Baseline data of all the two groups had no significant differences.
110 patients initially enrolled, of which 6 patients were excluded due to "dry aspiration", 4 cases with active rheumatoid arthritis, and 1 patient as acute PJI. Finally, a total of 99 patients were included in the present study, with 43 (43.4 %) as chronic PJI and 56 (56.6 %) were the aseptic failure. The baseline characteristics of the two groups were similar and shown in Table 2. There had no statistically significant difference in the demographics data of the two groups(P>0.05).
Table 2 The demographic data for the study population
characteristic
|
Septic(N=43)
|
Aseptic(N=56)
|
P value
|
Age (year)
|
70.58±5.26
|
68.98±6.05
|
0.17*
|
Gender
|
|
|
|
Male
|
22(51.2%)
|
27(48.2%)
|
0.77*
|
Female
|
21(48.8%)
|
29(51.8%)
|
|
Joint type
|
|
|
|
Hip
|
17(39.5%)
|
31(55.4%)
|
0.12*
|
Knee
|
26(60.5%)
|
25(44.6%)
|
|
BMI (Kg/m2)
|
22.74±3.91
|
23.21±4.66
|
0.59 #
|
Survival time of implatation(month)
|
67.91±29.06
|
76.82±21.55
|
0.08#
|
* chi-square test, # independent student’s T-test
Synovial fluid IL-4 had higher diagnose power than serum ESR and CRP
As shown in Table 3, the median value of serum ESR was 35.00mm/h in the chronic PJI group, which was higher than that in the aseptic group of 22.0 mm/h(P=0.001). Similarly, in the chronic PJI group, the serum CRP ranged from 5.80 to 91.20 mg/L, with a median value of 21.4 mg/L, which was higher than 6.75 mg/L of the aseptic group (P<0.0001). Results showed that the median level of SF-IL4 in the chronic PJI group is higher than that in the aseptic group( 3.30 vs.1.10 pg/ml, P<0.0001). Finally, the median SF-PMN% was higher in PJI patients (87.58%) than in the aseptic group(56.95%), with statistical significance(P<0.001).
Table 3 Analysis the character of single marker in chronic PJI and aseptic failure group
Marker
|
|
Chronic PJI(N=43)
|
Aseptic(N=56)
|
P-value
|
ESR(mm/hr)
|
Range
|
7.20~120.00
|
2.00~58.00
|
|
Median
|
35.00
|
22.00
|
|
P25, P75
|
22.00;50.00
|
17.50;32.00
|
|
Mean±SD
|
42.10±26.23
|
24.38±12.80
|
0.001#
|
Range
|
5.80~91.20
|
1.25~28.00
|
|
CRP(mg/L)
|
Median
|
21.40
|
6.75
|
|
P25, P75
|
13.20;33.40
|
3.29;16.00
|
|
Mean±SD
|
27.31±20.02
|
9.81±7.75
|
<0.001#
|
Range
|
1.20~14.10
|
0.09~2.88
|
|
SF-IL4(pg/mL)
|
Median
|
3.30
|
1.10
|
|
P25, P75
|
2.20;9.7
|
0.55;1.28
|
|
Mean±SD
|
5.57±4.37
|
1.00±0.53
|
<0.001#
|
Range
|
60.32-96.65
|
23-89.20
|
|
SF-PMN%
|
Median
|
87.58
|
56.95
|
|
P25, P75
|
83.50,90.50
|
53.40,56.95
|
|
Mean±SD
|
86.15±7.31
|
59.30±15.53
|
<0.001#
|
ESR, Erythrocyte sedimentation rate; CRP, C-reaction protein; SF-IL4, Synovial fluid Interleukin 4; SF-PMN%, Synovial fluid polymorphonuclear cell neutrophil percentage
SD: standard deviation; # Mann-Whitney U test
The ROC curves were used to measure the discriminatory strength of those indicators (Figure 1a-d). The AUC of SF-IL4 was 0.97(95% confidence interval,95%CI,0.92-0.99),is higher than serum ESR 0.72(95%CI,0.62-0.84)(P=0.0004), serum CRP 0.83(95%CI,0.74-0.90)(P<0.0001), and SF-PMN% 0.89(95%CI,0.82-0.95)(P=0.053)(Figure 1e). The sensitivity, specificity, PPV, NPV, +LR, -LR, and DOR of those markers were as Table 4 shown. The optimal cut-off value for SF-IL4 of 1.7pg/mL, with sensitivity of 93.02% (95% CI, 80.9% -98.5%), specificity of 94.64% (95% CI, 85.1%-98.9%), and with a high PPV,NPV,DOR of 93.0% ,94.6% and 248 individually. When the cut-off value for serum ESR was set as 34mm/hr, with a poor sensitivity of 58.14% (95%CI, 42.1%-73.0%) and moderate specificity of 83.93% (95%CI, 71.7%-92.4%). When 18mg/L was the cut-off value for serum CRP, with sensitivity about 62.79% (95%CI, 46.7%-77.0%) and specificity about 87.50% (95%CI, 75.9%-94.8%). And those values for SF-PMN% were 95.35% (95%CI, 84.2%-99.4%) and 78.57% (95%CI, 65.6%-88.4%) with the cutoff value of 75%.
