Patients characteristics
The characteristics of training and validation datasets are listed in Table 1. Compared the two datasets, the morbidities of severe complications were LCOS (9.46% vs 5.33%, P<0.05), AKI-rH (4.48% vs 7.29%, P<0.05) and MODS (4.95% vs 4.49%, P>0.05).
Table 1. Patient characteristics
Characteristics
|
Training dataset
|
Validation dataset
|
P
|
(n=930)
|
(n=713)
|
Demographics
|
|
|
|
Age (y)
|
47.91±13.83
|
49.68±15.00
|
0.001
|
Gender (female, No. %)
|
485(52.15%)
|
343(48.11%)
|
0.115
|
Height (cm)
|
160.73±8.18
|
160.49±10.69
|
0.939
|
Weight (kg)
|
54.66±10.39
|
56.96±11.95
|
<0.01
|
BMI
|
21.08±3.29
|
21.97±3.62
|
<0.01
|
BSA (m2)
|
6.87±1.29
|
7.15±1.49
|
<0.01
|
Smoke (No. %)
|
166 (17.85%)
|
95(13.32%)
|
0.016
|
Medical histories
|
|
|
|
CF (<4 weeks, No. %)
|
601(64.62%)
|
421(59.05%)
|
0.024
|
Endocarditis (No. %)
|
88(9.46%)
|
120(16.83%)
|
<0.01
|
Diabetes (No. %)
|
48(5.16%)
|
49(6.87%)
|
0.176
|
Hypertension (No. %)
|
122(13.12%)
|
129(18.09%)
|
0.007
|
Hepatitis (No. %)
|
78(8.39%)
|
23(3.23%)
|
<0.01
|
Pulmonary disease (No. %)
|
78(8.39%)
|
30(4.21%)
|
0.001
|
Dialysis (No. %)
|
0(0.00%)
|
0(0.00%)
|
<0.01
|
PVD (No. %)
|
0(0.00%)
|
0(0.00%)
|
<0.01
|
Re-operation (No. %)
|
58(6.24%)
|
52(7.29%)
|
0.453
|
Laboratory values
|
|
|
|
WBC (×109/l)
|
7.04±2.46
|
7.22±2.38
|
0.076
|
PLT (×1012/l)
|
213.26±66.44
|
217.74±82.53
|
0.789
|
RBC (×109/l)
|
4.68±0.70
|
4.58±0.76
|
0.003
|
RBC-DW
|
0.14±0.02
|
0.14±0.03
|
<0.001
|
<0.12
|
3(0.32%)
|
1(0.14%)
|
|
0.12-0.15
|
772(83.01%)
|
560(78.54%)
|
|
>0.15
|
148(15.91%)
|
151(21.18%)
|
|
Hb (g/l)
|
133.78±19.63
|
130.54±21.53
|
0.001
|
ALT (u/l)
|
25.79±33.6
|
26.94±29.51
|
0.998
|
ALB (g/l)
|
42.29±18.01
|
39.36±4.92
|
<0.01
|
TBil (mmol/l)
|
15.88±10.39
|
16.52±9.76
|
0.001
|
BUA (mg/l)
|
374.78±122.51
|
423.22±140.87
|
<0.01
|
BUN (mmol/l)
|
6.12±2.51
|
6.15±2.93
|
0.249
|
<2.9
|
24(2.58%)
|
22(3.09%)
|
|
2.9-8.6
|
802(86.24%)
|
606(84.99%)
|
|
>8.6
|
95(10.22%)
|
84(11.78%)
|
|
BCr (umol/l)
|
77.19±27.07
|
86.11±68.24
|
0.01
|
<50
|
56(6.02%)
|
43(6.03%)
|
|
50-115
|
808(86.88%)
|
609(85.41%)
|
|
116-200
|
55(5.91%)
|
51(7.15%)
|
|
>200
|
3(0.32%)
|
9(1.26%)
|
|
BUN/BCr
|
0.08±0.04
|
0.08±0.03
|
<0.01
|
<0.055 (No. %)
|
127(13.66%)
|
143(20.06%)
|
|
0.055-0.075 (No. %)
|
308(33.12%)
|
250(35.06%)
|
|
>0.075 (No. %)
|
486(52.26%)
|
319(44.74%)
|
|
CCr (ml/min/1.73m2, No.)
|
79.62±34.28
|
76.17±26.37
|
0.046
|
<50 (No. %)
|
116(12.47%)
|
98(13.74%)
|
|
50-80 (No. %)
|
396(42.58%)
|
338(47.41%)
|
|
>80 (No. %)
|
404(43.44%)
|
275(38.57%)
|
|
APTT (secs.)
