Development and validation of a noninvasive prediction model of autoimmune hepatitis in patients with liver diseases

Figures & data

Figure 1. Flow chart of included and excluded patients in the study.

Table 1. Demographic and clinical characteristics of patients.

Factor	AIH (N = 191)	Not AIH (N = 993)	P value
Sex			<0.001
Female	156 (81.68%)	540 (54.38%)
Male	35 (18.32%)	453 (45.62%)
Age (y)	53 (46 ∼ 62)	47 (36.5 ∼ 54)	<0.001
RBC (10^9/L)	4.06 (3.72 ∼ 4.31)	4.42 (3.99 ∼ 4.86)	<0.001
Hb (g/L)	125 (111 ∼ 135)	131 (120 ∼ 146)	<0.001
WBC (10^9/L)	5.25 (3.88 ∼ 8.15)	5.37 (4.19 ∼ 6.6)	0.352
PLT (10^9/L)	138 (101 ∼ 199)	174 (126 ∼ 225)	<0.001
TB (µmol/L)	28.1 (15.8 ∼ 70.0)	16.1 (11.1 ∼ 27.2)	<0.001
DB (µmol/L)	15.1 (6.5 ∼ 49.0)	5.8 (3.6 ∼ 10.8)	<0.001
ALB (g/L)	37.8 (33.4 ∼ 41.4)	42.5 (39.0 ∼ 45.8)	<0.001
GLO (g/L)	28.9 (25.4 ∼ 33.9)	26.9 (23.7 ∼ 31.1)	<0.001
ALT (U/L)	132.2 (55.3 ∼ 317.5)	72.7 (34.7 ∼ 145.9)	<0.001
AST (U/L)	123.6 (56.6 ∼ 270.7)	48.3 (27.35 ∼ 92.5)	<0.001
GGT (U/L)	121.6 (61 ∼ 209.0)	88.2 (37.85 ∼ 205.2)	0.005
ALP (IU/L)	127.6 (100.0 ∼ 179.0)	112.3 (80.6 ∼ 180.3)	0.01
LDH (IU/L)	202 (175 ∼ 258)	192 (163 ∼ 226)	<0.001
TBA (µmol/L)	32.1 (11.3 ∼ 133.5)	9.9 (4.7 ∼ 28.9)	<0.001
PT (s)	12.9 (11.9 ∼ 14.3)	11.8 (11.1 ∼ 12.6)	<0.001
INR	1.13 (1.04 ∼ 1.26)	1.03 (0.98 ∼ 1.11)	<0.001
FIB (g/L)	1.81 (1.53 ∼ 2.26)	2.33 (1.97 ∼ 2.80)	<0.001
IgA (g/L)	2.58 (1.97 ∼3.38)	2.36 (1.75 ∼ 3.13)	0.006
IgG (g/L)	16.3 (12.8 ∼ 19.6)	13.7 (10.8 ∼ 16.6)	<0.001
IgM (g/L)	1.34 (0.95 ∼ 2.07)	1.29 (0.88 ∼ 1.97)	0.318
ANA (+)	132 (69.11%)	512 (51.56%)	<0.001
SMA (+)	24 (12.57%)	78 (7.85%)	0.034
LKM (+)	0 (0%)	1 (0.10%)	1.00
LP (+)	0 (0%)	0 (0%)
AMA (+)	12 (6.28%)	158 (15.91%)	0.001
AMA M2 (+)	8 (4.19%)	139 (14.009%)	<0.001
HBsAg (+)	7 (3.66%)	109 (10.98%)	0.002

Values presented as Median (P25, P75) with Wilcoxon rank sum test, or N (%) with Pearson’s chi-square test and Fisher’s precision probability test.

Table 2. Histological features of patients.

Histological feature	AIH (N = 191)	Not AIH (N = 993)	p value
Interface Hepatitis	185 (96.86%)	358 (36.05%)	<0.001
Lymphocyte/plasma cell predominance	143 (74.87%)	226 (22.76%)	<0.001
Rosettes	97 (50.79%)	49 (4.93%)	<0.001
Emperipolesis	4 (2.09%)	0 (0%)	0.001
Bile duct changes	62 (32.46%)	311 (31.32%)	0.756
Steatosis	22 (11.52%)	384 (38.67%)	<0.001

Table 3. Predictors of autoimmune hepatitis diagnosis determined by univariate and multivariate logistic regression analysis.

