within the cannabis group. The revision rate was significantly greater in patients who used cannabis. Time to revision was also significantly increased in patients who used cannabis, with increased 30- and 90-day revision frequency compared with the non-cannabis group. Infection was the most common cause for revision in both groups (33.5% nonusers versus 36.6% cannabis users). Cannabis use can result in decreased implant survivorship and increased risk for revision within the 90-day global period compared with cannabis nonusers after primary TKA [30].

It has recently been reported that although opioids have been widely used for pain control following TKA, multiple level I and II studies have been published in recent years supporting the use of local infiltration analgesia and multimodal blood loss prevention approaches for improving pain control and accelerating recovery after TKA [31, 32]. In another recent report, Waldman et al. strongly recommended that institutions ensure non-opioid-based comprehensive pain management and multimodal and regional anesthesia strategies for TKA [33]. These approaches have been demonstrated to diminish opioid use, increase patient satisfaction, and shorten lengths of stay.

1.3.7 Smoking

In 2018, Bedard et al. investigated the potential impact of smoking on RTKA [34]. They found that smokers had a higher percentage of any wound complication (3.8% vs. 1.8%), deep PJI (2.5% vs. 1.0%), pneumonia (1.3% vs. 0.4%,), and reoperation (5.0% vs. 3.1%) compared with nonsmokers undergoing RTKA. A multivariate analysis identified current smokers as being at a significantly increased risk of any wound complication and deep PJI after RTKA. This study showed that smoking significantly augments the risk of PJI, wound complications, and reoperation following RTKA. The outcomes are even more exaggerated for revision procedures compared with published effects of smoking on primary TKA adverse events [34].

Rodriguez-Merchan reported that orthopedic perioperative complications of smoking include impaired wound healing, augmented PJI, and poorest TKA outcomes [35]. The adoption of smoking cessation methods such as transdermal patches, chewing gum, lozenges, inhalers, sprays, bupropion, and varenicline in the perioperative period should be advised. Perioperative smoking cessation appears to be an efficacious method to diminish postoperative complications, even if implemented as late as 4 weeks before TKA [35].

1.3.8 Metal Allergy (Nickel Sensitization)

Lionberger et al. have investigated the potential role of metal allergy sensitization in RTKA [36]. They hypothesized that nickel sensitization plays a role in the pathology of failed TKA in patients with unexplained dissatisfaction. Thirty-two patients with symptomatic TKA without obvious mechanical findings were tested prior to revision surgery. Nineteen nickel-sensitized and 13 nonsensitized patients were compared by cell counts of synovium surgical specimens for CD4⁺ and CD8⁺ cell lines. Patients were then revised with ceramic-coated implants. The nickel-sensitive patients showed a statistical increase in CD4⁺ reactivity compared with CD8⁺ reactivity. The ratio of CD4⁺/CD8⁺ T lymphocytes was 1.28 in nickel-sensitive patients versus 0.76 in the control. This study provided objective data via histological analysis in support of a nickel allergic sensitization in failed TKAs in which clinical and/or radiographic abnormalities might not be apparent [36].

Fröschen et al. have reported that the implantation of a cementless, hypoallergenic TKA might be a surgical treatment strategy in patients with evidence of allergies [37]. They reported six patients with aseptic loosening following TKA who underwent revision surgery after testing positive for benzoyl peroxide (BPO) hypersensitivity. After clarification of possible other causes of implant failure, epicutaneous testing was performed, and the implants were replaced in a two-stage procedure with cementless, diaphyseal anchoring, hypoallergenic (TiNb-coated) revision implants. Epicutaneous testing revealed a BPO allergy in all six patients and an additional nickel allergy in three of the six patients. There was no histopathological or microbiological evidence for a PJI. The clinical follow-up demonstrated a low level of pain with good function, a stable knee joint, and a proper implant position. Two implant-specific adverse events occurred: femoral stress shielding 2 years postoperatively with no further need for action and aseptic loosening of the tibial stem with the need of revision 3 years postoperatively. The regression of complaints after RTKA with cementless and nickel-free revision implants suggested allergic implant intolerance [37].