|
29.3±5.72
|
31.74±6.75
|
<0.01
|
Fbg (g/l)
|
3.08±1.16
|
3.08±1.20
|
0.419
|
ESR (mm)
|
23.49±20.45
|
26.16±23.34
|
0.192
|
ECG measurements
|
|
|
|
Atrial fibrillation (No. %)
|
389(41.83%)
|
235(32.96%)
|
<0.01
|
UCG measurements
|
|
|
|
LVD (mm)
|
54.51±11.10
|
53.65±10.97
|
0.220
|
EF (%)
|
62.71±10.14
|
63.95±9.93
|
0.005
|
>50
|
777(83.55%)
|
612(85.83%)
|
|
30-50
|
93(10%)
|
72(10.1%)
|
|
<30
|
4(0.43%)
|
1(0.14%)
|
|
PASP (mmHg)
|
21.86±27.54
|
45.59±17.4
|
<0.01
|
>60
|
80(8.6%)
|
86(12.06%)
|
|
30-60
|
310(33.33%)
|
380(53.3%)
|
|
<30
|
540(58.06%)
|
57(7.99%)
|
|
Intraoperative variables
|
|
|
|
AOT (min)
|
80.23±34.70
|
90.0±46.70
|
0.001
|
ACPBT (min)
|
37.7±22.50
|
58.4±40.20
|
<0.01
|
Defibrillation (freq.)
|
|
|
0.351
|
<1
|
773(83.12%)
|
605(84.85%)
|
|
≥1
|
157(16.88%)
|
108(14.79%)
|
|
Surgical approaches
|
|
|
|
AVR (No. %)
|
378(40.65%)
|
293(41.09%)
|
0.894
|
MVR (No. %)
|
684(73.55%)
|
432(60.59%)
|
<0.01
|
TVR (No. %)
|
33(3.55%)
|
31(4.35%)
|
0.483
|
MVP (No. %)
|
57(6.13%)
|
84(11.78%)
|
<0.001
|
TVP (No. %)
|
298(32.04%)
|
303(42.5%)
|
<0.001
|
CABG (No. %)
|
28(3.01%)
|
32(4.49%)
|
0.147
|
RFA (No. %)
|
27(2.9%)
|
33(4.63%)
|
0.086
|
other cardiac surgery (No. %)
|
60(6.45%)
|
210(29.45%)
|
<0.001
|
non-cardiac surgery (No. %)
|
4(0.43%)
|
0(0.00%)
|
0.138
|
Severe complications
|
|
|
|
LCOS (No. %)
|
88(9.46%)
|
38(5.33%)
|
0.002
|
AKI-rH (No. %)
|
45(4.84%)
|
52(7.29%)
|
0.047
|
MODS (No. %)
|
46(4.95%)
|
32(4.49%)
|
0.752
|
Mechanical assistant
|
|
|
|
IABP /ECMO (No. %)
|
33(3.55%)
|
29(4.07%)
|
0.677
|
Discharge status
|
|
|
|
Death (No. %)
|
61(6.56%)
|
47(6.59%)
|
1.000
|
The age in training dataset is younger than validation dataset (47.91±13.83 vs 49.68±15 years, P<0.05), cardiac failure (CF) was also shorter (64.62% vs 59.05%, P<0.05). However, the morbidities of preoperative pulmonary disease (PD) and hepatitis of training dataset were higher than that of validation dataset (PD: 8.39% vs 4.21%, P<0.05; hepatitis: 8.39% vs 3.23%, P<0.05).
More patients had history of endocarditis in validation dataset than training dataset (16.83% vs 9.46%, P<0.05), important organs, such as liver and kidney suffer more damages. In addition, according to the result of echocardiography, the validation dataset has higher pulmonary artery systolic pressure (PASP) than training dataset (45.59±17.4 mmHg vs 21.86±27.54 mmHg, P<0.05).
Furthermore, the AOT and ACPBT of training dataset are both shorter than validation dataset (AOT: 80.23±34.7 mins vs 90.0±46.7 mins, P<0.01; ACPBT: 37.7±22.5 mins vs 58.4±40.2 mins, P<0.01).
Prediction model for LCOS
The PRF model for LCOS includes BCr (OR 1.85; 95%CI 0.95-3.59), creatinine clearance rate (CCr)(OR 0.46; 95%CI 0.32-0.67 ), hemoglobin (Hb)(OR 0.73; 95%CI 0.58-0.91), PAH (OR 1.34; 95%CI 0.96-1.86), and hypertension (OR 1.70; 95%CI 0.94-3.05) (Table 2). As a comparison, the PIRF model only includes CCr (OR 0.38; 95%CI 0.27-0.53) and ACPBT (OR 1.80; 95%CI 1.52-2.12).We applied both models to the validation dataset. The AUC of the PIRF model is 0.821 (0.747, 0.896), which is statistically higher (P<0.01) than that 0.565 obtained in the PRF model (Figure 2, Table 5).
