In North America, two-stage reimplantation arthroplasty is widely considered the reference standard for treatment of chronic periprosthetic joint infections (PJIs) in the setting of THA and TKA . After component explantation and placement of an antibiotic-eluting spacer, the surgeon often uses a combination of inflammatory markers, culture results, and overall clinical appearance to assess treatment response and to guide reimplantation timing [18, 22].
Several recent studies have questioned the reliability and predictive capabilities of two commonly used serum inflammatory markers: erythrocyte sedimentation rate (ESR) and C-reactive protein (CRP), in determining the success of subsequent joint reimplantation after PJI [10, 13, 20]. These studies have suggested that the normalization of ESR and CRP was not associated with infection-free survivorship after reimplantation nor was consistent elevation of ESR and CRP associated with an infection-related return to the operating room. In a study specifically looking at PJI after THA, Shukla et al.  found that 62.5% of ESR and 27.5% of CRP values remained elevated compared with their cited threshold cutoff levels in the prereimplantation period in patients in whom the infection had been eradicated.
Because of this uncertainty surrounding serologic threshold values, we asked whether a percent improvement in serology markers might improve diagnostic accuracy to help determine the timing of second-stage reimplantation. We sought to determine whether (1) the percent, or delta, change in ESR and CRP values from preresection to prereimplantation (∆ESR, ∆CRP) is a useful marker of infection eradication and (2) whether the initial PJI causative organism (resistant, nonresistant, or culture-negative) is associated with serum ESR and CRP values before and after treatment with an antibiotic spacer and parenteral antibiotic therapy.
Patients and Methods
After obtaining requisite institutional review board approval, we reviewed the institutional databases at two tertiary referral total joint centers (OrthoCarolina, Charlotte, NC, USA, and the Hospital for Special Surgery, New York, NY, USA) for all patients with chronic TKA/THA PJI between 2005 and 2014. We excluded all acute hematogenous and perioperative infections occurring within 3 months of index surgery . A query of the databases resulted in 630 potential patients available to screen for eligibility who underwent a first-time two-stage revision arthroplasty (Stage 1 = explantation and antibiotic spacer placement; Stage 2 = removal of spacer and reimplantation arthroplasty). Preliminary exclusion criteria included those treated with a prior two-stage protocol for PJI, acute infections (< 4 weeks of symptoms), simultaneous multiple joint infections, PJI in the setting of documented inflammatory arthropathy (rheumatoid disease, systemic lupus erythematous, etc), fungal PJI, and those with PJI with existing hardware from previous fracture fixation or extensor mechanism disruption. Furthermore, we excluded patients who were referred to our institutions with an antibiotic-loaded cement spacer in place who did not have preresection inflammatory markers because of the discrepancies this creates in trying to standardize an antibiotic course and timing of reimplantation. We did, however, include six patients who underwent an interim spacer exchange at one of the two participating centers as a result of the persistence of infection as identified by preoperative laboratory values, culture testing, and/or intraoperative pathology. We treated the serology values before the final reimplantation as prereimplantation values. This left us with 375 eligible patients for analysis, nine of whom were lost to followup. Sixty-four participants were excluded for incomplete recordings of both serologic markers at the preresection time point and 11 patients were excluded for not having a record of prereimplantation values. Our final analysis included 291 patients with minimum 2-year followup and complete serologies at the two stated time points (Fig. 1).