1.3.9 Preoperative Valgus Deformity

In a prognostic study (level III of evidence), Mazzotti et al. have shown that preoperative valgus deformity has twice the risk of failure compared with varus deformity after TKA [38]. A total of 2327 TKA procedures performed from 2000 to 2016 were included in the study. A total of 640 primary TKAs with a diagnosis of valgus deformity were evaluated, with a median follow-up of 3.3 years; 1687 primary TKAs with a diagnosis of varus deformity were evaluated with a median follow-up of 2.5 years. Bi-compartmental, cemented, posterior-stabilized, fixed-bearing implants were preferred. For both diagnoses, the implant survivorship percentage was greater than 98% in the first year. However, the survival curve of the TKAs implanted for valgus deformity showed a greater slope in the first 3 years compared with the survival curve of those implanted for varus deformity. Valgus deformity had a 2.1-fold higher risk for RTKA compared with varus deformity. Infection was a major cause of implant failure in TKAs for varus deformity (9/24, 37.5%), whereas its rate was lower for valgus deformity (1/21, 4.8%) [38].

1.3.10 Hybrid vs. Standard Cemented Fixation

Gomez-Vallejo et al. compared the results of RTKA with hybrid vs. standard cemented fixation in a level III of evidence study [39]. During the period 2000–2013, RTKA was performed on 67 patients (29 cemented arthroplasty and 38 hybrid fixation). The average follow-up was 7 years (range 2–15). The main conclusion was that although the outcomes were analogous for the two groups, hybrid fixation tended to produce better outcomes than cemented fixation. In view of the risk of further loosening, these authors advised hybrid fixation [39].

1.3.11 Immediate Postoperative Mechanical Axis

In a 10-year follow-up study, Park et al. investigated whether immediate postoperative mechanical axis is associated with the revision rate of primary TKA [40]. They evaluated the association between the immediate postoperative mechanical alignment of the lower limb and the frequency of RTKA by comparing an adequate mechanical axis group (within ±3° from neutral alignment) and an outlier group (>3° deviation from neutral alignment). The main conclusion was that restoration of neutral limb alignment resulted in a lower revision percentage and higher longevity in TKA. However, there were no significant differences in clinical results between the two groups [40].

1.3.12 Physical Activity

In a prognostic study (level III of evidence), Ponzio et al. demonstrated that active patients have an elevated revision risk. The revision rate was higher for active patients (3.2%) compared with inactive patients (1.6%) at 5–10 years postoperatively. Accordingly, active patients should be carefully counseled regarding TKA to give them an understanding of its limitations and the potential risk of future revision. Active patients were defined by a Lower Extremity Activity Scale (LEAS) level of 13–18 [41].

1.4 Conclusions

Aseptic loosening, periprosthetic joint infection (PJI), and periprosthetic fracture are the most frequent causes of revision after total knee arthroplasty (TKA). Other less common causes of revision TKA (RTKA) are instability, pain, polyethylene wear, restriction of motion (arthrofibrosis), extensor mechanism insufficiency, mechanical defect, and metal allergy (nickel). Failed previous debridement, antibiotics, irrigation, and implant retention (DAIR) are related to higher failure percentages. Preoperative opioid use, younger age, obesity, and smoking are associated with early RTKA. Risk factors for RTKA are preoperative valgus deformity (valgus deformity has a 2.1-fold higher risk for RTKA compared with varus deformity), immediate abnormal postoperative mechanical axis (>3° deviation from neutral alignment), and physical activity (the revision rate is higher for active patients [3.2%] compared with inactive patients [1.6%] at 5–10 years postoperatively). Active patients are defined by a Lower Extremity Activity Scale (LEAS) level of 13–18. Smokers have a higher percentage of any wound complication (3.8% vs. 1.8%), deep infection (2.5% vs. 1.0%), pneumonia (1.3% vs. 0.4%,), and reoperation (5.0% vs. 3.1%) compared with nonsmokers undergoing RTKA. Patients who abuse drugs are at augmented risk for RTKA. Amphetamine users have the fastest mean time to revision (691 days). At 30 days, 90 days, and 6 months postoperatively, cocaine users have the highest rate of requiring RTKA (7%, 12%, and 20%, respectively); at 1 year, it is alcohol abusers (38%). PJI is the most common cause of RTKA in all drug abuse/drug-dependent groups.