	Univariate analysis		Multivariate analysis
Variables	OR (95% CI)	P value	Adjusted OR (95% CI)	P value
ALT/ULN
5–10	2.407 (1.573, 3.627)	<0.001	1.12 (0.63, 1.97)	0.703
>10	2.850 (1.766, 4.514)	<0.001	0.55 (0.24, 1.19)	0.134
IgG/ULN
1-1.5	1.746 (1.093, 2.883)	0.024	1.20 (0.67, 2.20)	0.539
1.5–2	3.479 (2.117, 5.881)	<0.001	2.08 (1.03, 4.23)	0.042
>2	5.308 (2.928, 9.745)	<0.001	2.48 (0.93, 6.54)	0.067
ALP/AST ratio
1.5–3	0.336 (0.229, 0.487)	<0.001	0.54 (0.33, 0.90)	0.017
>3	0.135 (0.084, 0.209)	<0.001	0.19 (0.11, 0.34)	<0.001
ANA (+)	2.093 (1.510, 2.932)	<0.001	1.36 (0.90, 2.06)	0.144
AMA (+)	0.352 (0.182, 0.620)	0.001	0.22 (0.10, 0.43)	<0.001
SMA (+)	1.686 (1.018, 2.703)	0.035	1.11 (0.61, 1.97)	0.735
HBsAg (+)	0.309 (0.129, 0.627)	0.003	0.31 (0.12, 0.69)	0.007
IgM	0.960 (0.845, 1.046)	0.468
IgA	1.184 (1.049, 1.338)	0.006	0.98 (0.82, 1.17)	0.818
FIB	0.534 (0.408, 0.689)	<0.001	0.88 (0.71, 1.03)	0.119
TBA	1.008 (1.006, 1.010)	<0.001	1.00 (1.00, 1.01)	0.385
LDH	1.002 (1.000, 1.003)	0.017	1.00 (1.00, 1.00)	0.176
INR	34.191 (12.735, 96.783)	<0.001	0.61 (0.03, 9.50)	0.737
PT	1.397 (1.281, 1.530)	<0.001	1.20 (0.94, 1.59)	0.184
GGT	1.000 (0.999, 1.001)	0.814
ALP	0.999 (0.998, 1.000)	0.221
AST	1.002 (1.002, 1.003)	<0.001	1.00 (1.00, 1.00)	0.113
GLO	1.000 (0.999, 1.001)	0.814
ALB	0.861 (0.835, 0.887)	<0.001	0.92 (0.88, 0.96)	<0.001
DB	1.007 (1.005, 1.010)	<0.001	1.01 (0.99, 1.02)	0.337
TB	1.005 (1.003, 1.006)	<0.001	0.99 (0.98, 1.01)	0.363
Age	1.047 (1.033, 1.061)	<0.001	1.03 (1.02, 1.05)	<0.001
Female	3.754 (2.579, 5.607)	<0.001	3.26 (2.06, 5.27)	<0.001

The bolded p-values represent statistically significant differences.

Figure 2. Nomogram model for predicting the presence of autoimmune hepatitis in patients with liver diseases.

Figure 3. The receiver operating characteristic curve estimating the accuracy of the autoimmune hepatitis predictive model. AUROC, area under the receiver operating characteristics curve.

Figure 4. The calibration curve of the nomogram model for predicting autoimmune hepatitis.

Table 4. Diagnostic parameters of the simplified criteria for diagnosis of autoimmune hepatitis.

Parameters	The simplified criteria
	≥6	≥7
Sensitivity	104/191 (0.545)	54/191 (0.283)
Specificity	800/993 (0.808)	891/993 (0.897)
Predictability	904/1184 (0.764)	945/1184 (0.798)
Positive predictive value	104/297 (0.350)	54/156 (0.346)
Negative predictive value	800/887 (0.902)	891/1028 (0.867)