Two-stage Total Knee Arthroplasty in the Treatment of Advanced Knee Septic Arthritis: A Single-center Experience

Background: Septic arthritis (SA) is dened as the joint inammation secondary to bacterial infection with a potentially fatal condition, but with no established approach treatment currently. Two-stage primary total knee arthroplasty (TKA) with an antibiotic-laden cement spacer has been shown to be effective in treating SA of the knee in individual cases. We reported our cases to contribute to current information on two-stage TKA for the treatment of advanced SA in the knee joint Methods: We retrospectively screened the clinical proles of 11 patients who were suffering from SA in their knee joints and underwent two-stage TKA with an antibiotic-laden cement spacer. Information on inammatory serological markers was summarised, and statistical comparisons of Hospital for Special Surgery (HSS) scores were performed. The degrees of the range of motion (ROM) in the pre-, interval- and post-operation periods were obtained, and the signicance level was set at a p value of <0.05. Results: Infection in the knee joint was eliminated in all 11 patients with the two-stage TKA with an antibiotic-laden cement spacer. C-reactive protein levels returned to normal after 3.4 ± 1.4 weeks (range, 1–6 weeks), whereas erythrocyte sedimentation rate returned to normal levels after 16.3 ± 3.0 weeks (range, 13–23 weeks). The means of HSS scores and ROM degrees signicantly increased after the rst and second stages of surgery (p < 0.05). No difference in HSS score was observed in three periods between age groups, genders, microorganism culture results or comorbidities in our study. Conclusions: Two-stage TKA appears to be an a reliable alternative for the treatment of advanced knee SA, providing an encouraging nal clinical outcome in controlling infection, relieving clinical severity and improving the function of the keen joint.


Background
Septic arthritis (SA) is an emergency condition of joint infection, having an approximately two cases per 100,000 persons per year and the knee being the most commonly affected joint accounting for 40%-50% of reported cases in adults [1][2][3] . Patients with SA are often misdiagnosed with rheumatic disease and thus receive inadequate treatment 4 , which often leads to irreversible joint destruction and even case-fatality (i.e., 3% to 29%) 5,6 . Therefore, diagnosing this condition early and providing effective treatment immediately are indeed important 1 . The standard diagnosis of SA is based on the detection of bacteria in the synovial uid through joint puncture. The primary therapy for SA includes the surgical interventions such as arthroscopic or open irrigation and debridement, in concerted with effective antibiotics utiliszation 1 . Arthroscopic surgery plays effective roles in SA treatment 6,7 , and open irrigation and debridement are usually considered for the patients with chronic infectious arthritis or persistent infection after arthroscopic surgery 4,8 . However, some cases show unsatisfactory recovery after such primary treatment, and eventually developed into non-salvageable destructive arthritis 3 .
Some prospective pilot studies suggested that using implementing interventional articulate cement spacer in the problematic joints after total knee arthroplasty (TKA) is effective in controlling infection and improving functional outcomes 3,[9][10][11] .
Therefore, this study aims assess the e cacy of application of a two-stage primary TKA to treat evolutive septic arthritis of knee joint, and contribute to existing literature and provide insights into this important topic for future research.

Participants
We retrospectively collected and investigated 11 eligible patients attending the Department of Orthopedic  14 Two-stage arthroplasty was performed for the treatment of advanced knee SA consecutively ( Figure 1). Surgery was conducted by a single surgical team. Deep infection was documented in all the 11 cases at the time of initial debridement. We did no apply any age or gender limitation. All the patients had their follow-up at our institution.

Surgical technique and perioperative care
First stage In the rst stage, a conventional medline knee skin incision was performed, followed by exposure of the joint through the parapatellar medial approach and complete removal of the synovial membrane, and infected and necrotic tissues. An intra-operative frozen section was obtained, and bacterial cultivation of the synovium and joint uid was performed for the con rmation of the diagnosis and selection of susceptible antibiotic. Three biopsies were obtained during the rst-stage procedure for postoperative pathological examination ( Figure 2). Bone cutting was performed in preparation for primary TKA according to the size of the spacer, and all cuts were performed under extramedullary cutting guides for the prevention of intramedullary contamination. Measurements were obtained using the full-length images of the lower limbs as shown in Figure 3. The line from point A (i.e., the center of the femoral condyle) to point D (i.e., the center of the femoral head) and the line from point B to ponint C (i.e., connecting line of the anterior superior iliac spine on both sides) were drawn for the cross point E, and the distance between point B and point E was measured and marked. During the operation, the line from points A and E were regarded as the virtual mechanical alignment of the lower limb. Finally, the cutting line of the distal femur was the vertical line of the line between points A and E. A spacer mold (Stage One; LINK Hamburg, Germany) was used in creating a dynamic antibiotic spacer from 80 gram of Palacos R + G bone cement (Heraeus Medica, Hanau, Germany) mixed with gentamycin (4.8 gram) and vancomycin (8 gram). The cement was mixed without vacuum with a "poor mixing technique" for the enhancement of antibiotic elution 15 . Antibiotic-loaded bone cement was poured into the spacer block mold at the liquid stage, shaped as the prosthesis and taken out at the curing period. The antibiotic spacer was then applied to the bone surfaces with a poor bonding technique, in which blood was allowed to get between the spacer and bone in order that a certain degree of motion in the knee is possible before cement hardening, to facilitate subsequent removal. The femoral spacer block was inserted rst and then an appropriate tibial spacer block adjusted with the extension gap was inserted in order that the limb was aligned and tissue tension was slightly relaxed for increasing post-operative range of motion ( Figure 4). After numerous irrigations the soft tissues were closed with the use of a suction drain for less than 48 hours.

