PDF Revision Total Knee Arthroplasty for Periprosthetic Joint Infection

Revision Total Knee Arthroplasty for Periprosthetic Joint Infection Is Associated With Increased Postoperative Morbidity and Mortality Relative to Noninfectious Revisions

Venkat Boddapati, BA

, Michael C. Fu, MD, MHS, David J. Mayman, MD, Edwin P. Su, MD, Peter K. Sculco, MD, Alexander S. McLawhorn, MD, MBA

Adult Reconstruction and Joint Replacement Service, Hospital for Special Surgery, New York, New York

a r t i c l e i n f o

Article history:

Received 8 August 2017 Received in revised form 30 August 2017

Accepted 6 September 2017 Available online xxx Keywords:

total knee arthroplasty revision periprosthetic joint infection ACS-NSQIP

infectious complications knee arthroplasty knee replacement

a b s t r a c t

Background: Periprosthetic joint infection (PJI) after primary total knee arthroplasty (TKA) is a devas- tating complication. The short-term morbidity profile of revision TKA performed for PJI relative to non-PJI revisions is poorly characterized. The purpose of this study is to determine 30-day postoperative outcomes after revision TKA for PJI, relative to primary TKA and aseptic revision TKA.

Methods: The American College of Surgeons National Surgical Quality Improvement Program from 2005 to 2015 was queried for primary and revision TKA cases. Revision TKA cases were categorized into PJI and non-PJI cohorts. Differences in 30-day outcomes including postoperative complications, readmissions, operative time, and length of stay were compared using bivariate and multivariate analyses.

Results: In total, 175,761 TKAs were included in this study, with 162,981 (92.7%) primary TKAs and 12,780 (7.3%) revision TKAs, of which 2196 (17.2%) revisions were performed for PJI. When compared to aseptic revision TKA, multivariate analysis demonstrated that PJI revisions had a significantly higher risk of major early postoperative complications including death (adjusted odds ratio [OR] 3.25) and sepsis (OR 8.73). In addition, nonhome discharge (OR 1.75), readmissions (OR 1.67), and length of stay (þ2.1 days) were all greater relative to non-PJI revisions.

Conclusion: Utilizing a large, prospectively collected, national database, we found that revision TKA for PJI has a greater risk of short-term morbidity and mortality and requires a higher utilization of healthcare resources. These results have implications for patient counseling and alternative payment models that may eventually include revision TKA.

©2017 Elsevier Inc. All rights reserved.

Primary total knee arthroplasty (TKA) has experienced a 231%

growth in volume between 2000 and 2010, with an estimated 680,886 cases performed in 2014 in the United States[1]. Outcomes after TKA are well characterized. Short-term complication rates range between 5.0% and 7.6%[2e4], mortality under 0.5%[2,3,5], and 30-day readmission rates between 3.3% and 4.6% in United

Statesebased hospitals [6e8]. Patient age, obesity, frailty, and smoking history among other risk factors have been associated with increased short-term complications[9e11].

Along with the rise in primary TKA cases, there has been a concomitant increase in revision TKA cases performed annually with the overall 10-year revision rate between 4.9% and 7.8%[12].

Relative to primary TKA, revision TKA is associated with more short-term complications; however, the morbidity profile for revision TKA is less well characterized compared to primary TKA [13]. In addition, there are little data on how revisions for peri- prosthetic joint infection (PJI) compare to aseptic revision TKA in regards to the postoperative morbidity, mortality, and healthcare utilization. A retrospective study of 3421 patients undergoing revision TKA showed a high incidence of short-term postoperative complications, with 5.5% of patients experiencing a major compli- cation such as respiratory failure and 2.2% experiencing minor One or more of the authors of this paper have disclosed potential or pertinent

conflicts of interest, which may include receipt of payment, either direct or indirect, institutional support, or association with an entity in the biomedicalfield which may be perceived to have potential conflict of interest with this work. For full disclosure statements refer tohttps://doi.org/10.1016/j.arth.2017.09.021.

Investigation performed at the Hospital for Special Surgery, New York, NY.

*Reprint requests: Venkat Boddapati, BA, Adult Reconstruction and Joint Replacement Service, Hospital for Special Surgery, 535 East 70th Street, New York, NY 10021.

Contents lists available atScienceDirect

The Journal of Arthroplasty

jo u rn a l h o m e p a g e :w w w . a r t h r o p l a s t y j o u r n a l . o r g

https://doi.org/10.1016/j.arth.2017.09.021 0883-5403/©2017 Elsevier Inc. All rights reserved.

complications such as urinary tract infection[14]. Revision TKA also has an estimated 6.2% readmission rate, significantly higher than primary TKA[15]. The most common indication for revision TKA is aseptic loosening, followed by PJI, instability, periprosthetic frac- ture, and arthrofibrosis[16,17]. However, PJI is the most common indication for revision arthroplasty within thefirst 2 years after primary TKA[16].

PJI is associated with significant morbidity and healthcare costs [18e21], due to prolonged hospitalization and the potential need for multiple operations. The management of PJI is complex and requires an interdisciplinary team of medical personnel, orthopedic surgeons, and infectious disease specialists. Accurate identification of the infectious organism and its antibiotic sensitivities along with aggressive surgical debridement of the involved soft tissues and bone improves the success of infection eradication.

