© Springer International Publishing AG 2018E. Carlos Rodríguez-Merchán and Sam Oussedik (eds.)The Infected Total Knee Arthroplastydoi.org/10.1007/978-3-319-66730-0_3
3. Preoperative Optimization to Prevent an Infected Total Knee Arthroplasty: Host Factors
Specialist Registrar, University College London Hospital, London, UK
Department of Orthopaedic Surgery, University College Hospital, 235 Euston Road, London, NW1 2BU, UK
A large number of host risk factors have been identified with an increasing number of papers demonstrating many times higher PJI rates in these patients. Some may be easier and faster to address than others. Patients who present with multiple risk factors need to have a clear understanding of the morbidity involved with PJI, and as clinicians there needs to be a planned approach in how to identify which of these conditions can be optimized and an appropriate time frame set to achieve this.
With a rise in obesity and diabetes, we may start to see an associated increase in PJI. Without good evidence to support bariatric surgery and many patients struggling with weight loss, it may be that the additional risk needs to be accepted. Identification of an early arthritic group who could be targeted for weight loss would seem logical.
When presented with an obese, diabetic patient with multiple other risk factors, there may be a point when optimization has been achieved, but the considerable increased risk of PJI remains. Further research will need to better quantify this risk in order to educate and consent patients, better prepare health services for the likelihood of complications arising, and ensure registry data is appropriately adjusted for risk.
Preoperative assessment of patients to identify groups at higher risk of knee prosthetic joint infection (PJI) is essential. Joint registry data looking at 56,216 total knee arthroplasties (TKAs) recorded deep surgical site infection (SSI) in 0.72% (404/56,216) and associated these with a body mass index (BMI) of ≥ 35 kg/m2 (hazard ratio (HR) = 1.47), diabetes mellitus (DM) (HR = 1.28), and American Society of Anesthesiologists (ASA) score of > – 3 (HR = 1.65) . When using data from a meta-analysis, similar findings are seen; of 12 cohorts or case-control studies which included 548 infected patients in 57,223 cases, BMI (BMI > 30: odds ratio (OR) = 2.53, 95% CI 1.25, 5.13; BMI > 40: OR = 4.00, 95% CI 1.23, 12.98), DM (OR = 3.72, 95% CI 2.30, 6.01), hypertension (OR = 2.53, 95% CI 1.07, 5.99), steroid therapy (OR = 2.04, 95% CI 1.11, 3.74), and rheumatoid arthritis (OR = 1.83; 95% CI 1.42, 2.36) were identified risk factors . These conditions and diseases, as well as others, are well known to affect tissue viability and healing; they need to be diagnosed, treated, and optimized as part of the preparation for joint replacement. This allows their effect to be minimized and the patient to be appropriately counseled.
Patient weight is a significant factor in the progression of knee osteoarthritis; it is therefore not surprising that a substantial number of patient undergoing TKA are obese . For clinical and research purposes, patients are grouped into weight categories based on their BMI. Those with BMI over 30 kg/m2 are termed obese and those over 40 kg/m2 are morbidly obese. Health economic studies now indicate that there is a worldwide increase in obesity levels , and this is reflected in TKA patients, with over 50% obese [4, 5].
Authors have broadly chosen two ways to look at obesity in TKA: reviewing all PJIs and looking at their demographic or looking at BMI and following these groups to see if they have a higher infection rate than a matched group.
A 2012 meta-analysis and systematic literature review identified 14 studies and 15,276 patients; overall infection occurred more in obese patients (OR = 1.90, 95% CI 1.46–2.47), and deep infection requiring surgical debridement was reported as increased in 9 studies including 5061 patients (OR = 2.38, 95% CI 1.28–4.55) .
Numerous studies have shown a link between obese patients experiencing longer operative times , worse knee score outcomes , increased overall complication rates [9, 10], and significantly both superficial and deep postoperative infections [4–6].