Serum ESR (mm/hr) and noncardiac CRP (mg/L) values were recorded at two designated points: (1) preresection and (2) after completion of 6 weeks of organism-specific intravenous (IV) antibiotic therapy with at least an additional 2-week antibiotic holiday with a drug-eluting spacer before reimplantation. All patients included met the modified Musculoskeletal Infection Society (MSIS) diagnostic criteria for PJI, which was also used to determine infection-related failures . We applied the conventional thresholds for upper limits of normal of 30 mm/hr for ESR and 10 mg/L for noncardiac CRP as references [1, 2, 27, 28]. A final joint aspiration was routinely performed before reimplantation and fluid was analyzed for synovial white blood cell count with differential and tested for aerobic and anaerobic cultures. All patients were treated with similar protocols for two-stage reimplantation by one of nine fellowship-trained arthroplasty surgeons at either institution. These included the regular use of high-dose antibiotic-laden polymethylmethacrylate (PMMA) spacers fashioned with 4 g of organism-specific antibiotics per package of PMMA for culture-positive cases and a combination of 2 g vancomycin and 2.4 g tobramycin per package of cement for culture-negative cases. Intravenous antibiotics were then continued for 6 weeks with at least a full 2-week antibiotic holiday period before rechecking serologies. All patients underwent thorough débridement of infected tissue and placement of an antibiotic-laden cement spacer, either static or articulating. Patients were followed by the infectious disease service and started on a 6-week course of organism-specific IV antibiotics. Patients with culture-negative infections received varied broad-spectrum antibiotic regimens at the discretion of the infectious disease consultant. The most common combination of broad-spectrum treatments included vancomycin and ceftriaxone. All surgeons routinely used antibiotic-laden cement for knee reimplantation and utilized cementless revision components for hip reimplantation. At one institution, the protocol was to wait 3 months before reimplantation, but at the other institution, this practice varied. It was not customary to prescribe oral suppressive antibiotics after reimplantation. At the time of Stage 2 reimplantation, frozen section histopathology was sent at the discretion of the treating surgeon [5, 9, 19].
Demographic and surgical variables recorded included age at the time of Stage 1 resection, sex, operative joint (hip or knee), body mass index (BMI), American Society of Anesthesiologists (ASA) classification, and microbial organism cultured at the time of resection. The median age at time of explantation for the cohort was 65 years (range, 34-90 years) and the median BMI was 32 kg/m2 (range, 18-62 kg/m2). Of the 291 patients included in final analysis, there were 145 hips (50%) and 146 knees (50%) and 159 were men (55%) and 132 (45%) women. Ninety-three percent of patients were either ASA Class II or III; five patients were classified as ASA Class I and 11 patients as ASA Class IV. The median time to reimplantation was 102 days (interquartile range [IQR], 84-133 days). Patient records were reviewed for return to the operating room, revision procedures, and recurrent or persistent PJI for a minimum of 2 years. Recurrent or persistent infection was defined as any reoperation for PJI (based on MSIS criteria), one positive aspiration after the drug holiday, culture-positive findings at the time of reimplantation, or two positive cultures obtained by aspiration after the second-stage reimplantation.
Forty-eight of 291 patients (16%) underwent a revision procedure for recurrent or persistent infection, whereas 31 patients (10%) were revised during the review period for noninfectious reasons. Noninfectious revision indications included aseptic loosening (12), instability (nine), periprosthetic fracture (four), wear-related complications (four), arthrofibrosis (one), and osteolysis (one).
Causative organisms for initial PJI included 59 (20%) coagulase-negative Staphylococcus, 32 (11%) methicillin-resistant Staphylococcus aureus (MRSA), 44 (15%) methicillin-sensitive S aureus, 37 (13%) Streptococcus species, 10 (3%) nonresistant Enterococcus, and 68 (24%) other organisms. Forty-one patients (14%) had culture-negative infections, as defined by the MSIS.
Changes in serum inflammatory markers were analyzed using a two-sided Wilcoxon rank-sum test for a nonparametric distribution of values across the groups. These ESR and CRP values were reported as medians and IQRs. Culture results and patient demographics were recorded and analyzed during subgroup analysis using receiver operator curve (ROC) testing for the area under the curve (AUC) regarding the changes in ESR and CRP while controlling for ASA class. ROC curves display the relationship between statistical tests in which false-negative results (1-specificity) are plotted on the X-axis against sensitivity (true-positive results) on the Y-axis. An AUC equal to 1 exhibits an ideal test with 100% sensitivity and specificity, whereas an AUC < 0.5 indicates a test no better than chance. Youden’s J-statistic was used to attempt to determine a threshold value for each serologic marker on its ROC curve. Statistical significance was set at p < 0.05. All statistical testing was conducted using SAS 9.4 for Windows (SAS Institute Inc, Cary, NC, USA).
Is There a Threshold ∆ESR or ∆CRP That Can Guide Reimplantation?
