General characteristics
Overall, A total of 1188 participants were included in the study with 583 and 605 participants for the TC and VC, respectively. Flow chart for participant inclusion/exclusion are provided in Fig. 1. The clinical characteristics of patients in the TC and VC are listed in Table 1. The median age of patients in the global cohort was 52 (25th to 75th percentiles, 42–59) years, among whom 701 (59.0%) were male, 1068 (89.9%) was Child-Pugh A class, 576 (48.5%) had varices, and 146 (12.3%) had HRV. The most common cause of cirrhosis was hepatitis B virus infection, which was found in 415 of the 583 patients (71.2%) in the TC and in 446 of the 605 participants (73.7%) in the VC. The VC had similar baseline characteristics to the TC. There was no statistical difference between the TC and VC. Comparison of patients with and without HRV in the entire cohort is also displayed in Supporting Table S1.
Table 1
General characteristics and comparison of TC and VC
Variable
|
All
(n = 1188)
|
TC
( n = 583)
|
VC
(n = 605)
|
Pvalue
|
Age,years
|
52 (42,59)
|
52 (42,59)
|
52 (41,60)
|
0.791
|
Man, n (%)
|
701,(59.0)
|
355,(60.9)
|
346,(57.2)
|
0.195
|
BMI, Kg/m2
|
24.9 (22.8,27.3)
|
25.0 (22.9,27.4)
|
24.8 (22.7,27.3)
|
0.250
|
Etiology, n (%)
|
-
|
-
|
-
|
0.552
|
HBV
|
861,(72.5)
|
415,(71.2)
|
446,(73.7)
|
-
|
HCV
|
169,(14.2)
|
89,(15.3)
|
80,(13.2)
|
-
|
Non-viral infectious
|
158,(13.3)
|
79,(13.5)
|
79,(13.1)
|
-
|
C-P A, n (%)
|
1068,(89.9)
|
530,(90.9)
|
538,(88.9)
|
0.257
|
ALT, U/L
|
49(26,118)
|
48(26,113)
|
49(25,125)
|
0.742
|
AST, U/L
|
45(26,102)
|
44(26,97)
|
47(26,111)
|
0.292
|
ALB, g/L
|
43.1(39.0,46.2)
|
43.5(39.4,46.1)
|
42.8(38.8,46.4)
|
0.189
|
TBIL, µmol/L
|
17.4(13.1,24.6)
|
17.4(13.2,23.9)
|
17.3(13.1,25.7)
|
0.753
|
INR
|
1.05(1.00,1.14)
|
1.05(1.00,1.13)
|
1.05(1.00,1.15)
|
0.666
|
PLT, ×109/L
|
135(98,174)
|
135(101,174)
|
135(98,172)
|
0.514
|
LS, kPa
|
14.3(8.8,24.5)
|
14.0(8.8,25.7)
|
15.0(9.0,23.6)
|
0.369
|
SA, cm2
|
42.94(34.65,56.25)
|
42.56(33.95,56.25)
|
43.32(35.12,55.90)
|
0.460
|
Varices, n (%)
|
576,(48.5)
|
282,(48.4)
|
294,(48.6)
|
0.938
|
HRV, n (%)
|
146,(12.3)
|
69,(11.8)
|
77,(12.7)
|
0.64
|
Abbreviations: TC, Training cohort; VC, Validating cohort; HRV, High-risk varices; BMI, Body mass index; HBV, Hepatitis B virus; HCV, Hepatitis C virus; C-P, Child-Pugh class; ALT, Alanine aminotransferase; AST, Aspartate aminotransferase; ALB, Albumin; TBIL, Total bilirubin; INR, International normalized ratio; PLT, Platelet count; LS, Liver stiffness; SA, Spleen area
Predictors of HRV in the TC
The univariate analysis was performed to detect the variables associated with the presence of HRV. Table 2 showed that the variables age, Albumin, total bilirubin, international normalized ratio, PLT, LS, and SA were significantly different (all P < 0.05). The ROC analysis showed that the area under the receiver operating characteristic (AUROC) of age was 0.605 (95% confidence interval [CI]: 0.564–0.645), and the best cut-off value was 52 years of age. To create an “easy-to-use” algorithm, 50 years of age was selected for the risk stratification. Age was evaluated as dichotomous data in the multivariate analysis. The risk of HRV in patients over 50 years was 2.387 times those under 50 years of age (Fig. 1-A).