An association between an even higher risk of PJI relative to an increase in BMI from >30 to >40 (morbidly obese) has also been demonstrated [11, 12].
A single-center study from Finland saw an overall infection rate of 0.79% in 3915 knee replacements; infected patients had a significantly higher median BMI (31.5 compared with 28.7 kg/m2; p = 0.001). Six of the 156 (3.85%) morbidly obese patients developed PJI . The same unit published a review of their experience . A further database review of 48 PJIs from a cohort of 4185 (1.1%) found that a BMI >40 kg/m2 was an independent predisposing factor for PJI (OR = 3.23 95%, CI 1.6–6.5 p = 0.001) . Findings from 41 consecutive TKAs in morbidly obese compared with a matched group with BMI <30 resulted in seven superficial infections and two deep versus no infections .
There is evidence to suggest that the higher the BMI, the greater your increased risk, with a complication rate including wound healing and infection higher in patients with a BMI >45, and with the risk rising for every 5 unit increase in BMI up to 70 . A number of studies have looked at this “super” obese category; from a group of 8494 total joint arthroplasty (TJA) patients, 30 deep infections were identified and found that a BMI > –50 had a massively increased odds ratio of infection of 21.3 (p = 0.001) , and 29 PJIs from a cohort of 1846 (1.5%) also found that a BMI > 50 was an independent risk factor (OR = 5.28, CI 1.38–17.1) . A study following the outcome of 101 matched TKAs in patients with a BMI of at least 50 found a variety of wound problems: 6 patients with wound necrosis, 1 superficial infection and 2 wound hematomas, and overall a 3.1 times higher odds ratio (95% CI 1.07–8.9) of complications compared with the matched group (p = 0.037) .
With such a large number of patients undergoing TKA, providing some evidence that they can experience good outcomes is essential. Survivorship of implants when comparing patients with either a BMI < 30 or > 30 was equivalent in a group of 535 . A large database review of over 90,000 patients with a BMI > 40 who underwent TKR found that there was an independent association with a higher risk of a small number of select in-hospital postoperative complications and mortality after matching with comorbid medical conditions linked to obesity, such as type 2 diabetes and hypertension. However, the independent impact of obesity appeared to be fairly modest .
3.2.1 Preoperative Weight Loss and Bariatric Surgery
With the increased rate of PJI seen, it would seem logical to recommend that patients undergo weight loss during the period of knee arthritis prior to TKA. How far this should be taken, and whether this should be a prerequisite to offering surgery is currently unknown. A study looking at 10,718 obese patients undergoing TKA over a 3-year period found no difference in SSI between those who gained or lost weight (decrease in 5% body weight) preoperatively .
The medical and social indications for bariatric surgery are broad and need to be taken into consideration when a patient is also indicated for TKA. A computer-based model reviewing outcomes from studies and registry data and using quality-adjusted life years and cost-effectiveness as an endpoint supported bariatric surgery prior to knee arthroplasty in obese patients .
Clinical studies however have resulted in conflicting conclusions [24–28]. A systematic review and meta-analysis concluded that, based on this moderate quality of evidence, an improvement in clinical outcome with pre-arthroplasty bariatric surgery needs to be questioned . After bariatric surgery, patients are at risk of nutritional deficiencies  which subsequently may also alter their susceptibility to PJI.
Malnutrition is a deficiency, excess, or imbalance of body protein and other nutrients. It causes adverse effects on tissue, function, and clinical outcome, with prolonged inflammation by reducing collagen synthesis and fibroblast proliferation. It may additionally impair the immune system through lymphocyte proliferation, antibody responses, interleukin-2, and interferon-gamma, as well as delayed hypersensitivity reactions . It can be seen in patients with different body weights and be related to an underlying condition or chronic disease. It has been seen to be prevalent in patients undergoing TJA .
Blood parameters that have been identified as markers of malnutrition and are subsequently studied are serum albumin (<3.5 g/dL), total lymphocyte count (<1500 cells/mm3), and transferrin (<200 mg/dL). Several studies have shown the adverse effects malnutrition can have when undergoing TJA [32–35].
