Multimarker model for non-invasive assessment of variceal bleeding in chronic hepatitis C cirrhotic patients

Table 1 Demographic and clinical characteristics of the study groups, n (%)/mean ± SD

Parameter	Group I (n = 31)	Group II (n = 31)	Group III (n = 31)	P value1
Age (years)	57 ± 9.46	60 ± 6.59	56 ± 6.52	0.06
FIB-4 index	4.5 ± 1.2	4.3 ± 1.1	NA	0.48
APRI score	2.1 ± 0.7	1.9 ± 0.6	NA	0.25
Sex (male/female)	22/9	19/12	14/17	0.08
Child-Pugh score				< 0.001
Class A	7 (22.6)	9 (29.1)
Class B	18 (58.0)	17 (54.8)
Class C	6 (19.4)	5 (16.1)

¹Comparisons were performed using one-way analysis of variance for continuous variables and the χ² test for categorical variables.

FIB-4: Fibrosis-4; APRI: Aspartate aminotransferase to platelet ratio index; NA: Not available.

Table 2 Laboratory findings across study groups, mean ± SD

Parameter	Group I (n = 31)	Group II (n = 31)	Group III (n = 31)	P value (I vs II)1	P value (I vs III)1	P value (II vs III)1
WBCs (× 103/μL)	8.78 ± 5.75	8.55 ± 6.94	7.58 ± 2.24	0.89 (NS)	0.28 (NS)	0.46 (NS)
Hb (g/dL)	9.10 ± 1.34	9.83 ± 1.49	12.31 ± 1.86	0.04 (S)	< 0.01 (HS)	< 0.01 (HS)
PLT (× 103/μL)	97.71 ± 35.82	123.00 ± 51.61	267.97 ± 71.24	0.03 (S)	< 0.01 (HS)	< 0.01 (HS)
INR	1.47 ± 0.28	1.39 ± 0.37	1.04 ± 0.15	0.36 (NS)	< 0.01 (HS)	< 0.01 (HS)
Albumin (g/dL)	2.34 ± 0.53	2.62 ± 0.68	4.05 ± 0.58	0.08 (NS)	< 0.01 (HS)	< 0.01 (HS)
Bilirubin (mg/dL)	2.14 ± 2.00	2.44 ± 2.64	0.88 ± 1.15	0.62 (NS)	0.004 (HS)	0.004 (HS)
ALT (U/L)	46.2 ± 15.8	42.5 ± 13.9	30.1 ± 9.8	0.28 (NS)	< 0.01 (HS)	0.02 (S)
AST (U/L)	58.7 ± 18.4	55.1 ± 16.2	33.9 ± 10.7	0.37 (NS)	< 0.01 (HS)	0.01 (S)
Creatinine (mg/dL)	1.32 ± 0.41	0.94 ± 0.22	0.82 ± 0.18	< 0.01 (HS)	< 0.01 (HS)	0.03 (S)
MELD score	16.8 ± 4.5	12.1 ± 3.8		< 0.01 (HS)

¹Comparisons were performed using one-way analysis of variance followed by pairwise comparisons performed using Scheffe’s post-hoc test.

NS: Not significant; S: Significant: HS: Highly significant; WBCs: White blood cells; Hb: Hemoglobin; PLT: Platelet count; INR: International normalized ratio; ALT: Alanine aminotransferase; AST: Aspartate aminotransferase; MELD: Model for end-stage liver disease.

Table 3 Thyroid function tests across study groups, mean ± SD

Parameter	Group I (n = 31)	Group II (n = 31)	Group III (n = 31)	P value (I vs II)1	P value (I vs III)1	P value (II vs III)1
TSH (μIU/mL)	1.22 ± 1.47	1.41 ± 1.75	1.70 ± 1.40	0.65 (NS)	0.47 (NS)	0.72 (NS)
FT3 (pmol/L)	2.34 ± 0.53	2.62 ± 0.68	4.05 ± 0.58	0.83 (NS)	< 0.01 (HS)	< 0.01 (HS)
FT4 (pmol/L)	14.35 ± 2.87	13.22 ± 5.48	15.29 ± 2.28	0.31 (NS)	0.15 (NS)	0.06 (NS)

¹Comparisons were performed using one-way analysis of variance followed by pairwise comparisons performed using Scheffe’s post-hoc test.

NS: Not significant; HS: Highly significant; TSH: Thyroid-stimulating hormone; FT3: Free triiodothyronine; FT4: Free thyroxine.