The definition of acute PJI is symptom onset within 3-6 weeks after primary TKA or after hematogenous seeding and can be treated with debridement, prolonged systemic antibiotic treat- ment, and implant retention strategies, usually with exchange of modular parts (polyethylene insert). The success of eradicating joint infection by this approach is variable and ranges between 31%-78%[22,23]. Chronic PJI is defined by a duration greater than 3-6 weeks from primary TKA, and treatment consists of debridement, antibiotics, and either 1-stage or 2-stage exchange arthroplasty of both the femoral and tibial components[23e25].

Previous studies assessing the outcomes of acute and chronic PJI have been limited by small sample size (<50 patients)[26e30], were from a single institution[27e33], or did not fully charac- terize short-term postoperative complications (eg, sepsis) or fully assess healthcare utilization in regards to length of stay (LOS) or readmissions[29,30,32].

The purpose of this study is to determine short-term compli- cation and readmission rates, operative time, and postoperative hospital LOS for revision TKA due to PJI relative to primary TKA and non-PJI revision TKA using a large, prospectively collected registry.

The secondary aim of this study is to determine if baseline patient characteristics were associated with adverse outcomes in patients undergoing revision arthroplasty for PJI.

Materials and Methods

Data from the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) from 2005 to 2015 were used in this study. ACS-NSQIP is a large multicenter registry that is prospectively collected by trained surgical clinical reviewers.

Rigorous collection of data and auditing of the data by the ACS has led to high inter-rater reliability [31,34]. In 2015, data were collected from a diverse set of over 600 participating institutions ranging from small community hospitals to large academic medical centers [35]. Data collected include preoperative demographical information and risk factors, perioperative events (eg, blood transfusion), and postoperative events up to 30 days after surgery, regardless of discharge date.

Three patient cohorts were defined in this study: (1) primary TKA, (2) revision TKA for non-PJI diagnoses, and (3) revision TKA for PJI. The above cohorts were defined based on primary Current Procedural Terminology (CPT) codes and operative International Classification of Disease, Ninth and Tenth edition diagnosis codes. Patients with primary CPT code 27447 with a diagnosis code for noninfectious etiology were included in the primary TKA group. The non-PJI revision TKA group included patients with CPT codes 27486, 27487, and 27488 with a noninfectious opera- tive diagnosis code. The PJI revision TKA cohort was defined as patients with primary CPT codes 27447, 27486, 27487, and 27488 with an infectious diagnosis code, as those with primary CPT

27447 with an infectious diagnosis code were presumably un- dergoing the second stage of a 2-stage revision for PJI. Emergency cases, primary TKA cases with nonclean wound classifications, and cases with missing patient demographic variables were excluded from this study.

Baseline patient variables included patient height and weight, and calculated body mass index (BMI). BMI was then stratified according to the World Health Organization classification. Addi- tionally, anesthesia type, preoperative functional status, gender, American Society of Anesthesiologists (ASA) class, history of smoking, diabetes, chronic obstructive pulmonary disease (COPD), and hypertension were collected. Anesthesia type was dichoto- mized into general or regional anesthesia, with regional anesthesia consisting of spinal, epidural, or other regional techniques.

Recorded 30-day postoperative complications included death, cardiac complications (consisting of myocardial infarction or car- diac arrest requiring cardiopulmonary resuscitation), renal com- plications (acute renal failure or progressive renal insufficiency), pulmonary complications (failure to wean from ventilator within 2 days postoperatively, unplanned reintubation, or pneumonia), deep vein thrombosis or pulmonary embolism, sepsis, deep surgical site infections (SSI), superficial SSI, wound dehiscence, urinary tract infection, and blood transfusions. Readmission within 30 days, operative time, and postoperative hospital LOS were also assessed.

Of note, readmission data were only collected from 2011 to 2015 in the NSQIP, and available for 90.1% of patients included in this study.

Cases prior to 2011 were excluded in all readmission analyses.

Statistical analysis comparing baseline patient characteristics was performed using Pearson’s chi-squared analysis. Postoperative outcomes and readmission rates between primary TKA and revision TKA were compared using bivariate analysis using Pearson’s chi-squared, as well as multivariate binary logistic regression con- trolling for baseline patient characteristics including age, gender, BMI, ASA class, functional status, anesthesia type, and comorbidities including hypertension, COPD, smoking history, and diabetes mel- litus. Multivariate binary logistic regression was also used to compare postoperative complications and readmission rates between PJI revision TKA relative to primary TKA and non-PJI revi- sion TKA. Finally, operative time and postoperative LOS were compared with multivariate linear regressions that controlled for baseline patient characteristics and compared revision TKA subgroups to primary TKA, as previously described by Basques et al [36]. Statistical significance was defined asP<.05; however, Bon- ferroni correction was applied to all baseline patient characteristic and outcome comparisons to correct for multiple-group analyses.

Statistical analysis was performed with SPSS version 23 (IBM Corp., Armonk, NY).