In 1991, a study found that TKA patients had a five times greater risk of developing major wound complications if their preoperative total lymphocyte count was <1500 cells/mm3 and a seven times greater risk if their albumin was <3.5 g/dL . A number of studies since then have also shown increased complication rates and PJIs. A preoperative serum albumin < 3.5 g/dL has been demonstrated to increase overall complication rate following TKA including SSI . Two studies have analyzed large numbers of patient from ACS-NSQIP data. The first found significantly higher rates of SSI and pneumonia, an extended length of stay, and readmission in the 4% of patients with hypoalbuminemia . The second showed higher superficial SSI, organ space SSI, and deep SSI. These patients also had a higher rate of postoperative pneumonia, acute renal failure, cardiac arrest requiring pulmonary resuscitation, septic shock, unplanned intubation, and blood transfusion. Their overall risk was calculated to be higher when compared to morbidly obese patients, and when compared to patients with normal albumin, there was a 3.19-fold increased risk of mortality . A prospective study found that patients with hypoalbuminemia also had an increased chance of unplanned intensive care admission postoperatively: 15.4 and 3.8% in patients with a serum albumin of <3.0 and <3.5 g/dL, respectively .
Like biochemical parameters, triceps skinfold (TSF) can be used as a marker of nutritional status. A prospective study of 213 TKAs where 11 patients subsequently became infected (5 deep and 6 superficial) found this to be the only significant risk factor for postoperative infection risk .
These findings have been seen in studies on revision TKA, with malnutrition independently associated with chronic septic failure requiring further surgery [39–42].
Malnutrition can be seen in underweight patients but is also common in obese patients and is likely more significant . When looking at low patient BMI, a study which included 1315 TKA patients found that while there was an increase rate of wound hematomas and seromas, and postoperative anemia was more frequent, this patient group had in fact demonstrated a lower rate of PJI .
Given the findings of the studies described, it would be logical to identify malnourished patients and correct this preoperatively. Benefits of perioperative nutritional support have been seen in an elderly fracture neck of femur population [45, 46], but there are no studies to date specifically looking at the benefit of this in TJA patients. Nutritional optimization is an area of future research with a recent randomized pilot study demonstrating a decreased length of stay and postoperative C-reactive protein (CRP) in total hip arthroplasty (THA) patients who received multimodal perioperative care and immunonutrition .
Preoperative anemia is reported to occur in 15–33% of patients undergoing TJA . Anemia is associated with a large number of comorbid disorders and it is therefore unsuprising that, when present preoperatively, there is also an association with postoperative complications following TJA .
A review of 6371 TKA patients found 1249 to be anemic. This study also included 7230 THA patients with 1286 anemic in this group. The results combined both arthroplasty groups and subsequently demonstrated an increase in PJI when compared to a matched group with normal hemoglobin levels preoperatively (4.3 versus 2%) .
There is a decrease in the rate of postoperative allogenic blood transfusion if anemia is corrected [14, 51], and with an association between transfusion and SSI, this may also be a relevant factor . Preoperative correction will depend on the cause of the anemia, with iron replacement and recombinant human erythropoietin therapy proven to be useful options after discussion with a hematologist [53–55].
3.5 Diabetes Mellitus
The rise in obesity levels has also seen an associated increase in type 2 DM [56, 57]. With this rise comes a projected increase in PJI . The effect of DM in wound healing is well known, and more focused research into patients with DM undergoing TJA has shown in general that these patients are at increased risk of wound complications and deep infection [58, 59], as well as a variety of other complications such as deep vein thrombosis, hospital cost, readmission, and mortality [60–66].