Table 4 Radiological findings across study groups, mean ± SD

Parameter	Group I (n = 31)	Group II (n = 31)	Group III (n = 31)	P value (I vs II)1	P value (I vs III)1	P value (II vs III)1
Spleen size (cm)	16.95 ± 1.99	14.10 ± 1.87	12.47 ± 0.68	< 0.01 (HS)	< 0.01 (HS)	< 0.01 (HS)
PVD (mm)	13.33 ± 2.05	11.95 ± 2.01	10.45 ± 0.59	< 0.01 (HS)	< 0.01 (HS)	< 0.01 (HS)

¹Comparisons were performed using one-way analysis of variance followed by pairwise comparisons were performed using Scheffe’s post-hoc test.

HS: Highly significant; PVD: Portal vein diameter.

Table 5 Distribution and classification of varices in the cirrhotic cohort, n (%)

Variceal pattern	Group I (with bleeding) (n = 31)	Group II (without bleeding) (n = 31)	Total (n = 62)
Esophageal varices only	22 (71.0)	27 (87.1)	49 (79.0)
Fundal varices only (IGV1)	2 (6.5)	1 (3.2)	3 (4.8)
Both esophageal and fundal varices (GOV2)	7 (22.6)	3 (9.7)	10 (16.1)
Total patients with fundal varices	9 (29.0)	4 (12.9)	13 (21.0)

GOV2: Gastroesophageal varices type 2; IGV1: Isolated gastric varices type 1.

Table 6 Endoscopic findings in group I and group II, n (%)

Parameter1	Group I (n = 31)	Group II (n = 31)	P value2
OV grades			< 0.01 (HS)
Grade 0	0 (0)	5 (16.1)
Grade 1	5 (16.1)	10 (32.3)
Grade 2	12 (38.7)	13 (41.9)
Grade 3	12 (38.7)	3 (9.7)
Grade 4	2 (6.5)	0 (0)
FV			< 0.01 (HS)
Yes	9 (29)	4 (12.9)
No	22 (71)	27 (87.1)

¹Endoscopic characteristics of cirrhotic patients with and without variceal bleeding, including variceal grading distribution and prevalence of fundal varices.

²Comparisons were performed using the χ² test.

OV: Oesophageal varices; FV: Fundal varices; HS: Highly significant.

Table 7 Relationship between oesophageal varices grades and different parameters in group I and group II (including post hoc analysis)

Parameter	Group I			Group II
	F value	P value	Significant pairs (Scheffe test)	F value	P value	Significant pairs (Scheffe test)
TSH (μU/mL)	1.16	0.35	NS	1.76	0.18	NS
FT3 (pmol/L)	7.96	0.001	Grade 1 vs 3, grade 2 vs 3, grade 3 vs 4 (P < 0.05)	4.55	0.01	Grade 0 vs 3, grade 1 vs 3 (P < 0.05)
FT4 (pmol/L)	1.62	0.21	NS	1.32	0.29	NS
PLT (× 103/μL)	3.96	0.02	Grade 1 vs 3 (P < 0.05)	7.60	0.00	Grade 0 vs 3, grade 1 vs 3 (P < 0.05)
INR	3.52	0.03	Grade 2 vs 4 (P < 0.05)	6.88	0.00	Grade 1 vs 4, grade 2 vs 4 (P < 0.05)
Spleen size (cm)	3.31	0.04	Grade 1 vs 3 (P < 0.05)	6.58	0.00	Grade 0 vs 3, grade 1 vs 3 (P < 0.05)
PVD (mm)	3.64	0.02	Grade 1 vs 4 (P < 0.05)	8.40	0.00	Grade 0 vs 3, grade 1 vs 3 (P < 0.05)

Analysis of variance of clinical parameters across esophageal variceal grades within group I and group II. The P-value represents the overall significance of the difference across all grades. Significant differences between specific grade pairs were determined using the Scheffe post-hoc test. NS: Not significant; PLT: Platelet count; INR: International normalized ratio; TSH: Thyroid-stimulating hormone; FT3: Free triiodothyronine; FT4: Free thyroxine; PVD: Portal vein diameter.