Results

In total, 175,761 patients were included in this study after application of inclusion and exclusion criteria. Of these, 162,981 (92.7%) patients underwent primary TKAs and 12,780 (7.27%) un- derwent revision TKAs (Table 1). Relative to primary TKA, patients undergoing revision TKA were more likely to be younger, male, have a higher BMI, ASA class, dependent functional status, undergo a procedure with general anesthesia, and have more medical comorbidities including diabetes mellitus, cigarette smoking, and COPD (P<.001 for all comparisons).

Complications and Readmissions After Revision TKA

The rate of any complication was 3.54% among primary TKA and 6.26% among revision TKA (Table 2; adjusted odds ratio [OR]

1.66). With multivariate analysis, there was also an increased

rate of death (OR 1.89), respiratory complications (OR 1.40), renal complications (OR 2.06), sepsis (OR 5.45), deep SSI (OR 5.12), and wound dehiscence (OR 1.83) in the revision group.

Revision TKA patients were also more likely to have a blood transfusion (OR 1.68) and be readmitted within 30 days (6.11%), compared to the primary TKA group (3.44%, OR 1.69,P<.001 for all comparisons).

Complications and Readmissions After Revision due to PJI and Non-PJI

Of the 12,780 revision cases, 10,785 (82.8%) cases were for non-PJI and 1456 (17.2%) cases were for PJI indications (Table 3).

Compared to primary TKA, the rate of any complication was significantly increased in non-PJI revisions at 4.72% (OR 1.26) and in PJI revisions at 13.66% (OR 3.64). In both revision groups, there was also an increased risk of sepsis (OR_nonPJI 2.27, OR_PJI 19.9), deep SSI (ORnonPJI3.74, ORPJI 11.0), wound dehiscence (ORnonPJI

1.68, ORPJI2.42), blood transfusions (ORnonPJI1.44, ORPJI3.08), and hospital readmission (OR_nonPJI1.51, OR_PJI2.57) relative to primary TKA. Additionally, the PJI revision group had an increased risk of death (OR 5.00), respiratory complications (OR 2.68), renal complications (OR 5.3), and nonhome discharge (OR 1.67) that were not seen in the non-PJI revision group (P < .001 for all comparisons).

Table 1

Comparison of Patient and Operative Characteristics Between Primary and Revision Knee Arthroplasty.

Number of Patients All Patients, 175,761

Primary TKA, 162,981

Revision TKA P-Value

All, 12,780 Non-PJI, 10,584

PJI, 2196 Primary vs All Revisions

Primary vs Non-PJI

Primary vs PJI

Age <.001^a <.001^a <.001^a

<60 24.0% 23.6% 29.6% 30.1% 27.1%

60-70 40.9% 41.2% 38.0% 37.9% 38.5%

71-80 27.0% 27.2% 24.1% 24.0% 25.0%

>80 8.0% 8.0% 8.2% 8.0% 9.3%

Female % 62.0% 62.3% 58.0% 60.5% 46.3% <.001^a <.001^a <.001^a

Body mass index (kg/m²) <.001^a <.001^a <.001^a

Non-obese (<30) 37.6% 37.6% 37.4% 36.8% 40.4%

Obese I (30-34.9) 28.4% 28.4% 27.8% 28.1% 26.1%

Obese II (35-39.9) 18.8% 18.8% 18.1% 18.4% 16.3%

Obese III (>40) 15.2% 15.1% 16.8% 16.7% 17.1%

Comorbidities

Diabetes mellitus 17.6% 17.3% 20.5% 19.7% 24.1% <.001^a <.001^a <.001^a

Smoking history 8.86% 8.62% 11.9% 11.8% 12.0% <.001^a <.001^a <.001^a

COPD 3.7% 3.6% 5.4% 5.2% 6.6% <.001^a <.001^a <.001^a

Hypertension 65.9% 65.8% 66.7% 66.2% 69.0% .044 .401 .002^a

Anesthesia type <.001^a <.001^a <.001^a

General 52.3% 51.5% 63.1% 61.0% 73.7%

Regional 47.7% 48.5% 36.9% 39.0% 26.3%

Functional status <.001^a <.001^a <.001^a

Independent 98.2% 98.4% 95.8% 96.6% 92.1%

Dependent 1.8% 1.6% 4.2% 3.4% 7.9%

ASA class <.001^a <.001^a <.001^a

I 2.1% 2.1% 1.3% 1.4% 0.7%

II 49.5% 50.3% 39.5% 41.8% 28.5%

III 46.7% 46.0% 55.7% 53.9% 64.4%

IV 1.7% 1.6% 3.5% 2.9% 6.4%

aSignificance defined asP<.0045 after Bonferroni correction, significant values are in bold.

Table 2

Comparison of Adverse Outcomes After Primary and Revision Knee Arthroplasty.