Table 2
Variables
|
Univariate analysis
|
Multivariate analysis
|
HRV +(n = 69)
|
HRV -(n = 514)
|
P value
|
OR, 95% CI
|
Pvalue
|
Age,years
|
55(48,62)
|
51(41,58)
|
0.004
|
-
|
-
|
Age ≥ 50, n (%)
|
49(71.0)
|
287(55.8)
|
0.017
|
2.387,1.282–4.443
|
0.006
|
Man, n (%)
|
43(62.3)
|
312(60.7)
|
0.796
|
-
|
-
|
BMI,kg/m2
|
24.9(22.9,27.6)
|
25.0(22.9,27.3)
|
0.964
|
-
|
-
|
C-P A, n (%)
|
63,(91.3)
|
467, (90.9)
|
0.903
|
-
|
-
|
ALT, U/L
|
42(27,77)
|
49(26,119)
|
0.258
|
-
|
-
|
AST, U/L
|
52(28,92)
|
44(26,99)
|
0.515
|
-
|
-
|
ALB, g/L
|
40.7(37.4,45.4)
|
43.7(39.8,46.3)
|
0.002
|
-
|
-
|
TBIL, µmol/L
|
18.9(15.3,27.1)
|
17.1(12.9,23.6)
|
0.017
|
-
|
-
|
INR
|
1.11(1.05,1.27)
|
1.04(0.99,1.12)
|
< 0.001
|
-
|
-
|
PLT, ×109/L
|
85(65,116)
|
141(108,179)
|
< 0.001
|
0.986,0.979–0.993
|
0.000
|
LS, kPa
|
27.0(16.9,36.8)
|
13.1(8.3,21.6)
|
< 0.001
|
1.034,1.014–1.053
|
0.001
|
SA, cm2
|
59.04(47.18,72.8)
|
41.19(32.99,53.75)
|
< 0.001
|
1.018,1.003–1.033
|
0.017
|
Abbreviations: HRV, High-risk varices; TC, Training cohort; OR, Odds ratio; CI, Confidence interval; BMI, Body mass index; C-P, Child-Pugh class; ALT, Alanine aminotransferase; AST, Aspartate aminotransferase; ALB, Albumin; TBIL, Total bilirubin; INR, International normalized ratio; PLT, Platelet count; LS, Liver stiffness; SA, Spleen area.
Multivariateanalysis showed that age ≥ 50 years, LS, PLT, and SA were independently associated with HRV (Supporting Table S4). In the TC, the predicting performance of PLT, SA and LS for HRV was assessed by a receiver operating curve (ROC) analysis. The AUROC of PLT, SA and LS were 0.781 ([CI]: 0.745–0.814), 0.731 ([CI]: 0.694–0.767), and 0.735 ([CI]: 0.697–0.770), respectively. The Delong test showed that there was no statistical difference between PLT, SA, and LS (Fig. 1-B). Therefore, SA could replace LS to establish a new criterion to predict HRV.
The modified Baveno VI criteria based on age stratification
Table 3 shows the performance of the exploratory data and the new age-based criteria by adjusting the cutoff value of LS and PLT in each age-related subgroup. In patients < 50 years of age, the criteria of PLT > 100×109/L and LS < 30kPa maximized the number of potential EGDs that were avoided while keeping the risk of missing a HRV below the 5% threshold. For patients aged ≥ 50 years, the same result was obtained with a combination of PLT > 125×109/L and LS < 20 kPa. We have named these new classification rules the MB6C.
Table 3
Performance of the exploratory data and MB6C and AB6C classification rules based on age in the TC
|
|
Patients age < 50
( n = 247)
|
|
Patients age ≥ 50
( n = 336)
|
Variables
|
|
EGD spared,
n, (%)
|
HRV missed,
n, (%)
|
|
EGD spared,
n, (%)
|
HRV missed,
n, (%)
|
PLT
|
> 115
|
171 (69.2)
|
3 (1.8)
|
> 125
|
179 (53.3)
|
12 (6.7)
|
|
> 110
|
175 (70.9)
|
3 (1.7)
|
> 120
|
192 (57.1)
|
13 (6.8)
|
|
> 100
|
193 (78.1)
|
3 (1.6)
|
> 115
|
207 (61.6)
|
14 (6.8)
|
LS
|
< 30
|
204 (82.6)
|
9 (4.4)
|
< 25
|
251 (74.7)
|
23 (9.2)
|
|
< 25
|
182 (73.7)
|
5 (2.7)
|
< 20
|
217 (64.6)
|
18 (8.3)
|
|
< 20
|
164 (66.4)
|
4 (2.4)
|
-
|
-
|
-
|
SA
|
< 55
|
172 (69.6)
|
6 (3.5)
|
< 55
|
249 (74.1)
|
21 (8.4)
|
|
< 50
|
155 (62.8)
|
3 (1.9)
|
< 50
|
222 (66.1)
|
17 (7.7)
|
|
< 44
|
125 (50.6)
|
2 (1.6)
|
< 44
|
178 (53.0)
|
11 (6.2)
|
MB6C
|
PLT > 100
+ LS < 30
|
171 (69.2)
|
1 (0.6)
|
PLT > 125
+ LS < 20
|
145 (43.2)
|
6 (4.1)
|
AB6C
|
PLT > 100
+ SA < 55
|
158 (64.0)
|
3 (1.9)
|
PLT > 125
+ SA < 44
|
129 (38.4)
|
6 (4.7)
|
Abbreviations: MB6C, Modified Baveno VI criteria;AB6C, Alternative Baveno VI criteria; HRV,High-risk varices; TC, Training cohort; EGD, Esophagogastroduodenoscopy; PLT, Platelet count (× 109/L); LS, Liver stiffness (kPa); SA, Spleen area (cm2).