Receiver operator characteristic AUCs demonstrated that both the ∆ESR (AUC = 0.581) and ∆CRP (AUC = 0.539) percentages were poor markers of recurrent or persistent infection risk (Fig. 2 A-B). When comparing preresection with prereimplantation values, the median percent ∆ESR was 50% (IQR, 17%-77%) for those patients who remained infection-free versus 59% (IQR, 29%-78%) for those who became reinfected (p = 0.540). The median percent ∆CRP was 77% (IQR, 47%-92%) for those patients who remained infection-free versus 79% (IQR, 46%-95%) for those who experienced reinfection (p = 0.634; Table 1). For the overall cohort, the median preresection ESR value was 58 mm/hr (IQR, 34-84 mm/hr) and CRP was 51 mg/L (IQR, 25-132 mg/L), whereas the prereimplantation ESR was 23 mm/hr (IQR, 11-40 mg/L) and the CRP was 9 mg/L (IQR, 7-20 mg/L). When evaluated independently, there were no differences noted between preresection and prereimplantation absolute ESR or CRP values in patients who became reinfected and those who did not. As a result of the low sensitivity and specificity caused by the wide distribution of changes in inflammatory levels for those who did and did not experience a reinfection, a threshold value could not be calculated by Youden’s J-statistic.
Are ESR and CRP Levels Associated With Organism-specific Differences?
When performing a subgroup analysis on the causative organisms associated with the primary PJI, the only serologic difference noted was higher prereimplantation ESR values in those afflicted with a MRSA PJI (Table 2). No statistically significant difference was noted when ∆CRP values between resistant and nonresistant PJIs were compared (p = 0.241).
Patients with culture-negative infections exhibited no differences in serum inflammatory markers at both recorded points compared with those with identifiable organism-driven PJI (Table 3). Recurrent or persistent PJI risk was not associated with initial resistant versus nonresistant speciation (odds ratio, 1.192; p = 0.800; 95% confidence interval, 0.4624-3.0743). There were no associations noted between initial causative PJI species and recurrent infection given the broad distribution of organisms (Table 4).
Given that the goal of treating PJI with two-stage reimplantation arthroplasty is to eradicate the infection, properly interpreting laboratory tests between stages is critical to defining the time for reimplantation. Previous authors have been unable to identify a serologic threshold to guide the timing of reimplantation [4, 10, 13, 25]. However, no study of which we are aware has sought to investigate if the percent change in serum ESR and CRP might serve as a more accurate marker of reinfection risk than an absolute threshold value. By investigating the ∆ESR and ∆CRP before and after two-stage revision arthroplasty, we discovered that the percent change was not associated with reinfection. Secondarily, we found that causative PJI organisms were neither associated with changes in ESR and CRP nor the incidence of persistent or recurrent infection.
Although we have reviewed many patients treated with an established two-stage exchange protocol over a 10-year period, multiple study limitations exist. Our records did not have the granularity to adequately capture information regarding systemic host factors and limb status, both of which have been noted to influence clinical success when treating PJIs [6, 16]. The importance of classifying the host status has recently been recognized by the MSIS host classification system and has shown associations between poor host grade and reinfection risks . Although this is now recognized as being an important factor in any study regarding PJI, it was not possible to cite this accurately in a retrospective manner. We did, however, record ASA classification as a crude measure of overall health. Additionally, our data set did not capture whether an articulating versus static antibiotic spacer was placed at the time of explantation. Controversy remains regarding the clinical benefits of each spacer type, but to date, neither has been shown to have superior infection eradication rates [3, 7, 11, 17]. A third limitation in our analysis is that the multiple causative infectious species precluded us from comparing individual organisms as a result of the low overall effect size of each bacteria type. We chose to include six patients with antibiotic spacers who underwent a spacer exchange and treated them as recurrent/persistent infections. It was the decision of the treating surgeon to perform repeat débridement and spacer exchange, which may add additional bias to the study because a specific serologic threshold was not used. Decisions of this nature do not solely depend on one test, but are a compilation of wound appearance, host status, remaining bone stock, serology, pain patterns, and causative organisms that drive a surgeon to redébride or reimplant. An additional limitation that we continue to learn more about is chronic antibiotic use. We do not have accurate records of whether patients were continued on prolonged oral suppressive antibiotics after two-stage reimplantation, as is more commonly done at the present time given early findings that it may increase infection-free survivorship [8, 26]. We do not think this limitation alters the underlying findings of our study, however, because suppressive antibiotics are typically given for a variable time course. Thus, if an underlying infection were to be present, it might be expected to reappear once the oral regimen is stopped.