Primary TKA, 162,981

Revision TKA, 12,780

Bivariate Analysis Multivariate Analysis^a

OR P-Value OR P-Value

Any complication (major or minor) 3.54% 6.26% 1.77 <.001^b 1.66 <.001^b

Major complication 2.08% 4.50% 2.17 <.001^b 1.98 <.001^b

Death 0.11% 0.27% 2.48 <.001^b 1.89 .001^b

Cardiac complications 0.27% 0.40% 1.46 .011 1.22 .185

Renal complications 0.18% 0.46% 2.54 <.001^b 2.06 <.001^b

Respiratory complications 0.46% 0.80% 1.73 <.001^b 1.40 .002^b

Deep vein thrombosis 0.92% 0.94% 1.03 .789 1.00 .970

Stroke/cerebrovascular accident 0.09% 0.09% 1.09 .767 1.02 .940

Sepsis 0.22% 1.38% 6.36 <.001^b 5.45 <.001^b

Deep surgical site infection 0.15% 0.90% 5.87 <.001^b 5.12 <.001^b

Minor complication 1.65% 2.10% 1.27 <.001^b 1.20 .007

Urinary tract infection 0.93% 0.97% 1.05 .614 1.01 .881

Wound dehiscence 0.20% 0.42% 2.10 <.001^b 1.83 <.001^b

Superficial surgical site infection 0.55% 0.76% 1.37 .003 1.25 .040

Nonhome discharge 28.01% 30.03% 1.07 <.001^b 1.04 .091

Blood transfusion 8.16% 13.94% 1.71 <.001^b 1.69 <.001^b

Thirty-day readmission 3.42% 6.30% 1.84 <.001^b 1.69 <.001^b

aBinary logistic multivariate regression.

b Significance defined asP<.0029 after Bonferroni correction, significant values are in bold.

With further subanalysis comparing PJI revisions to the non- PJI revision cohort, multivariate analysis revealed that PJI revisions had an increased risk of any complication (OR 2.81), death (OR 3.25), respiratory complications (OR 2.43), renal complications (OR 5.08), sepsis (OR 8.73), deep SSI (OR 3.16), blood transfusions (OR 2.07), nonhome discharge (OR 1.75), and hospital readmissions (OR 1.67) relative to non-PJI revision TKA (P.001 for all comparisons).

Predictors of Adverse Outcomes in PJI Revision TKA

COPD was found to be an independent predictor of any complication (OR 2.08,P¼.001). General anesthesia was associated with an increased risk of both sepsis (OR 2.53, P ¼.001) and

nonhome discharge (OR 1.47,P¼.001) (Table 4). Additional pre- dictors of nonhome discharge were older age, female gender (OR 1.37,P¼.002), BMI greater than 40 kg/m²(OR 1.77,P<.001), ASA class III (OR 1.66,P<.001), and ASA class IV (OR 2.44,P<.001). ASA class III (OR 1.87,P¼.003) and IV (OR 2.85,P¼.001) also increased the risk of hospital readmission.

Operative Time and Postoperative Length of Stay

Operative time was increased in revision TKA compared to pri- mary TKA (þ38.4 minutes, P< .001), as was postoperative LOS (þ0.6 days,P<.001) (Table 5). Non-PJI revisions had an increased operative time (þ37.5 minutes, P < .001) and LOS (þ0.2 days, P<.001), as did PJI revisions (þ43.2 minutes,þ2.4 days,P<.001) Table 3

Multivariate Analysis Comparing Adverse Outcomes After Primary, PJI, and Non-PJI Revision Knee Arthroplasty.

Primary TKA, 162,981

Revisions Multivariate Analysis^a

Non-PJI, 10,584

PJI, 2196 Non-PJI vs Primary PJI vs Primary PJI vs Non-PJI

OR P-Value OR P-Value OR P-Value

Any complication (major or minor) 3.54% 4.72% 13.66% 1.26 <.001^b 3.64 <.001^b 2.81 <.001^b

Major complication 2.08% 3.00% 11.7% 1.35 <.001^b 5.00 <.001^b 3.64 <.001^b

Death 0.11% 0.16% 0.16% 1.24 .398 4.30 <.001^b 3.25 .001^b

Cardiac complications 0.27% 0.35% 0.64% 1.14 .457 1.48 .156 1.52 .196

Renal complications 0.18% 0.25% 1.50% 1.15 .497 5.83 <.001^b 5.08 <.001^b

Respiratory complications 0.46% 0.58% 1.87% 1.07 .600 2.68 <.001^b 2.43 <.001^b

Deep vein thrombosis 0.92% 0.85% 1.37% 0.92 .422 1.42 .060 1.49 .066

Stroke/cerebrovascular accident 0.09% 0.09% 0.09% 1.07 .834 0.85 .819 0.68 .629

Sepsis 0.22% 0.55% 5.37% 2.27 <.001^b 19.9 <.001^b 8.73 <.001^b

Deep surgical site infection 0.15% 0.64% 2.14% 3.74 <.001^b 11.0 <.001^b 3.16 <.001^b

Minor complication 1.65% 1.98% 2.64% 1.14 .068 1.44 .007 1.27 .128

Urinary tract infection 0.93% 0.90% 1.32% 0.94 .588 1.36 .107 1.43 .103

Wound dehiscence 0.20% 0.38% 0.64% 1.68 .002^b 2.42 .002^b 1.67 .113

Superficial surgical site infection 0.55% 0.76% 0.77% 1.26 .049 1.19 .482 0.96 .878

Nonhome discharge 28.01% 27.73% 41.53% 0.94 .010 1.67 <.001^b 1.75 <.001^b

Blood transfusion 8.16% 11.87% 23.95% 1.44 <.001^b 3.08 <.001^b 2.07 <.001^b

Thirty-day readmission 3.42% 5.49% 10.31% 1.51 <.001^b 2.57 <.001^b 1.67 <.001^b

aBinary logistic multivariate regression.

bSignificance defined asP<.0029 after Bonferroni correction, significant values are in bold.