Second stage
Total knee arthroplasty was performed at least 12 weeks after laboratory in ammation markers had normalised and there was no further clinical evidence of infection (reddened and hyperthermic joint, drainage from the wound and unreasonable knee pain) was observed. Intraoperative selection cemented posterior stabilised prosthesis (Genesis II™ total knee system, Smith & Nephew, Memphis, TN, USA) was implanted, and patellar resurfacing was performed instead of patella replacement ( Figure 5).

Antibiotic treatment and perioperative care
After the rst stage of surgery, a two-week intravenous anti-biotherapy was administered according to bacterial culture, and sensitive results obtained after vancomycin was used for those without culture and sensitive results. Oral antibiotics (rifampicin + optimal antibiotics/levo oxacin) were administered for six weeks until two successive CRP results were normal and ESR results were below twice the value of the high threshold or had a continuous downtrend with no positive signs and symptoms. The CRP and ESR were monitored every 3 days in the rst 2 weeks and every 2 weeks thereafter.
During the second stage of surgery, we applied 1 g of vancomycin for prophylaxis in the implementation of primary TKA surgical procedures 16 , and postoperatively vancomycin adminstration was continued until the wounds were dry. Then treatment was changed to oral rifampicin and levo oxacin for another 24 weeks 17 . Professional rehabilitation and functional training were implemented by rehabilitation Page 5/15 physiatrists in our hospital after the two-stage surgery, and all patients were followed up every 3 months in the rst year, and once per year thereafter.

Statistical analysis
Patient's les were screened by two authors for residual pain post-operative complications (including hematoma, seroma, blood transfusion, deep venous thrombosis, and revision surgery). Hospital for Special Surgery Knee (HSS) scores were obtained in pre-operation period (i.e., before the rst-stage of TKA surgery), interval period (i.e., after the rst-stage and before the second stage TKA surgery), and postoperation period (i.e., the latest follow-up) and were compared using analysis of variance (ANOVA) and post-hoc test of multiple group comparison. The effect of a two-stage operation was determined by comparing the degrees of pre-, interval-and post-knee range of motion (ROM) with paired t-tests. Twotailed independent t-test was applied to the comparisons between groups in terms of age (i.e., before or after middle age, following PubMed age ltration of 65 years old), gender, microorganism culture result (i.e., staphylococcus or negative) and comorbidities (i.e., have or not have). Data were reported as means ± standard deviation (SD). All statistical analyses were performed using SPSS 22.0 (IBM Corp., Armonk, NY, USA). The level of signi cance was set at p value of < 0.05.

Results
Eleven patients comprising six males and ve females were reported in this case-series (Table 1). Their mean age was 64.5 ± 5.4 years (range, 56-74 years). All patients had invasive treatments, such as knee injection (i.e., in 7/11), arthroscopic debridement (i.e., in 6/11) and open debridement (i.e., in 9/11) previously. Four patients had comorbidities that might compromise immunity: one patient was diabetic, two patients were hypertensive and another had both. Regarding the micro-organism culture, six were positive for staphylococcus and the other ve were negative, although utilized antibiotics were utilised variously.

CRP and ESR level
Infection was controlled successfully in all the patients by using the two-stage method for TKA. CRP level returned to normal level after 3.4 ± 1.4 weeks (range, 1-6 weeks), whereas ESR returned to a normal level after 16.3 ± 3.0 weeks (range, 13-23 weeks). The interval period was 14.4 ± 4.2weeks (range, 10-24 weeks) ( Table 1).
Comparisons of the HSS scores in these periods are shown in the Fig. 6a. The scores signi cantly increased (p<0.01) after both the rst stage (p<0.01)and second stage (p<0.01; Fig. 6a). There was no signi cant difference identi ed of the HSS score in these three timings in age, gender, microorganism culture result and comorbidities.