An additional limitation to consider is the fact that we report on the outcomes of two-stage protocols conducted at two high-volume, tertiary referral centers. Thus, the different bacterial profiles of these regions and subtle differences in the timing of laboratory draws and antibiotic regimens as part of the two-stage treatment with multiple providers may limit the generalizability of our results. Furthermore, the variable timing to ultimate reimplantation may confound the reproducibility of our findings. Although we reported on an average of 14 weeks between stages, it is common practice for physicians to space surgeries outside of the 90-day global period given scheduling issues with a busy clinical practice and concerns with reimbursement for the workload performed. With this said, we strongly believe that the strength of our study rests in the large sample and effect size that provide a strong case that the change in serologies is no better than chance to predict infection-related failures. Although there are other important variables to control for, given our available sample size, the regression model would only support one covariate for analysis. We chose ASA as a surrogate for patient health status because this was the most consistent such measure across time.
We found that both the ∆ESR and ∆CRP percentages were poor markers of recurrent or persistent infection risk. Several previous studies have investigated how to best monitor infection status using serologic testing in the period before reimplantation [10, 13, 14, 25]. However, all studies of this nature suffer from the fact that two-stage reimplantation is a relatively successful procedure. Therefore, when comparing laboratory values of successfully treated patients with those who became reinfected, a very large sample size is needed to reach adequate power. In their small case series, Lindsay et al.  reported no reinfections in 14 of 19 patients with PJI whose ESR and CRP both normalized before reimplantation. Ghanem et al.  studied 109 patients with infected TKAs who underwent two-stage revision and were unable to define an absolute CRP or ESR threshold for infection eradication despite a 21% (23 of 109) reinfection rate at an average of 2 years. The authors did note, however, that all successfully treated PJIs had downtrending serum inflammatory values. Kusuma et al.  also reported on 76 infected TKAs treated with two-stage exchange in which 12% of patients (eight patients) experienced a recurrent PJI. They discovered that both ESR and CPR normalized in only 37% (28 of 76) of patients, but neither serum marker was accurate nor had an acceptable positive predictive value to diagnose persistent infection. Shukla et al.  investigated 86 consecutive patients with chronic hip PJI treated with two-stage exchange and noted a reinfection rate of 10% (nine patients). They also found no difference in mean ESR or CRP for those patients who remained persistently infected and noted that 19 of 86 patients (23.8%) had both ESR and CRP remain persistently elevated despite not experiencing a PJI recurrence. None of these three larger studies were able to establish a reproducible ESR or CRP threshold level to indicate infection eradication. Although our study population was larger than the four aforementioned studies combined, we were unable to determine a percent change threshold in either serology value to accurately predict reinfection.
We found that the preexplantation ESR was significantly higher in PJI caused by resistant versus nonresistant bacteria. This is not necessarily surprising given the virulence of MRSA compared with nonresistant bacteria. Regarding MRSA, Ryu et al.  found increased preresection CRP serologies were associated with an increased risk of reinfection (13.9% versus 4%) and a longer duration of IV antibiotic treatment needed for CRP to normalize. Unfortunately, in our series, the prereimplantation ESR and CRP values in patients with resistant and nonresistant PJI afforded no added utility when trying to predict reinfection in at-risk patients. Culture-negative PJIs in our series did not vary in preresection or prereimplantation inflammatory markers after 6 weeks of IV antibiotic treatment from those with positive cultures. Huang et al.  found no difference between infection eradication rates when treating culture-negative PJI with parenteral vancomycin as part of their 6-week postresection protocol by using the assumption that there was an underlying Gram-negative cause. However, their study cohort also considered débridement and retention protocols, which may decrease the generalizability of their findings. Choi et al.  and Malekzadeh et al.  found that patients with culture-negative PJI commonly had higher rates of previous antibiotic use yet lower rates of infection recurrence after two-stage arthroplasty. This is in contrast to Mortazavi et al. , who reported a 4.5 times increased odds of infection with culture-negative PJI. We did not find a higher risk of reinfection with culture-negative patients in our series nor did we have complete information regarding prior antimicrobial therapies. If prior antibiotic use is the presumed cause of some culture-negative infections, it does not seem to affect recurrence rates.