Table 4

Independent Risk Factors for Adverse Outcomes by Multivariate Logistic Regression in Patients Undergoing PJI Revision TKA.

Variable Any Complication Sepsis Nonhome Discharge Read Mission

OR (95% CI) P-Value OR (95% CI) P-Value OR (95% CI) P-Value OR (95% CI) P-Value

Age

<60 Reference e Reference e Reference e Reference e

60-70 0.97 (0.70-1.36) .872 1.14 (0.68-1.91) .611 1.43 (1.10-1.85) .008 0.92 (0.62-1.37) .682

71-80 1.04 (0.72-1.51) .829 0.90 (0.50-1.64) .738 2.87 (2.14-3.83) <.001^a 1.00 (0.64-1.56) .994

>80 1.87 (1.18-2.97) .008 2.02 (1.01-4.05) .047 7.59 (4.95-11.65) <.001^a 1.43 (0.81-2.53) .214

Female 0.86 (0.67-0.10) .233 0.89 (0.61-1.31) .4549 1.37 (1.12-1.67) .002^a 0.94 (0.69-1.27) .674

Body mass index (kg/m²)

Non-obese (<30) Reference e Reference e Reference e Reference e

Obese I (30-34.9) 1.26 (0.91-1.74) .162 1.31 (0.80-2.15) .286 1.11 (0.86-1.43) .438 1.01 (0.68-1.50) .970 Obese II (35-39.9) 1.58 (1.10-2.26) .014 1.87 (1.09-3.18) .022 1.09 (0.81-1.47) .563 1.28 (0.83-1.99) .269 Obese III (>40) 1.09 (0.73-1.63) .690 0.89 (0.47-1.71) .733 1.77 (131-2.39) <.001^a 1.05 (0.65-168) .844 Comorbidities

Diabetes mellitus 1.30 (0.98-1.73) .073 1.65 (1.09-2.50) .019 1.28 (1.01-1.61) .040 1.09 (0.77-1.54) .630 Current smoker 1.04 (0.69-1.56) .857 0.92 (0.48-1.74) .794 1.16 (0.84-1.61) .363 0.90 (0.54-1.50) .690

COPD 2.08 (1.37-3.18) .001^a 1.86 (0.99-3.53) .056 1.35 (0.91-2.01) .131 1.77 (1.07-2.93) .027

Hypertension 1.36 (1.00-1.85) .051 1.28 (0.78-2.06) .315 1.12 (0.89-1.41) .349 0.94 (0.66-1.34) .726

Dependent functional status 1.18 (0.77-1.81) .449 0.94 (0.47-1.87) .850 1.36 (0.93-2.00) .117 1.04 (0.60-1.79) .903 General anesthesia 1.36 (1.01-1.84) .043 2.53 (1.45-4.40) .001^a 1.47 (1.17-1.85) .001^a 1.05 (0.74-1.48) .804 ASA class

I or II Reference e Reference e Reference e Reference e

III 1.07 (0.78-1.46) .671 1.14 (0.70-1.86) .603 1.66 (1.31-2.12) <.001^a 1.87 (1.23-2.85) .003^a

IV 1.22 (0.72-2.07) .460 0.90 (0.37-2.15) .805 2.44 (1.54-3.84) <.001^a 2.85 (1.52-5.37) .001^a

CI, confidence interval.

aSignificance defined asP<.005 after Bonferroni correction, significant values are in bold.

relative to primary TKA. Operative time was also increased (þ3.7 minutes,P¼.021) in PJI revisions relative to non-PJI revisions, as was LOS (þ2.1 days,P<.001).

Discussion

PJI is a catastrophic complication following primary TKA; how- ever, short-term complication rates, readmission rates, operative time, and postoperative hospital LOS for this condition have not been clearly elucidated. This study used a large, multicenter, and prospectively collected registry to determine these outcomes in revision TKA and revision TKA subgroups, including PJI and non-PJI revisions. Revision TKA was associated with increased post- operative complications relative to primary TKA. In addition, revision TKA for PJI had a significantly worse postoperative morbidity profile than non-PJI revision TKA including death, sepsis, readmission, and prolonged hospital LOS.

In a prospective study of 117 patients with acute PJIs, Cobo et al [37]described clinical outcomes after treatment with debridement, antibiotics, and implant retention and exchange of removable components (eg, liner exchange). The authors reported a success rate of 57.3% and a mortality rate due to infection of 3.6% at 2.5 years of follow-up. Notably, patients infected with methicillin- resistant Staphylococcus aureus had the highest rates of failure.

Other perioperative complications were not assessed. Another study by Lange et al[38]assessed outcomes of chronic hip PJI in 130 patients, 82 of whom had their prosthesis removed and eventually reimplanted. In this cohort of 2-stage revision patients, the 1-year mortality rate was 2% and the 5-year reinfection rate was 14.6%.

