Reinfection Rate and Its Associated Factors after Two-Stage Revision for Infected Total Knee Arthroplasty


 Background: Two-stage revision is the gold standard for treatment of infected total knee arthroplasty. The purpose of our study was to evaluate the reinfection rate of two-stage revision and to analyze the factors affecting the prognosis of two-stage revision for infected total knee arthroplasty.Methods: One hundred seven cases of two-stage revision for infected total knee arthroplasty were reviewed retrospectively from March 2006 to November 2019. We evaluated possible risk factors between success and reinfection groups. Statistical analyses included multivariable logistic regression analysis to examine the relative contribution of risk factors to the success of two-stage revision. Results: There were 19 cases of reinfection (17.8%) after two-stage revision in our center. Between the success and reinfection groups, there was a significant difference in history of cancer (p=0.015). Also, multivariable logistic regression analysis of risk factors demonstrated history of cancer (HR 5.928, p=0.015). There were no statistically significant differences in reinfection relative to other risk factors. Conclusions: In subjects undergoing two-stage revision for infected total knee arthroplasty, history of cancer was a risk factor for reinfection, though no other significant differences between risk factors was shown for reinfection.Trial registration: Retrospectively registeredLevel of evidence: IV

However, recurrence of reinfection after two-stage revision occurs in up to 19% of cases [26][27][28], which elucidates the socioeconomic and individual burden of infected TKA. Previous studies have attempted to identify risk factors associated with reinfection after two-stage revision [29][30][31][32][33]. However, there were no common risk factors among studies. The purpose of the present study was to evaluate the reinfection rate of two-stage revision and to analyze the factors affecting prognosis of two-stage revision for infected TKA.

Methods
The design of this retrospective study was approved by the institutional review board (IRB) at the author's hospital, and all patients provided informed consent.
One hundred forty-two cases (136 patients) of infected TKA from 2008 to 2019 at a single center were reviewed retrospectively. Among these, 35 cases were successfully treated without need for component removal or antibiotic-loaded cement spacer insertion, and four failure cases required two-stage revision after infection control. Therefore, a total of 107 cases (104 patients) of two-stage revision for infected TKA was analyzed including four failure cases (Fig. 1).
The diagnosis of periprosthetic joint infection prior to two-stage revision was con rmed when at least three of the following criteria were met: 1) CRP > 1 mg/dl; 2) ESR > 30 mm/h; 3) positive culture from joint aspirate; 4) pus at operation; and 5) positive intra-operative culture [36].
Two-stage revision consisted of removal of all prosthetic components, insertion of a vancomycin-impregnated cement (2 g vancomycin per 40 g cement) articulating spacer, and delayed reimplantation. After the rst stage, patients underwent physiotherapy to encourage passive knee movement and preserve quadriceps strength. Delayed reimplantation was performed when the wound was healthy and the patient was clinically stable with normal CRP. Clinical and radiological data from all patients who underwent joint replacement were collected retrospectively from the joint registries of our institutions. Patients had been examined before surgery and at six weeks, six months, and one year after surgery and yearly thereafter.
Reinfection was diagnosed by clinical signs, blood work (ESR and CRP), and positive culture of synovial aspiration. The mean follow-up period after revision TKA was 62.2 ± 36.9 months. In the reinfection group, 16 patients were females, three patients were male, and the mean age was 70.7 ± 7.8 years. The mean time from revision TKA to diagnosis of reinfection was 15.0 ± 18.8 months. We evaluated possible risk factors between success and reinfection groups.
Differences in continuous variables between the two groups were evaluated using the unpaired t-test or Mann-Whitney rank test. The data are shown as mean ± standard deviation. For discrete variables, differences are shown as count and percentage and analyzed with the x2 outcome (= reinfection) occurred over 15 years of follow-up. Multivariate logistic regression analysis was used to assess the independent impact factor for reinfection. A two-tailed p-value < 0.05 was considered statistically signi cant.