Table 8 Relationship between fundal varices and different parameters in group I and group II, mean ± SD

Parameter	Group I1		P value2	Group II1		P value2
	FV present	FV absent		FV present	FV absent
TSH (μIU/mL)	1.62 ± 2.51	1.06 ± 0.77	0.35	2.65 ± 3.23	1.23 ± 1.44	0.45
FT3 (pmol/L)	2.40 ± 0.93	2.72 ± 0.91	0.39	3.01 ± 1.38	2.51 ± 0.91	0.34
FT4 (pmol/L)	13.82 ± 1.35	14.56 ± 3.29	0.39	17.45 ± 6.72	12.59 ± 5.13	0.09
PLT (× 103/μL)	100.56 ± 37.12	96.55 ± 36.06	0.78	100.75 ± 30.61	126.29 ± 53.64	0.36
INR	1.46 ± 0.12	1.48 ± 0.33	0.84	1.74 ± 0.62	1.35 ± 0.31	0.40
Spleen size (cm)	17.69 ± 1.79	16.65 ± 2.02	0.19	15.07 ± 2.33	13.96 ± 1.80	0.27
PVD (mm)	13.66 ± 1.67	13.19 ± 2.21	0.58	12.38 ± 1.60	11.89 ± 2.08	0.66

¹Comparison of clinical parameters between patients with and without fundal varices in group I and group II.

²The P-value represents statistical significance based on the t-test (or Mann-Whitney U test if data were non-parametric). A P-value < 0.05 was considered statistically significant.

FV: Fundal varices; PLT: Platelet count; INR: International normalized ratio; TSH: Thyroid-stimulating hormone; FT3: Free triiodothyronine; FT4: Free thyroxine; PVD: Portal vein diameter.

Table 9 Relationship between Child-Pugh scores and thyroid function parameters in group I and group II, mean ± SD

Child-Pugh class	Group I1			P value2	Significant pairs (Scheffe test)	Group II1			P value2	Significant pairs (Scheffe test)
	TSH	FT3	FT4			TSH	FT3	FT4
Class A	1.22 ± 1.47	2.99 ± 0.91	14.35 ± 2.87	< 0.01	FT3 A vs C (P < 0.01)	1.35 ± 1.52	3.25 ± 0.85	14.80 ± 2.45	< 0.01	FT3 A vs C (P < 0.01)
Class B	1.41 ± 1.75	2.55 ± 0.89	13.22 ± 5.48	< 0.01	FT3 B vs C (P < 0.01)	1.58 ± 1.68	2.75 ± 0.78	13.65 ± 5.10	< 0.01	FT3 B vs C (P < 0.01)
Class C	1.70 ± 1.40	1.75 ± 0.77	15.29 ± 2.28	< 0.01	FT3 A vs C, B vs C (P < 0.01)	1.85 ± 1.60	2.05 ± 0.70	15.50 ± 2.15	< 0.01	FT3 A vs C, B vs C (P < 0.01)

¹Relationship between Child-Pugh scores and thyroid function parameters in group I and group II.

²The P-value represents statistical significance based on one-way analysis of variance. Significant differences between specific groups were determined using the Scheffe test, with significant pairs listed in the last column.

TSH: Thyroid-stimulating hormone; FT3: Free triiodothyronine; FT4: Free thyroxine.

Table 10 Spearman's correlation between thyroid function and clinical parameters in group I and group II¹

Parameter	TSH1		FT31		FT41		TSH2		FT32		FT42
	ρ	P value	ρ	P value	ρ	P value	ρ	P value	ρ	P value	ρ	P value
OV grades	0.25	0.17	-0.61	< 0.001	-0.22	0.23	0.38	0.03	-0.51	< 0.001	-0.32	0.07
INR	-0.11	0.55	-0.35	0.04	-0.22	0.22	0.17	0.35	0.10	0.57	0.16	0.37
Child-Pugh score	0.26	0.15	-0.43	0.01	-0.16	0.37	0.33	0.06	-0.33	0.05	-0.31	0.08
PLT (× 103/μL)	-0.03	0.86	0.28	0.12	0.21	0.25	-0.23	0.20	0.40	0.02	0.19	0.29
Spleen size (cm)	-0.19	0.29	-0.05	0.77	-0.15	0.41	0.05	0.77	-0.15	0.40	0.04	0.81

¹Group I.

²Group II.

Data are expressed as Spearman’s rank correlation coefficients (ρ) and corresponding P-values. TSH: Thyroid-stimulating hormone; OV: Oesophageal varices; FT3: Free triiodothyronine; FT4: Free thyroxine; PLT: Platelet count; INR: International normalized ratio.

Table 11 Spearman’s correlation between varices and different parameters in group I and group II

Parameter	OV1		FV1		OV2		FV2
	ρ	P value	ρ	P value	ρ	P value	ρ	P value
TSH (μIU/mL)	0.25	0.17	0.15	0.42	0.38	0.03	-0.25	0.16
FT3 (pmol/L)	-0.61	< 0.001	0.14	0.44	-0.51	< 0.001	-0.15	0.41
FT4 (pmol/L)	-0.22	0.23	0.10	0.58	-0.32	0.07	-0.27	0.13
PLT (× 103/μL)	-0.52	< 0.001	0.10	0.58	-0.52	< 0.001	0.14	0.43
INR	0.35	0.04	0.02	0.90	0.36	0.04	-0.33	0.05
Spleen size (cm)	0.44	0.01	0.02	0.90	0.43	0.01	0.18	0.32
PVD (mm)	0.49	0.005	0.10	0.58	0.46	0.008	0.06	0.73
Child-Pugh score	0.54	0.001	0.16	0.38	0.38	0.03	0.06	0.73
MELD score	0.56	< 0.001	0.13	0.47	0.45	0.01	0.08	0.65

¹Group I.