In our study of 162,981 primary, 10,584 non-PJI revision, and 2196 PJI revision TKA cases, we found that PJI revisions had significantly worse 30-day morbidity and mortality compared to non-PJI revisions. Most importantly, there was an increased risk of death, sepsis, readmissions, and an increased LOS in PJI revisions relative to non-PJI revisions. PJI likely has such poor outcomes relative to non-PJI revisions because it is more difficult to completely eradicate a PJI than it is to definitively treat non-PJI revision indications such as prosthesis loosening. For instance, the failure rate at 2 years for all revision TKAs in aggregate is under 5%[39], but in one retrospective study with a minimum follow-up of 2 years, a failure rate of 84% was observed for acute PJI due to methicillin-resistant S aureus[40]. The pathophysiology of PJI is multifactorial, and in chronic cases the bacterium is able to form biofilm that serves as protection against host immune responses and antibiotic penetrance [41]. Furthermore, certain bacteria within the core of the biofilm can enter a stationary phase characterized by limited growth, also known as“persister cells,”in the setting of stressors such as nutrient deprivation or antibiotics [42,43]. This allows the bacteria to remain latent when in the presence of bacteriostatic antibiotics and then resume growth in the absence of antibiotics. Biofilm repopulation of a joint only requires a very small number of microbes remaining after incom- plete eradication of PJI[43].

This study also identified numerous independent predictors of adverse outcomes, including a significantly increased risk of sepsis and nonhome discharge in patients who underwent general anesthesia, compared to spinal or regional anesthesia.

Previous studies have identified increased rates of adverse outcomes in primary joint replacements, however, not in their revision counterparts [44e46]. For instance, Pugely et al [46]

found that patients who underwent primary total knee replacement with spinal anesthesia had lower rates of superfi- cial wound infections, blood transfusions, and overall compli- cations. The mechanism for this adverse association between general anesthesia and perioperative outcomes in patients undergoing revision TKA for PJI is unclear; however, numerous theories have been postulated. Epidural anesthesia is associated with fewer adverse outcomes, namely infectious complications, perhaps secondary to a decreased need for blood transfusions in patients who receive hypotensive epidural anesthesia. This technique utilizes epidural anesthesia with systemic hypoten- sion via a catecholamine infusion, leading to decreased blood loss, hematoma formation, and transfusion requirements[47].

Given that intraoperative blood transfusions are associated with an increased rate of postoperative bacterial infections [48], decreasing the usage of these products through neuraxial anesthesia, tranexamic acid, and potentially lower hemoglobin thresholds prior to transfusion may reduce serious infectious complications such as SSI and sepsis in patients undergoing revision arthroplasties for PJI.

This study has several limitations, most of which are inherent to ACS-NSQIP. First, we were unable to study results beyond 30 days, as ACS-NSQIP only contains complications and readmissions up to 30 days postoperatively. Therefore, we were unable to measure longer term failure rates and outcomes. In addition, we did not have additional variables relevant to the study of PJIs, including the time intervals between primary TKAs and the presentation of PJIs, iden- tities and sensitivities of infectious organisms, antibiotic treatment regimens, or any previous procedures that patients may have un- dergone. Finally, we were unable to differentiate between debride- ment, antibiotics, and implant retention, 1-stage, and 2-stage exchange arthroplasties, given the variability in procedure coding practices between surgeons, as well as the lack of CPT modifier codes in ACS-NSQIP.

Conclusion

In conclusion, this study illustrates that PJI revisions are associated with a significantly higher rate of postoperative complications, hospital readmissions, and prolonged post- operative LOS compared to non-PJI revisions and primary TKA controls. Furthermore, we identified baseline patient charac- teristics that are predictive of adverse outcomes, which may be useful for preoperative risk stratification and patient counseling.

Table 5

Association of Revision Total Knee Arthroplasty With Operative Time and Postoperative Length of Stay by Multivariate Regression.

Primary, Mean±SD All Revision TKAs Non-PJI Revision TKAs PJI Revision TKAs PJI vs Non-PJI Mean±SD Beta vs

pTKA^a

P-Value Mean±SD Beta vs pTKA^a

P-Value Mean±SD Beta vs pTKA^a

P-Value Beta^a P-Value

Operative time (min) 94.5±38.6 135±67.0 þ38.4 <.001^b 134±66.9 þ37.5 <.001^b 141±66.7 þ43.2 <.001^b þ3.7 .021^b Length of stay (d) 3.1±3.8 3.8±4.1 þ0.6 <.001^b 3.4±3.3 þ0.2 <.001^b 5.7±6.4 þ2.4 <.001^b þ2.1 <.001^b SD, standard deviation; pTKA, primary total knee arthroplasty.

aUnstandardized beta represents unit change in operation time or length of stay on multivariate linear regression.

b Significance defined asP<.025 after Bonferroni correction, significant values are in bold.

References

[1] Healthcare Cost and Utilization Project. Agency for Healthcare Research and Quality.https://hcupnet.ahrq.gov[accessed 14.03.17].

[2] Belmont Jr PJ, Goodman GP, Waterman BR, Bader JO, Schoenfeld AJ. Thirty-day postoperative complications and mortality following total knee arthroplasty:

incidence and risk factors among a national sample of 15,321 patients. J Bone Joint Surg Am 2014;96:20e6.