We have shown that the delta change in ESR and CRP, regardless of causative organism, provides no additional diagnostic accuracy to guide the timing of reimplantation. The decision to reimplant the definitive prosthesis in the setting of PJI must therefore take into account a number of variables rather than a specific threshold for serum inflammatory markers. At this time, we believe surgeons should continue to use a variety of tools to guide the timing of reimplantation, including clinical suspicion, wound appearance, and pain relief in addition to serology, synovial fluid analysis, and intraoperative findings.
We thank the staff at the OrthoCarolina Research Institute for their help with data management and securing patient followup.
1. Alijanipour P, Bakhshi H, Parvizi J. Diagnosis of periprosthetic joint infection: the threshold for serological markers. Clin Orthop Relat Res. 2013;471:3186–3195.
2. Berbari E, Mabry T, Tsaras G, Spangehl M, Erwin PJ, Murad MH, Steckelberg J, Osmon D. Inflammatory blood laboratory levels as markers of prosthetic joint infection: a systematic review and meta-analysis. J Bone Joint Surg Am. 2010;92:2102–2109.
3. Chiang E-R, Su Y-P, Chen T-H, Chiu F-Y, Chen W-M. Comparison of articulating and static spacers regarding infection with resistant organisms in total knee arthroplasty. Acta Orthop. 2011;82:460–464.
4. Choi H-R, Kwon Y-M, Freiberg AA, Nelson SB, Malchau H. Periprosthetic joint infection with negative culture results: clinical characteristics and treatment outcome. J Arthroplasty. 2013;28:899–903.
5. Della Valle CJ, Bogner E, Desai P, Lonner JH, Adler E, Zuckerman JD, Di Cesare PE. Analysis of frozen sections of intraoperative specimens obtained at the time of reoperation after hip or knee resection arthroplasty for the treatment of infection. J Bone Joint Surg Am. 1999;81:684–689.
6. Fehring KA, Abdel MP, Ollivier M, Mabry TM, Hanssen AD. Repeat two-stage exchange arthroplasty for periprosthetic knee infection is dependent on host grade. J Bone Joint Surg Am. 2017;99:19–24.
7. Fehring TK, Odum S, Calton TF, Mason JB. Articulating versus static spacers in revision total knee arthroplasty for sepsis. The Ranawat Award. Clin Orthop Relat Res. 2000;380:9–16.
8. Frank JM, Kayupov E, Moric M, Segreti J, Hansen E, Hartman C, Okroj K, Belden K, Roslund B, Silibovsky R, Parvizi J, Della Valle CJ; Knee Society Research Group. The Mark Coventry, MD, Award: Oral antibiotics reduce reinfection after two-stage exchange: a multicenter, randomized controlled trial. Clin Orthop Relat Res. 2017;475:56–61.
9. George J, Kwiecien G, Klika AK, Ramanathan D, Bauer TW, Barsoum WK, Higuera CA. Are frozen sections and MSIS criteria reliable at the time of reimplantation of two-stage revision arthroplasty? Clin Orthop Relat Res. 2016;474:1619–1626.
10. Ghanem E, Azzam K, Seeley M, Joshi A, Parvizi J. Staged revision for knee arthroplasty infection: what is the role of serologic tests before reimplantation? Clin Orthop Relat Res. 2009;467:1699–1705.
11. Hart WJ, Jones RS. Two-stage revision of infected total knee replacements using articulating cement spacers and short-term antibiotic therapy. J Bone Joint Surg Br. 2006;88:1011–1015.