Table 2. Prognostic models for LCOS in development dataset
Variables
|
PRF model
(n=930)
|
PIRF model
(n=930)
|
β
|
OR (95% CI)
|
P
|
β
|
OR (95% CI)
|
P
|
Intercept
|
-2.4909
|
0.08
|
0.015
|
-0.3645
|
0.69
|
0.311
|
BCr
|
0.6152
|
1.85 (0.95-3.59)
|
0.068
|
|
|
|
CCr
|
-0.7756
|
0.46 (0.32-0.67)
|
<0.01
|
-0.9801
|
0.38 (0.27-0.53)
|
<0.01
|
Hb
|
-0.3191
|
0.73 (0.58-0.91)
|
0.006
|
|
|
|
PASP
|
0.2929
|
1.34 (0.96-1.86)
|
0.082
|
|
|
|
Hypertension
|
0.5281
|
1.70 (0.94-3.05)
|
0.079
|
|
|
|
ACPBT
|
|
|
|
0.5855
|
1.80 (1.52-2.12)
|
<0.01
|
Prediction model for AKI-rH
The PRF model for AKI-rH includes CCr (OR 0.33; 95%CI 0.21-0.52), red blood cell distribution width (RBC-DW)(OR 2.59; 95%CI 1.31-5.13) and total bilirubin (TBil) (OR 1.51; 95%CI 1.20-1.90)(Table 3). As a comparison, the PIRF model includes CCr (OR 0.36; 95%CI 0.22-0.57), RBC-DW(OR 2.19; 95%CI 1.08-4.43), TBil (OR 1.52; 95%CI 1.21-1.92) and ACPBT (OR 1.50; 95%CI 1.23-1.82). We applied both models to the validation dataset. The AUC of the PIRF model is 0.78 (0.717, 0.843), which is statistically higher (P<0.01) than that 0.688 obtained in the PRF model (Figure 2, Table 5).
Table 3. Prognostic models for AKI-rH in development dataset
Variables
|
PRF model
(n=930)
|
PIRF model
(n=930)
|
β
|
OR (95%CI)
|
P
|
β
|
OR (95% CI)
|
P
|
Intercept
|
-2.9392
|
0.05
|
0.002
|
-2.8605
|
0.06
|
0.004
|
CCr
|
-1.1041
|
0.33 (0.21-0.52)
|
<0.01
|
-1.0247
|
0.36 (0.22-0.57)
|
<0.01
|
RBC-DW
|
0.9530
|
2.59 (1.31-5.13)
|
0.006
|
0.7835
|
2.19 (1.08-4.43)
|
0.030
|
TBil
|
0.4093
|
1.51 (1.20-1.90)
|
0.001
|
0.4206
|
1.52 (1.21-1.92)
|
<0.01
|
ACPBT
|
|
|
|
0.4042
|
1.50 (1.23-1.82)
|
<0.01
|
Prediction model for MODS
The PRF model for MODS includes CCr (OR 0.28; 95%CI 0.18-0.45), BUN/BCr (OR 1.81; 95%CI 1.11-2.95), Hb (OR 0.74; 95%CI 0.55-1.01), heart failure history (OR 1.84; 95%CI 0.82-4.16) and PD (OR 3.33; 95%CI 1.55-7.16) (Table 4). As a comparison, the PIRF model includes CCr (OR 0.29; 95%CI 0.17-0.48), BUN/BCr (OR 1.86; 95%CI 1.1-3.14), CF (<4 weeks) (OR 1.95; 95%CI 0.83-4.58), PD (OR 4.69; 95%CI 2.10-10.47), ACPBT (OR 1.71; 95%CI 1.41-2.09) and combined with tricuspid valve replacement (cTVR) (OR 3.69; 95%CI 1.16-11.47). We applied both models to the validation dataset. The AUC of the PIRF model is 0.774 (0.70, 0.847), which is statistically higher (P<0.01) than that 0.657 obtained in the PRF model (Figure 2, Table 5).
Table 4. Prognostic models for MODS in development dataset
Variables
|
PRF model
(n=930)
|
PIRF model
(n=930)
|
β
|
OR (95% CI)
|
P
|
β
|
OR (95% CI)
|
P
|
|
-2.5690
|
0.08
|
0.002
|
-3.0533
|
0.05
|
0.001
|
CCr
|
-1.2645
|
0.28 (0.18-0.45)
|
<0.01
|
-1.2457
|
0.29 (0.17-0.48)
|
<0.01
|
BUN/BCr
|
0.5907
|
1.81 (1.11-2.95)
|
0.018
|
0.6219
|
1.86 (1.10-3.14)
|
0.020
|
Hb
|
-0.296
|
0.74 (0.55-1.01)
|
0.057
|
|
|
|
CF
|
0.6110
|
1.84 (0.82-4.16)
|
0.141
|
0.6660
|
1.95 (0.83-4.58)
|
0.127
|
PD
|
1.2038
|
3.33 (1.55-7.16)
|
0.002
|
1.5459
|
4.69 (2.10-10.47)
|
<0.01
|
ACPBT
|
|
|
|
0.5381
|
1.71 (1.41-2.09)
|
<0.01
|
cTVR
|
|
|
|
1.3049
|
3.69 (1.16-11.47)
|
0.027
|
Table 5. Comparisons of PRF and PIRF models for three complications in validation dataset
Complications
|
AUC
|
PRF model
(n=713)
|
PIRF model
(n=713)
|
P
|
LCOS
|
0.565 (0.466, 0.664)
|
0.821 (0.747, 0.896)
|
<0.01
|
AKI-rH
|
0.688 (0.62, 0.757)
|
0.78 (0.717, 0.843)
|
<0.01
|
MODS
|
0.657 (0.563, 0.751)
|
0.774 (0.7, 0.847)
|
0.003
|