[3] Pulido L, Parvizi J, Macgibeny M, Sharkey PF, Purtill JJ, Rothman RH, et al. In hospital complications after total joint arthroplasty. J Arthroplasty 2008;23(6 Suppl 1):139e45.

[4] Fehringer EV, Mikuls TR, Michaud KD, Henderson WG, O'Dell JR. Shoulder arthroplasties have fewer complications than hip or knee arthroplasties in US veterans. Clin Orthop Relat Res 2010;468:717e22.

[5] Memtsoudis SG, Della Valle AG, Besculides MC, Gaber L, Laskin R. Trends in demographics, comorbidity profiles, in-hospital complications and mortality associated with primary knee arthroplasty. J Arthroplasty 2009;24:518e27.

[6] Pugely AJ, Martin CT, Gao Y, Schweizer ML, Callaghan JJ. The incidence of and risk factors for 30-day surgical site infections following primary and revision total joint arthroplasty. J Arthroplasty 2015;30(9 Suppl):47e50.

[7] Schairer WW, Vail TP, Bozic KJ. What are the rates and causes of hospital readmission after total knee arthroplasty? Clin Orthop Relat Res 2014;472:181e7.

[8] Ramkumar PN, Chu CT, Harris JD, Athiviraham A, Harrington MA, White DL, et al. Causes and rates of unplanned readmissions after elective primary total joint arthroplasty: a systematic review and meta-analysis. Am J Orthop (Belle Mead NJ) 2015;44:397e405.

[9] Huddleston JI, Wang Y, Uquillas C, Herndon JH, Maloney WJ. Age and obesity are risk factors for adverse events after total hip arthroplasty. Clin Orthop Relat Res 2012;470:490e6.

[10] Shin JI, Keswani A, Lovy AJ, Moucha CS. Simplified frailty index as a predictor of adverse outcomes in total hip and knee arthroplasty. J Arthroplasty 2016;31:2389e94.

[11] Duchman KR, Gao Y, Pugely AJ, Martin CT, Noiseux NO, Callaghan JJ. The effect of smoking on short-term complications following total hip and knee arthroplasty. J Bone Joint Surg Am 2015;97:1049e58.

[12] Pabinger C, Berghold A, Boehler N, Labek G. Revision rates after knee replacement. Cumulative results from worldwide clinical studies versus joint registers. Osteoarthritis Cartilage 2013;21:263e8.

[13] Saleh KJ, Hoeffel DP, Kassim RA, Burstein G. Complications after revision total knee arthroplasty. J Bone Joint Surg Am 2003;85-A(Suppl 1):S71e4.

[14] Dieterich JD, Fields AC, Moucha CS. Short term outcomes of revision total knee arthroplasty. J Arthroplasty 2014;29:2163e6.

[15] Belmont Jr PJ, Goodman GP, Rodriguez M, Bader JO, Waterman BR, Schoenfeld AJ. Predictors of hospital readmission following revision total knee arthroplasty. Knee Surg Sports Traumatol Arthrosc 2016;24:3329e38.

[16] Sharkey PF, Lichstein PM, Shen C, Tokarski AT, Parvizi J. Why are total knee arthroplasties failing todaydhas anything changed after 10 years?

J Arthroplasty 2014;29:1774e8.

[17] Sadoghi P, Liebensteiner M, Agreiter M, Leithner A, Bohler N, Labek G. Revi- sion surgery after total joint arthroplasty: a complication-based analysis using worldwide arthroplasty registers. J Arthroplasty 2013;28:1329e32.

[18] Ritter MA, Farris A. Outcome of infected total joint replacement. Orthopedics 2010;33.

[19] Tande AJ, Patel R. Prosthetic joint infection. Clin Microbiol Rev 2014;27:

302e45.

[20] Haenle M, Skripitz C, Mittelmeier W, Skripitz R. Economic impact of infected total knee arthroplasty. ScientificWorldJournal 2012;2012:196515.

[21] Hernandez-Vaquero D, Fernandez-Fairen M, Torres A, Menzie AM, Fernandez- Carreira JM, Murcia-Mazon A, et al. Treatment of periprosthetic infections: an economic analysis. ScientificWorldJournal 2013;2013:821650.

[22] Kuiper JW, Willink RT, Moojen DJ, van den Bekerom MP, Colen S. Treatment of acute periprosthetic infections with prosthesis retention: review of current concepts. World J Orthop 2014;5:667e76.

[23] Gehrke T, Alijanipour P, Parvizi J. The management of an infected total knee arthroplasty. Bone Joint J 2015;97-B(10 Suppl A):20e9.

[24] Petretta R, Phillips J, Toms A. Management of acute periprosthetic joint infection of the kneedalgorithms for the on call surgeon. Surgeon 2017;15:

83e92.

[25] Jamsen E, Stogiannidis I, Malmivaara A, Pajamaki J, Puolakka T, Konttinen YT.

Outcome of prosthesis exchange for infected knee arthroplasty: the effect of treatment approach. Acta Orthop 2009;80:67e77.

[26] Fehring TK, Odum SM, Berend KR, Jiranek WA, Parvizi J, Bozic KJ, et al. Failure of irrigation and debridement for early postoperative periprosthetic infection.

