Sequential algorithms combining non-invasive markers and biopsy for the assessment of liver fibrosis in chronic hepatitis B

Table 1 Demographic, laboratory and histological characteristics of the 110 patients with chronic hepatitis B

Males n (%)	80 (72.7)
Age (mean yr ± SD)	42.6 ± 11.3
BMI (kg/m2)	24.2 ± 3.3
AST (mean IU/L ± SD)	73.4 ± 61.2
AST / ULN ratio (mean ± SD)	1.75 ± 1.47
ALT (mean IU/L ± SD)	144.5 ± 148.0
ALT / ULN ratio (mean ± SD)	3.14 ± 3.3
γGT (mean IU/L ± SD)	46.4 ± 49.1
γGT / ULN ratio (mean ± SD)	0.81 ± 0.84
Bilirubin (mean μmol/L ± SD)	13.9 ± 6.48
PLT (mean 109/L ± SD)	194.6 ± 56.9
Albumin (mean g/L ± SD)	42.7 ± 5
Cholesterol (mean mg/dL ± SD)	177.5 ± 32.9
INR (mean value ± SD)	1.12 ± 0.1
Haptoglobin (mean g/L ± SD)	1.05 ± 0.6
α2M (mean g/L ± SD)	2.67 ± 0.84
ApoA1 (mean g/L ± SD)	1.48 ± 6.48
Viral load (mean log10 cp/mL ± SD)	2.15 ± 1.18
HBeAg positive cases (%)	20 (18.2)
HDV co-infected cases (%)	8 (7.3)
Staging n (%)
F0	15 (13.6)
F1	20 (18.2)
F2	40 (36.4)
F3	13 (11.8)
F4	22 (20.0)

SD: standard deviation; ULN: upper limits of normal;PLT: platelets; INR: international normalised ratio; α₂M: alpha-2-macroglobulin; ApoA1: apolipoprotein A1.

Table 2 Performance of the non-invasive methods and of the algorithm A in detecting significant fibrosis (≥ F2 by METAVIR) in patients with CHB

	Fibrotest	Forns		APRI		GUCI	Hui’s model	Algorithm A
Classified cases (%)	100	63.3		66.2		100	100	100
Cut-off	F2	4.2	6.9	0.5	1.5	0.2	0.15	na
Sensitivity (%)	80.8	58.3	14.6	70.8	27.1	66.7	50	100
Specificity (%)	90	78.3	100	87	95.7	95.7	90.9	91.3
PPV (%)	95.5	90.6	100	94.1	97.9	97.9	96.3	96
NPV (%)	64.3	53.5	35.9	62.2	39.7	58.9	45.8	100
LR +	8.1	2.69	0.146	5.45	6.3	15.5	5.49	11.49
LR -	0.21	0.53	0.85	0.36	0.76	0.35	0.55	0
Accuracy (%)	83.3	64.8	42.3	76.1	49.3	76.1	62.2	97.2
AUC (95% CI)	0.85 (0.75-0.95)	0.63 (0.50-0.76)		0.72 (0.58-0.86)		0.81 (0.70-0.92)	0.71 (0.56-0.86)	0.96 (0.92-1)

APRI: aspartate aminotransferase to platelets ratio; GUCI: Goteborg University Cirrhosis Index; na: not available; PPV: positive predictive value; NPV: negative predictive value; LR+: positive likelihood ratio; LR-: negative likelihood ratio; AUC: area under the curve; CI: confidence interval.

Table 3 Performance of the non-invasive methods and of the algorithm B in detecting cirrhosis (F4 by METAVIR) in patients with CHB

	Fibrotest	APRI	AAR	GUCI	Algorithm B
Classified cases (%)	100	66.2	100	100	100
Cut-off	F4	2	1	1	na
Sensitivity (%)	55.6	42.9	7.1	21.4	92.9
Specificity (%)	96.3	85.4	94.7	91.2	96.5
PPV (%)	90	53.8	82.4	73.7	87.5
NPV (%)	87.1	91.1	81.4	83.8	98.3
LR +	15	2.94	1.34	2.43	26.5
LR -	0.46	0.67	0.98	0.86	0.07
Accuracy (%)	86.1	79.2	77.5	77.5	95.8
AUC (95% CI)	0.76 (0.67-0.85)	0.64 (0.53-0.75)	0.51 (0.39-0.62)	0.56 (0.47-0.65)	0.95 (0.90-1)

Table 4 Features of clinical interest of stepwise algorithms in chronic hepatitis B

	Algorithm A	Algorithm B
Saved biopsies (%)	48	81
APRI performed (%)	100	100
Fibrotest performed (%)	34	52
Under-diagnosed and unclassified (%)	0	0
Over-diagnosed (%)	3	3

Algorithm A: algorithm for significant fibrosis (F ≥ 2 by METAVIR); Algorithm B: algorithm for cirrhosis (F4 by METAVIR); APRI: aspartate aminotransferase to platelets ratio.