12. Huang R, Hu C-C, Adeli B, Mortazavi J, Parvizi J. Culture-negative periprosthetic joint infection does not preclude infection control. Clin Orthop Relat Res. 2012;470:2717–2723.
13. Kusuma SK, Ward J, Jacofsky M, Sporer SM, Della Valle CJ. What is the role of serological testing between stages of two-stage reconstruction of the infected prosthetic knee? Clin Orthop Relat Res. 2011;469:1002–1008.
14. Lindsay CP, Olcott CW, Gaizo DJD. ESR and CRP are useful between stages of 2-stage revision for periprosthetic joint infection. Arthroplasty Today. 2017;3:183–186.
15. Malekzadeh D, Osmon DR, Lahr BD, Hanssen AD, Berbari EF. Prior use of antimicrobial therapy is a risk factor for culture-negative prosthetic joint infection. Clin Orthop Relat Res. 2010;468:2039–2045.
16. McPherson EJ, Woodson C, Holtom P, Roidis N, Shufelt C, Patzakis M. Periprosthetic total hip infection: outcomes using a staging system. Clin Orthop Relat Res. 2002;403:8–15.
17. Meek RMD, Masri BA, Dunlop D, Garbuz DS, Greidanus NV, McGraw R, Duncan CP. Patient satisfaction and functional status after treatment of infection at the site of a total knee arthroplasty with use of the PROSTALAC articulating spacer. J Bone Joint Surg Am. 2003;85:1888–1892.
18. Mont MA, Waldman BJ, Hungerford DS. Evaluation of preoperative cultures before second-stage reimplantation of a total knee prosthesis complicated by infection. A comparison-group study. J Bone Joint Surg Am. 2000;82:1552–1557.
19. Moore GA, Hill MV, Kuo YF, Stephenson K, Lindsey RW. First Place: The utility of frozen sections in hip and knee reimplantation surgery. Curr Orthop Pract. 2015;26:332–335.
20. Mortazavi SMJ, Vegari D, Ho A, Zmistowski B, Parvizi J. Two-stage exchange arthroplasty for infected total knee arthroplasty: predictors of failure. Clin Orthop Relat Res. 2011;469:3049–3054.
21. Parvizi J, Gehrke T; International Consensus Group on Periprosthetic Joint Infection. Definition of periprosthetic joint infection. J Arthroplasty. 2014;29:1331.
22. Pignatti G, Nitta S, Rani N, Dallari D, Sabbioni G, Stagni C, Giunti A. Two stage hip revision in periprosthetic infection: results of 41 cases. Open Orthop J. 2010;4:193–200.
23. Ryu DJ, Kang JS, Moon KH, Kim MK, Kwon DG. Clinical characteristics of methicillin-resistant Staphylococcus aureus
infection for chronic periprosthetic hip and knee Infection. Hip Pelvis. 2014;26:235–242.
24. Sayeed Z, Anoushiravani AA, El-Othmani MM, Chambers MC, Mihalko WM, Jiranek WA, Paprosky WG, Saleh KJ. Two-stage revision total knee arthroplasty in the setting of periprosthetic knee infection. Instr Course Lect. 2017;66:249–262.
25. Shukla SK, Ward JP, Jacofsky MC, Sporer SM, Paprosky WG, Della Valle CJ. Perioperative testing for persistent sepsis following resection arthroplasty of the hip for periprosthetic infection. J Arthroplasty. 2010;25:87–91.
26. Siqueira MBP, Saleh A, Klika AK, O’Rourke C, Schmitt S, Higuera CA, Barsoum WK. Chronic suppression of periprosthetic joint infections with oral antibiotics increases infection-free survivorship. J Bone Joint Surg Am. 2015;97:1220–1232.
27. Spangehl MJ, Masri BA, O’Connell JX, Duncan CP. Prospective analysis of preoperative and intraoperative investigations for the diagnosis of infection at the sites of two hundred and two revision total hip arthroplasties. J Bone Joint Surg Am. 1999;81:672–683.
28. Ting NT, Della Valle CJ. Diagnosis of periprosthetic joint infection—an algorithm-based approach. J Arthroplasty. 2017;32:2047–2050.