Table 4 The AUC、Cut-off level、Sensitivity, Specificity, PPV, NPV, +LR、-LR、and DOR of each single marker
Test
|
CRP (mg/L)
|
ESR (mm/hr)
|
CRP (mg/L)
|
ESR (mm/hr)
|
SF-IL4(pg/mL)
|
SF-PMN%
|
AUC(95%CI)
|
|
|
0.83(0.74-0.90)
|
0.72(0.62-0.84)
|
0.97(0.92-0.99)
|
0.89(0.82-0.95)
|
Cutoff level
|
10
|
30
|
18
|
34
|
1.7
|
75%
|
Sensitivity(95%CI)
|
81.4
|
67.4
|
62.79(46.7-77.0)
|
58.14(42.1-73.0)
|
93.02 (80.9- 98.5)
|
95.35(84.2-99.4)
|
Specificity(95%CI)
|
58.9
|
62.6
|
87.50(75.9-94.8)
|
83.93(71.7-92.4)
|
94.64(85.1-98.9)
|
78.57(65.6-88.4)
|
PPV
|
-
|
-
|
79.4(65.0- 88.9)
|
73.5(59.2-84.2)
|
93.0(81.6 -97.6)
|
77.4(67.3-85.0)
|
NPV
|
-
|
|
75.4(67.0- 82.10)
|
72.3(64.3 -79.1)
|
94.6(85.6 -98.1)
|
95.7(85.0-98.8)
|
+LR
|
-
|
--
|
5.02(2.40- 10.40)
|
3.62(1.9-6.9)
|
17.36(5.8 -52.4)
|
4.45(2.7-7.4)
|
-LR
|
-
|
-
|
0.43(0.3-0.6)
|
0.50(0.3- 0.7)
|
0.07(0.02 -0.2)
|
0.06(0.02-0.2)
|
DOR
|
-
|
-
|
11.67
|
7.24
|
248.00
|
74.17
|
Accurancy
|
58.6
|
50.5
|
72.7
|
68.7
|
92.9
|
84.8
|
ESR, Erythrocyte sedimentation rate; CRP, C-reaction protein; SF_IL4, Synovial fluid Interleukin 4; SF-PMN%, Synovial fluid polymorphonuclear cell neutrophil percentage; PPV, Positive Predictive Value; NPV, Negative Predictive Value; +LR, Positive Likelihood Ratio; -LR, Negative Likelihood Ratio; DOR, Diagnostic Odds Ratio;
Combination SF-IL4 with SF-PMN% improved diagnostic ability of chronic PJI
Next, We combined SF-IL4 with other biomarkers. As results are shown in Table 5, when met the cut-off of SF-IL-4 and SF-PMN% at the same time, which can maximize the specificity to 97% and accuracy to 96%, but decreased the sensitivity to 91% for chronic PJI diagnosis.
Table 5 The Sensitivity, Specificity, PPV, NPV, and Accuracy of Combination the Different Markers
Conbination
|
SF-IL4>1.7 +CRP>18
|
SF-IL4>1.7+ SF-PMN%>75%
|
CRP>18+SF-IL4>1.7+ SF-PMN%>75%
|
Sensitivity
|
0.88
|
0.91
|
0.53
|
Specificity
|
0.98
|
0.97
|
0.98
|
PPV
|
0.96
|
0.94
|
0.95
|
NPV
|
0.75
|
0.92
|
0.73
|
Accuracy
|
80.1%
|
96%
|
78.8%
|
CRP, C-reaction protein; SF-IL4, Synovial fluid Interleukin 4; SF-PMN%, Synovial fluid polymorphonuclear cell neutrophil percentage; PPV, Positive Predictive Value; NPV, Negative Predictive Value.