We tested the MB6C in the VC (Table 4). In the patients < 50 years of age (n = 269), 187 (69.5%) met the MB6C, among whom 3 (1.6%) had HRV. In patients age ≥ 50, 123 (36.6%) met the MB6C and 4 (3.3%) among them had HRV. Totally, MB6C could help avoid EGD in 310 (51.2%) patients with 7 (2.3%) cases of HRV that were missed.
Table 4
Performance of the MB6C and AB6C in the VC
|
VC
(n = 605)
|
Patients age < 50
(n = 269)
|
Patients age ≥ 50
(n = 336)
|
|
EGD
spared,
n, (%)
|
HRV missed, n, (%)
|
EGD spared,
n, (%)
|
HRV missed, n, (%)
|
EGD
spared,
n, (%)
|
HRV missed, n, (%)
|
MB6C
|
310 (51.2)
|
7 (2.3)
|
187 (69.5)
|
3 (1.6)
|
123 (36.6)
|
4 (3.3)
|
AB6C
|
297 (49.1)
|
8 (2.7)
|
174 (64.7)
|
2 (1.1)
|
123 (36.6)
|
6 (4.9)
|
Abbreviations: MB6C, Modified Baveno VI criteria; AB6C, Alternative Baveno VI criteria; HRV: High-risk varices; VC, Validating cohort; EGD,Esophagogastroduodenoscopy;
Construct the alternative criteria based on age stratification
The SA was selected to replace the LS and build the AB6C. In patients < 50 years of age, thecriteriaof PLT > 100×109/L and SA < 55 cm2 maximized the number of potential EGDs avoided while keeping the risk of HRV missed below 5%. In patients aged ≥ 50, AB6C was the combination of PLT > 125×109/L and SA < 44 cm2 (Table 3).
The performance of AB6C in the VC is shown in Table 4. 174 (67.4%) cases met the AB6C for the subgroup age < 50 (n = 269), among them 2 (1.1%) had HRV. The alternative criteria for the subgroup age ≥ 50 was met by 123 (36.6%)cases, among them 6 (4.9%) had HRV. Overall, the AB6C could avoid 297 (49.1%) EGDs with 8 (2.7%) cases of HRV that were missed.
Validation of the B6C, EB6C, MB6C, and AB6C, and a comparison of the different criteria
In the entire cohort (n = 1188), 377 (31.7%) patients met the B6C. Of these 377 patients, 245 (64.9%) had no varices, 126 (33.4%) were low-risk varices (LRV) (diameter ≤ 5mm), and only 6 (1.6%) had HRV. Among the 651 (54.8%) patients that met the EB6C, 402 (61.8%) did not have varices, 225 (34.6%) had LRV, and 24 (3.7%) had HRV (Supporting Table S2).
Next, we evaluated the performance of these criteria in the entire cohort and age-related subgroups (Supporting Table S2 and S3).In patients < 50 years of age (n = 516), the B6C, EB6C, and MB6C could safely help avoid 36.8%, 59.1%, and 69.4%EGDs, respectively. In patients ≥ 50 years of age (n = 672), B6C could safely avoid 27.8% of EGDs with the risk of missing HRV being < 5%. Although the EB6C could avoid more EGDs (51.5%), 6.4% of HRV were missed. The MB6C could not only avoid more EGDs (39.9%) than B6C (27.8%), but also keep a low risk of missing HRV (3.7%). Overall, in the entire cohort, the MB6C helped avoid 52.7% of potential EGDs, and only 2.2% HRV were missed
Besides, AB6C also had a good performance in the entire cohort. By using the corresponding AB6C in the patients < 50 (n = 516) and ≥ 50 years of age (n = 672), 64.3% and 37.5% of the patients, respectively, could avoid EGD with 1.5% and 4.8% HRV being missed. Overall, the AB6C could safely spare 49.2% EGDs with only2.9% HRV being missed(Fig. 3).