Analysis of 3915 TKA patients saw an infection rate increase (from 0.66%, 95% CI 0.43–0.99% to 1.59%, 95% CI 0.84–2.99%) in those diagnosed with DM in comparison with those without. This was independent of obesity. When combined with morbid obesity, that figure rose to a PJI in 5 out of 51 patients (9.8%, 95% CI, 4.26–20.98%). Patients not diagnosed with diabetes but presenting with a blood glucose of >6.9 nmol/L also had a higher PJI rate (1.15%, 95% CI, 0.56–2.35% compared with 0.28%, 95% CI 0.15–0.53%) suggesting that this is an important part of the preoperative assessment . A retrospective review of 167 TKAs in type 2 DM patients found an overall wound complication rate of 6.6% (n = 11), and logistic regression revealed an HbA1c > –8 as an independent risk factor . A prospective review of 1214 TKAs found an overall infection rate of 1.5% with an increase in diabetic patients (OR = 6.87; 95% CI 2.42–19.56), and significantly in this study, there were no PJIs in patients with diabetes who were not also obese . In a Korean study comparing 222 TKAs in patients with DM versus a matched group, a wound complication rate of 9.5% compared to 5% (p < 0.05) was seen, and the 1 superficial infection and 2 deep infections did not reach significance .
The view that all DM patients are at an increased risk of PJI is not universal, however. In a large multicenter retrospective cohort study, 7567 (18.7%) DM patients were identified from a joint registry cohort of 40,491 TKAs. From the 287 patients who developed deep infection, there was no significant increased risk in the DM group, even when adjusting for BMI. The study went on to look at patient’s diabetic control; no significant difference could be found when subdividing DM patients into uncontrolled (HbA1c <7%) and controlled (HbA1c >7%) .
Analysis of HbA1c gives a long-term record of a patient’s diabetic control. Whether it is a useful marker in TKA and subsequent PJI is unclear. A study found no correlation between HbA1c and infection rates although in this study the overall deep PJI was higher in the DM group versus normal controls: 2.6 versus 0.87% .
With a large number of patients with DM undergoing TJA, perioperative control of blood glucose levels has also been studied. In 1565 TKAs, a rate of PJI during the 1-year follow-up of 0.44, 0.93, and 2.42% was recorded in groups with preoperative plasma glucose of <6.1 mmol/L, 6.1–6.9 mmol/L, and >–7.0 mmol/L, respectively. The patients with highest glucose levels had a fourfold risk for infection when compared to patients with the lowest glucose . A further study demonstrated a twofold increase in PJI in patients with DM who had postoperative hyperglycemia and interestingly a rate three times higher in non-DM patients with hyperglycemia representing a physiological response to surgery .
3.6 Other Comorbidities
This chapter has highlighted a number of specific conditions that have been studied and subsequently identified as risk factors for PJI. It is often the combination of risk factors that place a patient at the highest risk. It is therefore useful to try and quantify the significance of comorbidities. A review of 83,011 patients undergoing TKR looked at a broad range of disorders and found (in decreasing order of significance) congestive heart failure, chronic pulmonary disease, preoperative anemia, DM, depression, renal disease, pulmonary circulation disorders, obesity, rheumatologic disease, psychoses, metastatic tumor, peripheral vascular disease, and valvular disease as risk factors for PJI . In a 2001 single-center study of 6489 TKAs, 116 PJIs became infected. Of the 12 identified significant comorbidities, prior open surgical procedure, immunosuppressive therapy, hypopotassemia, and poor nutrition were the 4 most significant factors (each p < 0.001) .
The American Society of Anesthesiologists (ASA) is accepted as an overall summary of patient’s preoperative risk and encompasses all comorbidities. A review of 4185 TKAs identified 48 (1.1%) PJIs and after univariate analysis found ASA > 2 to be an independent factor (p = 0.002) and a more significant one than morbid obesity (p = 0.03) .
3.6.1 Inflammatory Joint Disease and Immunosuppression
Patients with inflammatory joint disease may have a number of different reasons to have an increased rate of PJI including overall disability, disease activity, and the use of disease-modifying antirheumatic drugs (DMARDs). With over 80% of rheumatoid arthritis (RA) patients now prescribed with DMARDs  and 24% undergoing large joint replacement , optimal perioperative management of these medications will become an increasing challenge.