Results
There were 19 cases of reinfection (17.8%) after two-stage revision in our center. Three patients were treated by I & D, and seven patients underwent a twostage re-revision TKA. Seven patients who were in poor physical condition or refused additional surgical procedures were treated with antimicrobial therapy, and two cases were lost to follow-up.
The demographic characteristics and univariate analysis of infected TKA patients are shown in Table 1. The mean durations from primary TKA to infection and from infection to implant removal in all patients were 34.3 ± 42.3 months and 20.7 ± 44.5 weeks, respectively.  In 19 reinfection cases, those with cancer had a nal collection period of 3 to 99 months, and those without had a nal collection period of 0 to 157 months. There were a total of 10 cancer patients: ve were reinfected and ve were observed without reinfection. In the reinfection-cancer group, there were two uterine cancers, two gastric cancers, and one colorectal cancer (Table 3).  Figure 2 depicts the cumulative survival rates between the cancer and non-cancer groups during a one-year period. There was a signi cantly higher chance that subjects in the cancer group were reinfected within one year compared to that in subjects in the non-cancer group (33.3% vs. 7.2%, p = 0.002).

Discussion
Infection remains one of the most serious complications of TKA. Two-stage revision is the standard treatment of infected TKA and seems to eradicate infection and provide a functional outcome.
This case-control study aimed to evaluate the reinfection rate of two-stage revision and to analyze the factors affecting prognosis of two-stage revision for infected TKA. The most important nding of our study is that history of cancer was the only risk factor for reinfection after two-stage revision for infected TKA. There was a signi cantly higher chance that a subject with cancer would be reinfected within one year than a subject without. In our study, there were 19 cases of reinfection (17.8%) after two-stage revision. This is similar to previous studies that showed a 10-12% incidence of reinfection after two-stage revision [31,37,38].
Several studies have reported risk factors and reinfection rate of two-stage revision for infected TKA [26,33,[39][40][41]. Sabry et al. reported the following as preoperative predictors of failure following two-stage revision: duration of symptoms, time from index surgery, number of previous surgeries, high markers of in ammation levels (CRP, ESR, and peripheral WBC count), lower hemoglobin and hematocrit, need for soft tissue coverage, time to reimplantation, previous infection in the same joint, higher the American society of anesthesiologists (ASA) score, DM, anemia, heart disease, and infection with a gram-negative organism in the absence of malignancy [39]. Kubista et al. reported chronic lymphedema, revision between resection and de nitive reimplantation, and intravenously administered cefazolin as the strongest positive predictors of treatment failure [33]. Fashingbauer et al. reported revision during or after a twostage exchange, number of surgeries, and alcohol abuse as risk factors for recurrence, and that recurrence rates did not differ among organisms [32].
Fehring et al reported repeat two-stage exchange arthroplasty for periprosthetic knee infection is dependent on host grade(Musculoskeletal Infection Society, MSIS) [40]. Also, Vadiee et al found a higher incidence of failure in those patients with poor general health based on MSIS score, inadequate soft tissue envelope and resistant bacteria [41].
Previous studies attempted to identify risk factors associated with reinfection after two-stage revision [29][30][31][32][33]. However, little information is available concerning the prognosis and risk factors in reinfection after two-stage revision for infected TKA. In this study, that undergoing revision operation for infected TKA with history of cancer (10 patients) showed a relatively high re-infection rate of 50% (5 patients).
This study had some limitations. First, it was retrospectively designed and may have introduced bias when data were not accurately reported in the medical chart. In particular, medical history was recorded through patient statements. Second, in the cancer group, we could not assess cancer recurrence or immunosuppressant use during the follow-up period. Third, we could not analyze antibiotics used due to individualization of treatment regimens without a standardized protocol. Fourth, the nding that there was no organism-dependent difference in reinfection after two-stage revision con icts with the current literature [42][43][44]. One possible reason for this discrepancy is that our study had an insu cient number of cases to achieve adequate power analyses.
Nevertheless, our study says that history of cancer might be especially important among systemic host compromising factors and signi cantly higher chance that subjects in the cancer group were reinfected within one year compared to that in subjects in the non-cancer group.

Conclusions
In conclusion, patients with identi ed history of cancer undergoing revision operation for infected TKA showed worse outcomes. Therefore, staged revision arthroplasty in patients with history of cancer should be closely observed to minimize re-infection, and these patients should be informed of the high probability of infection. Consent for publication: Authors agree to publication. This manuscript has not been published in any journal Availability of data and material: All data generated or analysed during this study are included in this published article.
Competing interests: Each author certi es that he or she has no commercial association (eg. consultancies, stock ownership, equity interest, patent, licensing arrangements, etc.) that might pose a con ict of interest in connection with the submitted article.
Funding: Not applicable.
Authors' contributions: Chang Hyun Nam, Ji-Hoon Baek: writing and revision of article. Su Chan Lee, Kyungwon Choi, Hye Sun Ahn: data collection and statistical analysis.