General information and patients’ follow-up
Demographic information and follow-up results are shown in Table 1. There were 54 (49.5%) males in non-reinfection group and 12 (54.55%) in reinfection group. The mean age was 58.44 ± 13.28 years in the success group and 64.50 ± 11.07 years in the re-infected group. The mean body mass index (BMI) was 25.04 ± 3.60 kg/m2 in the success group compared with 24.92 ± 3.31 kg/m2 in the reinfection group. Eighty-one hips and sixty knees were included in the analysis. There was no significant difference in the prevalence of sinus occurrence between the reinfection group (50.00%) and the non-reinfection group (35.29%, p=0.191).
The mean follow-up year was 3.45 (range, 1.32 to 8.70 years) years in the success group and 3.62 (range, 0.65 to 7.32 years) years in re-infected group. The interval of spacer insertion was 23.39 (range, 4.57 to 110.86 weeks) weeks in the success group and 26.00 (range, 5.29 to 143.71 weeks) weeks in reinfection group. There were 22 patients re-infected after reimplantation in two-stage exchange arthroplasty, and the total success rate was 84.40%.
Details of the causative organisms in 22 infection patients are shown in Table 2. Coagulase-negative Staphylococcus (CNS) species were the most common causative bacteria for recurrent PJI (6, 27.27%). The rest of the causative organisms included 1 (4.55%) Staphylococcus aureus, 1 (4.55%) Enterococcus faecalis, 1 (4.55%) Gram-negative Bacillus, 1 methicillin-resistant Staphylococcus aureus (MRSA), 5 (22.73%) polymicrobial organisms and 2 (9.09%) other organisms. Five (22.73%) patients had negative culture results but were diagnosed with PJI according to the MSIS criteria at the time of reinfection.
Was there any difference in values of serum markers between the success and reinfection groups?
The values of the ESR, IL-6, CRP and fibrinogen were compared between success patients and reinfection patients. The details of each serum marker are shown in Table 3. No significant difference was found in these four serum markers between the success group and the reinfection group at the time of resection and reimplantation. At the time of resection, the median CRP was 18.00 mg/L (IQR, 7.95-35.50 mg/L) in the success group versus 25.50 mg/L (IQR, 16.89-38.18 mg/L) in the reinfection group (p= 0.087), and the median ESR was 35 mm/hr (IQR, 23.00-58.00 mm/hr) for those patients who remained infection-free versus 48 mm/hr (IQR, 30.25-75.75 mm/hr) for reinfection patients (p = 0.124). At the time of reimplantation, the median CRP was 3.48 mg/L (IQR, 2.03-8.40 mg/L) in the success group versus 3.32 mg/L (IQR, 1.60-15.30 mg/L) in the reinfection group (p= 0.643), and the median ESR was 12.00 mm/hr (IQR, 7.00-18.00 mm/hr) for those patients who remained infection-free versus 12.00 mm/hr (IQR, 7.00-42.50 mm/hr) for reinfection patients (p = 0.214).
Can the percent changes or value changes in serum markers guide the timing of reimplantation?
Comparison of percent changes and value changes in non-reinfection group versus reinfection group was shown in Table 4. No significant difference was found in either the value change or percent change in these four serum markers. When comparing preresection with prereimplantation values, the median value changes in the CRP was 11.95 mg/L (IQR, 2.01-28.31 mg/L) in the non-reinfection group and 17.58 mg/L (IQR, 0.45-36.61 mg/L) in the reinfection group (p=0.627). The median value changes in the IL-6 was 7.91 pg/L (IQR, 4.26-13.76 pg/L) for those patients who remained infection-free and 8.81 pg/L (IQR, 3.55-15.56 pg/L) for those who experienced reinfection (p=0.948). The median value changes in the fibrinogen was 1.32 g/L (IQR, 0.69-1.95 g/L) in the infection-free group versus 1.85 g/L (IQR, 0.37-2.95 g/L) in the reinfection group (p=0.200). In addition, the median value changes in the ESR was 25.00 mm/hr (IQR, 9.00-40.00 mm/hr) for those patients who remained infection-free and 26.00 mm/hr (IQR, 5.75-45.50 mm/hr) for those who re-infected (p=0.901).
With regard to percent changes from the resection to the reimplantation, the median percent change in the CRP was 29.99% (IQR, 11.12-74.81%) for those patients who remained infection-free and 16.04% (IQR, 7.02-107.32%) for those who experienced reinfection (p=0.595). The median percent changes in the ESR was 30.77% (IQR, 16.85-55.39%) in the non-reinfection group and 29.91% (IQR, 20.49-78.31%) in the reinfection group (p=0.611). The median value changes in the fibrinogen was 71.66% (IQR, 61.69-83.50%) in the infection-free group versus 63.03 (IQR, 54.03-92.08%) in the reinfection group (p=0.242). In addition, the median value changes in the IL-6 was 29.58% (IQR, 16.37-55.39%) for those patients who remained infection-free and 36.45% (IQR, 16.00-55.56%) for those who re-infected (p=0.611).
Furthermore, we generated a receiver operator curve (ROC) for value changes (figure 1) and percent changes (figure 2) in these four serum markers. The area under the receiver operator curve (AUC) was 0.533 for CRP, 0.504 for IL-6, 0.508 for the ESR and 0.586 for fibrinogen for predicting failure based on value changes, demonstrating that value changes in these four serum markers were poor indicators. With regard to percent changes for determining the timing of reimplantation, the AUC indicated that percent changes in the CRP (AUC=0.464), the IL-6 (AUC=0.534), the ESR (AUC=0.527) and the fibrinogen (AUC=0.586) were all poor markers when predicted persistent or recurrent infection. As a result of the low sensitivity and specificity caused by the wide distribution of changes in inflammatory levels, a threshold value could not be calculated by Youden’s J-statistic.