Discussion
Although therapeutic options for knee SA are limited, several studies applied two-stage primary TKA and attempted to improve its e ciency 2,3,11,18,19 , which encourage us to use the two-stage approach with a much shorter interval for the treatment of evolutive knee SA. In our study, elimination of infection, relief of symptoms and recovery of functions were ful lled. Moreover, we reported experiences and endings that were the same as those reported in previous studies despite our small sample size. Therefore, two-stage TKA is a promising option for severe knee SA management.
Patients enrolled to this study were all diagnosed with knee SA, and ful lled the criteria, particularly badly damaged knee joint, exfoliated cartilage and failure in controlling infection through arthroscopy , open debridement and appropriate antibiotic treatment. The operation mode was based from that of B Hochreiter et al. 10 . Regarding the femoral cutting, we used a simple extramedullary guiding method in rst stage to decrease the risk of surgery-related infections and improve function recovery. A previous study 20 reported that valgus angles in 30% to 51% of patients who received TKA exceeded the valgus cut angle (i.e., 5-7 degrees) and are thus unsuitable for intramedullary guide tools and may result in the malalignment of the knee after TKA. Thus, in the rst stage, we used a simple extramedullary guiding method in cutting the distal femur in all the 11 patients, which was equivalent to individualised valgus, and the alignment in these patients were appropriate. In addition, the extramedullary guiding method reduced the spread of infection and decreased the risk of bleeding and fat embolism. Further more, this extramedullary guiding method does not require C-arm uoroscopy for the femoral head during the operation and can thus reduce radiation injury.
The CRP levels of the 11 patients were returned to normal levels within 6 weeks after the rst stage, although ESR decreasd more slowly. Previous studies suggested that bio lm formed in the spacer surface due to the long interval period, which casued infection recurrence. Our patients were suggested to undergo the second stage of surgery within 3 months for them to satisfy the following criteria: CRP test results that are more than twice the normal range, ESR hal y decreased and absence of symptoms of infection such as joint redness, fever, wound drainage and unreasonable knee pain.
Infection were eliminated in the 11 patients, and their knee joint functions also returned to a level at which they were able to perform daily activities. However, patient 3 and 6 did not fully follow our advice to perform functional exercise after the rst stage and thus had rigid knee joints surrounded by scar tissues. Thus, we had to perform the tibial tubercle osteotomy to expose the joint clearly and complete the operation in the second stage of surgery. Nevertheless, they had smaller ROM than other patients after surgery. We emphasised the advantages of functional exercises to them, including knee exion and extension, lower limb muscle strength exercise and partial weight-bearing exercises, which accelerated their recovery post-operation. Knee joint function (i.e., HSS score) signi cantly increased after the twstage surgery. Therefore, function exercise should be performed in the entire process, especially within 1 month after each operation, because such activity affects the outcome of this surgery.
Previous studies on age and gender mainly focused on patients with osteoarthritis undergoing primary TKA and reported that females received primary TKA 3 years later than male patients and had lower knee function scores preoperatively and postoperatively 22 . Such ndings demonstrated that the early initiation of treatment may enhance post-operative outcomes in women. Another study 23 revealed that female gender, younger age and worsened preoperative pain predict increased risk of moderate to severe pain post-operatively in patients with primary and revision TKA.
However, information on factors in uencing the functional outcomes of two-stage TKA for knee SA is indeed scant. We did comparisons in term of age, gender, micro-organism culture result and comorbidities but failed to identify differences.
However, similar to many previous studies 2, 11,19,25 , our study has some limitations. First, our study was a retrospective study with a limited number of patients. The reason is that advanced knee arthritis combined with severe knee joint disability is uncommon and is di cult to treat. Second, we had no control group. Third, the average follow-up is not long enough, and we will continue to follow up further. Despite the small sample size, our study provided valuable information on this rare clinical challenge. We could concluded that the two-stage TKA with antibiotic-laden cement spacer is a potential option in the treatment for non-salvageable knee SA, and that extramedullary guiding method should be applied.

Conclusion
These results show that two-stage TKA is an a reliable alternative for the treatment of advanced knee SA, providing an encouraging nal clinical outcome in term of infection control, clinical severity relief and knee joint function improvement.
Abbreviations SA:septic arthritis; TKA:total knee arthroplasty; HSS: Hospital for Special Surgery; ROM:range of motion; CRP:C-reactive protein; ESR:erythrocyte sedimentation rate; Table   Table 1 is not available in this version of the manuscript. Figure 1 Images before surgery Images before surgery. (a) MRI PD-SPAIR image in Sagittal view; (b) X-ray image inanteroposterior (AP) view; (c) X-ray image in lateral view.

Figure 2
Open surgery on knee joint in rst stage A: intra-operative images of advanced septic arthritis of knee before debriding; B: image post debridement, extensive synovectomy and bone cutting.