²Group II.

Data are expressed as Spearman’s rank correlation coefficients (ρ) and corresponding P-values. TSH: Thyroid-stimulating hormone; OV: Oesophageal varices; FV: Fundal varices; FT3: Free triiodothyronine; FT4: Free thyroxine; PLT: Platelet count; INR: International normalized ratio; MELD: Model for end-stage liver disease; PVD: Portal vein diameter.

Table 12 Multivariate ordinal logistic regression for predictors of higher variceal grade in group I and group II

Variable	Group I-model 1 (Child-Pugh)		Group I-model 2 (MELD)		Group II-model 1 (Child-Pugh)		Group II-model 2 (MELD)
	aOR (95%CI)	P value	aOR (95%CI)	P value	aOR (95%CI)	P value	aOR (95%CI)	P value
FT3 (per 1 pmol/L decrease)	2.25 (1.52-3.33)	< 0.001	2.18 (1.48-3.21)	< 0.001	2.10 (1.45-3.05)	< 0.001	2.05 (1.42-2.98)	< 0.001
PLT (per 50 × 103/μL decrease)	1.75 (1.28-2.39)	< 0.001	1.70 (1.25-2.32)	0.001	1.65 (1.22-2.24)	0.001	1.62 (1.20-2.19)	0.002
INR (per 0.1 unit increase)	1.15 (1.03-1.28)	0.012	1.13 (1.02-1.26)	0.023	1.14 (1.02–1.27)	0.018	1.12 (1.01-1.25)	0.038
Child-Pugh score	1.45 (1.08-1.95)	0.013			1.40 (1.05-1.88)	0.023
MELD score			1.12 (1.02-1.23)	0.015			1.10 (1.01-1.20)	0.035
Age (years)	1.02 (0.98-1.06)	0.280	1.02 (0.98-1.06)	0.305	1.01 (0.97-1.05)	0.550	1.01 (0.97-1.05)	0.585
Sex (male)	1.20 (0.87-1.66)	0.260	1.18 (0.85-1.63)	0.322	1.15 (0.84–1.58)	0.380	1.14 (0.83-1.56)	0.415

Continuous predictors (free triiodothyronine, platelet count, international normalized ratio) were scaled to clinically relevant increments (a 1 pmol/L decrease, a 50 × 10³/μL decrease, and a 0.1 unit increase, respectively) to facilitate the interpretation of odds ratios. An adjusted odds ratio > 1 indicates increased odds of a higher variceal grade. Models are adjusted for the variables listed. aOR: Adjusted odds ratio; FT3: Free triiodothyronine; PLT: Platelet count; INR: International normalized ratio; MELD: Model for end-stage liver disease.

Table 13 Final multivariate ordinal logistic regression for predictors of higher variceal grade in the combined cirrhotic cohort (groups I and II)

Variable1	aOR (95%CI)	P value
FT3 (per 1 pmol/L decrease)	2.25 (1.68-3.01)	< 0.001
PLT (per 50 × 103/μL decrease)	1.72 (1.38-2.15)	< 0.001
INR (per 0.1 unit increase)	1.15 (1.06-1.25)	0.001
Bleeding status (yes)	2.85 (1.90-4.28)	< 0.001
Age (years)	1.01 (0.98-1.04)	0.480
Sex (male)	1.15 (0.91-1.45)	0.240

¹Continuous predictors (free triiodothyronine, platelet count, international normalized ratio) were scaled to clinically relevant increments (a 1 pmol/L decrease, a 50 × 10³/μL decrease, and a 0.1 unit increase, respectively) to facilitate interpretation.

Variance inflation factors for all variables were < 3, indicating no significant multicollinearity. A sensitivity analysis excluding patients with nonhepatic comorbidities (n = 4) yielded nearly identical results: Free triiodothyronine adjusted odds ratio (aOR) = 2.22 (1.61-3.06), platelet count aOR = 1.71 (1.35-2.16), international normalized ratio aOR = 1.15 (1.05-1.25), supporting model robustness. aOR: Adjusted odds ratio; FT3: Free triiodothyronine; PLT: Platelet count; INR: International normalized ratio.