Clin Orthop Relat Res 2013;471:250e7.

[27] Marculescu CE, Berbari EF, Hanssen AD, Steckelberg JM, Harmsen SW, Mandrekar JN, et al. Outcome of prosthetic joint infections treated with debridement and retention of components. Clin Infect Dis 2006;42:471e8.

[28] Deirmengian C, Greenbaum J, Stern J, Braffman M, Lotke PA, Booth Jr RE, et al.

Open debridement of acute gram-positive infections after total knee arthro- plasty. Clin Orthop Relat Res 2003:129e34.

[29] Pitto RP, Castelli CC, Ferrari R, Munro J. Pre-formed articulating knee spacer in two-stage revision for the infected total knee arthroplasty. Int Orthop 2005;29:305e8.

[30] MacAvoy MC, Ries MD. The ball and socket articulating spacer for infected total knee arthroplasty. J Arthroplasty 2005;20:757e62.

[31] Tornero E, Garcia-Oltra E, Garcia-Ramiro S, Martinez-Pastor JC, Bosch J, Climent C, et al. Prosthetic joint infections due toStaphylococcus aureusand coagulase-negative staphylococci. Int J Artif Organs 2012;35:884e92.

[32] Hofmann AA, Goldberg T, Tanner AM, Kurtin SM. Treatment of infected total knee arthroplasty using an articulating spacer: 2- to 12-year experience. Clin Orthop Relat Res 2005;430:125e31.

[33] Meek RMD, Dunlop D, Garbuz DS, McGraw R, Greidanus NV, Masri BA. Patient satisfaction and functional status after aseptic versus septic revision total knee arthroplasty using the PROSTALAC articulating spacer. J Arthroplasty 2004;19:

874e9.

[34] Trickey AW, Wright JM, Donovan J, Reines HD, Dort JM, Prentice HA, et al.

Interrater reliability of hospital readmission evaluations for surgical patients.

Am J Med Qual 2017;32:201e7.

[35] ACS-NSQIP. Program specifics. https://www.facs.org/quality-programs/acs- nsqip/program-specifics[accessed 14.03.17].

[36] Basques BA, Ondeck NT, Geiger EJ, Samuel AM, Lukasiewicz AM, Webb ML, et al. Differences in short-term outcomes between primary and revision anterior cervical discectomy and fusion. Spine 2017;42:253e60.

[37] Cobo J, Miguel LG, Euba G, Rodriguez D, Garcia-Lechuz JM, Riera M, et al.

Early prosthetic joint infection: outcomes with debridement and implant retention followed by antibiotic therapy. Clin Microbiol Infect 2011;17:

1632e7.

[38] Lange J, Troelsen A, Soballe K. Chronic periprosthetic hip joint infection. A retrospective, observational study on the treatment strategy and prognosis in 130 non-selected patients. PLoS One 2016;11:e0163457.

[39] Sheng PY, Konttinen L, Lehto M, Ogino D, Jamsen E, Nevalainen J, et al.

Revision total knee arthroplasty: 1990 through 2002da review of the Finnish Arthroplasty Registry. J Bone Joint Surg Am 2006;88a:1425e30.

[40] Bradbury T, Fehring TK, Taunton M, Hanssen A, Azzam K, Parvizi J, et al. The fate of acute methicillin-resistantStaphylococcus aureusperiprosthetic knee infections treated by open debridement and retention of components.

J Arthroplasty 2009;24(6 Suppl):101e4.

[41] Zimmerli W, Trampuz A, Oschsner PE. Prosthetic-joint infections. N Engl J Med 2004;351:1645e54.

[42] Anderl JN, Zahller J, Roe F, Stewart PS. Role of nutrient limitation and stationary-phase existence in Klebsiella pneumoniae biofilm resistance to ampicillin and ciprofloxacin. Antimicrob Agents Chemother 2003;47:1251e6.

[43] del Pozo JL, Patel R. The challenge of treating biofilm-associated bacterial infections. Clin Pharmacol Ther 2007;82:204e9.

[44] Liu J, Ma C, Elkassabany N, Fleisher LA, Neuman MD. Neuraxial anesthesia decreases postoperative systemic infection risk compared with general anesthesia in knee arthroplasty. Anesth Analg 2013;117:1010e6.

[45] Memtsoudis SG, Sun X, Chiu YL, Stundner O, Liu SS, Banerjee S, et al. Peri- operative comparative effectiveness of anesthetic technique in orthopedic patients. Anesthesiology 2013;118:1046e58.

[46] Pugely AJ, Martin CT, Gao YB, Mendoza-Lattes S, Callaghan JJ. Differences in short-term complications between spinal and general anesthesia for primary total knee arthroplasty. J Bone Joint Surg Am 2013;95a:193e9.

[47] Sharrock NE, Salvati EA. Hypotensive epidural anesthesia for total hip arthroplasty: a review. Acta Orthop Scand 1996;67:91e107.

[48] Hill GE, Frawley WH, Griffith KE, Forestner JE, Minei JP. Allogeneic blood transfusion increases the risk of postoperative bacterial infection: a meta- analysis. J Trauma 2003;54:908e14.