In a health insurance database case-control study of 2212 patients identified with a PJI, risk factors were established; those with a diagnosis of RA (OR = 1.47, p = 0.031) while identified were not as significant as obesity (OR = 1.66, p < 0.001) or DM (OR = 1.38, p = 0.001). A prescription for prednisolone had a significantly increased risk (OR = 1.59, p < 0.001), while no significant difference was found for the use of DMARDs including tumor necrosis factor-alpha (anti-TNFa) inhibitors .
The decision to stop any DMARDs needs to be a balance between the risk of a perioperative disease flare and risk of PJI, and there is no current consensus on the safety of perioperative continuation of these medications . A 2009 systematic review on the perioperative use of methotrexate concluded that low doses seem to be safe without a significant increase risk of PJI and an avoidance of disease flare . A further controlled study with a 10-year follow-up again found no PJIs in either a group taking perioperative methotrexate or one which stopped (n = 65) .
Review articles also looking at anti-TNFa and other biologics have concluded that despite inconclusive data, the trend was to an increased risk of PJI. The advice currently is to withhold anti-TNFa and other biologics prior to surgery based on the dosing interval and continue methotrexate and hydroxychloroquine through the perioperative period. There was no consensus regarding medications such as leflunomide and rituximab [82–84]. Without compelling evidence and with new medications and drug combinations being prescribed, a perioperative plan needs to be made with guidance from local rheumatologists.
3.6.2 Glucocorticoid Steroids
There is a strong association between systemic glucocorticoid therapy and the risk of infection, demonstrated in a 2011 systematic review and meta-analysis  with higher doses conferring greater risk. There is evidence that perioperative supplemental doses to counter the adrenal insufficiency precipitated by surgical stress are not required [86–88].
Intra-articular injection of steroids prior to TKA was not shown to increase either deep or superficial infection in systematic review and meta-analyses [89, 90]. A review of a national database of TKAs found a significant difference in PJI between patients administered with intra-articular steroid within 3 months (2.6%, OR = 1.6–2.5, p < 0.0001) and 6 months (3.41%, OR = 1.2–1.8, p < 0.0001) of TKA , suggesting that there is a period of time where steroid concentration is high enough to be a risk factor.
3.6.3 Human Immunodeficiency Virus (HIV)
With the advent of antiretroviral medication, patients with HIV can mount an immune response. A 2001 study of 28 patients with HIV saw 4 (14%) develop a PJI, 2 of these patients also had a history of intravenous drug use .
3.6.4 Renal, Gastrointestinal, and Hepatic Disorders
Optimization of organ function with particular emphasis on how function may be affecting anemia and nutritional status has been recommended . Using data from the Scottish Arthroplasty Project, an analysis of 59,288 TKAs found 652 PJI infections within 90 days. There was a significant association with those diagnosed with renal failure (p = 0.002), and dialysis prior to TKA appeared to eliminate this risk .
There are numerous studies linking smoking with wound healing and infection rates in surgery. A 2012 systematic review and meta-analysis found 140 cohort studies including 479,150 patients on the subject. The conclusion was that postoperative healing complications occur significantly more often in smokers compared with nonsmokers and in former smokers compared with those who never smoked. Perioperative smoking cessation intervention reduces surgical site infections, but not other healing complications .
A 2011 systematic review looking at outcomes in smokers after TKA and THA found they had a significantly higher rate of reoperation or revision, implant loosening, deep infection, and mortality compared to those who did not smoke . When focusing on TKA patients, studies have shown an increase in hospital length of stay and intensive care unit admissions  and revision rates, including PJI . A history of smoking was one of the significant risk factors identified of the 116 infected knees from a cohort of 6489 , and analysis from ACS-NSQIP data showed a substantially higher risk of morbidity and mortality with an increase in complications as a dose-dependent relationship with the number of pack-years .