Table 14 Diagnostic performance of free triiodothyronine levels for predicting high-grade varices in group I and group II

Parameter	Group I		Group II
	Value	95%CI	Value	95%CI
Cut-off FT3 (pmol/L)	2.5		2.6
Sensitivity (%)	85.7	72.8-93.9	84.0	70.0-93.0
Specificity (%)	78.3	64.2-88.7	77.0	62.0-88.0
PPV (%)	76.9	63.2-87.5	75.0	60.0-86.0
NPV (%)	86.7	73.2-94.9	85.0	70.0-94.0
AUC	0.88	0.79-0.93	0.86	0.77-0.92

Data are presented as diagnostic accuracy measures from receiver operating characteristic curve analysis. Cut-off free triiodothyronine (pmol/L): The threshold value provides the best balance between sensitivity and specificity. Sensitivity: The percentage of patients with high-grade varices correctly identified by free triiodothyronine levels. Specificity: The percentage of patients without high-grade varices correctly excluded by free triiodothyronine levels. AUC: Area under the curve; FT3: Free triiodothyronine; PPV: Positive predictive value; NPV: Negative predictive value.

Table 15 Diagnostic performance of variceal grade for predicting high-risk varices in group I and group II

Parameter	Group I		Group II
	Value	95%CI	Value	95%CI
Cut-off variceal grade	Grade II		Grade II
Sensitivity (%)	85.0	72.0-93.0	84.0	70.0-93.0
Specificity (%)	80.0	66.0-90.0	78.0	63.0-89.0
PPV (%)	82.0	69.0-91.0	80.0	65.0-90.0
NPV (%)	83.0	69.0-92.0	82.0	67.0-92.0
AUC	0.88	0.80-0.94	0.87	0.78-0.93

Data are presented as diagnostic accuracy measures from receiver operating characteristic curve analysis. Cut-off: The variceal grade threshold provides the best balance between sensitivity and specificity. Sensitivity: The ability to correctly identify patients with high-risk varices. Specificity: The ability to correctly exclude patients without high-risk varices. AUC: Area under the curve; PPV: Positive predictive value; NPV: Negative predictive value.

Table 16 Diagnostic performance of portal vein diameter for predicting high-grade varices in group I and group II

Parameter	Group I		Group II
	Value	95%CI	Value	95%CI
Cut-off PVD (mm)	13		12.5
Sensitivity (%)	80.0	66.0-90.0	78.0	63.0-89.0
Specificity (%)	75.0	60.0-86.0	74.0	59.0-86.0
PPV (%)	78.0	64.0-88.0	76.0	61.0-87.0
NPV (%)	77.0	62.0-88.0	76.0	61.0-87.0
AUC	0.85	0.77-0.91	0.84	0.76-0.90

Data are presented as diagnostic accuracy measures from receiver operating characteristic curve analysis. Cut-off: The portal vein diameter threshold provides the best balance between sensitivity and specificity. Sensitivity: The ability to correctly identify patients with high-grade varices. Specificity: The ability to correctly exclude patients without high-grade varices. AUC: Area under the curve; PPV: Positive predictive value; NPV: Negative predictive value; PVD: Portal vein diameter.

Table 17 Diagnostic performance of the variceal risk score for predicting high-grade varices

Parameter2	Apparent performance (95%CI)1	Bootstrap-corrected (1000 samples) (95%CI)1
AUC	0.94 (0.89-0.97)	0.92 (0.86-0.96)
Cut-off value	> 0.45
Sensitivity (%)	91.4 (81.5-96.6)	89.8 (79.5-95.2)
Specificity (%)	88.9 (78.4-95.4)	87.2 (76.8-93.8)
PPV (%)	87.5	85.9
NPV (%)	92.3	90.7

¹The “Apparent performance” column shows the model’s diagnostic accuracy on the original dataset. The “Bootstrap-corrected” column shows performance after internal validation using 1000 bootstrap samples, which corrects for overoptimism and estimates the model’s expected performance on new data. Area under the curve, sensitivity, and specificity are presented with 95% confidence intervals.

²The variceal risk score (VRS) was derived from a binary logistic regression model incorporating free triiodothyronine (FT3), platelet count (PLT), international normalized ratio (INR), and portal vein diameter (PVD): VRS = (-1.15 × FT3) + (-0.018 × PLT) + (1.92 × INR) + (0.28 × PVD). The cutoff (> 0.45) was selected using Youden's index. Bootstrap validation served as an internal sensitivity analysis, confirming the score's stability against sampling variability.

AUC: Area under the curve; PPV: Positive predictive value; NPV: Negative predictive value.