The development of esophageal varices (EV) is the most common complication in liver cirrhosis. Depending on the clinical degree of liver cirrhosis, EVs are present in 30%-40% of patients with compensated cirrhosis, but they can also be present in up to 85% of decompensated cirrhotic patients [1].

Variceal bleeding is a life-threatening disease occurring in 5% of small EV patients and up to 15% of patients with large esophageal varices. The death risk per episode of bleeding is between 10% to 20% [2, 3]; the higher difference in the percentage of variceal bleeding between Egypt and western countries (which might reach 56.29% [4] in the elderly) could be explained by the fact that Egypt has the highest prevalence rate of hepatitis C viral infection (HCV) worldwide (15%–20%) in addition to hepatitis B (3.2%) and bilharzial infestation, which are considered major risk factors for the high incidence of chronic liver diseases with a sequel of variceal bleeding [5, 6]. Thus, EV screening in patients with liver cirrhosis is highly recommended in all consensus reports [7-9]. The standard screening procedure is endoscopy in patients without EV at 2-3 years and in patients with small EV at 1-2 years [10].

The non-invasive methods for EV prediction consist of serum markers, imaging, measurement of liver stiffness, and a combination of these methods. Serum markers mostly consist of alanine and aspartate aminotransferase (ALT and AST), alkaline phosphatase, gamma-glutamyltranspeptidase, platelet count, and clinical variables. Imaging markers mostly consist of portal vein diameter, spleen diameter, and PC/SD [11].

Transient elastography is utilized to assess hepatic stiffness as it is non-invasive, easily performed with immediate results, and has good reproducibility [12].

Insulin resistance is the real determinant of both nonalcoholic fatty liver disease (NAFLD) and diabetes and can facilitate the accumulation of triglycerides in the liver [13]. The relationship between insulin resistance (IR) and hepatic hemodynamics as more reliable predictors of portal hypertension has been studied in several studies [14]. Increased resistance to insulin is also associated with chronic disease of the liver [15].

Insulin resistance is characterized by the reduction in the glucose disposal rate caused by the concentration of insulin [16].

According to Erice et al. [5], secondary hyperinsulinemia (abnormally high levels of insulin in the body relative to the level of glucose.) may develop due to one of the two mechanisms. If a hepatic venous pressure gradient (10 mmHg) [clinically significant PH (CSPH)] is developed; either pancreatic hyper-secretory state, as suggested by moderately higher C-peptide levels in patients with IR, or reduced insulin clearance as a result of Porto-systemic PH leads to secondary hyperinsulinemia. However, it is not linked to any underlying hepatocellular or endothelial disease [17].

The “gold standard” for insulin-sensitivity assessment is hyperinsulinemic euglycaemic clamp (HEC(. However, it is time-consuming and expensive, leading to the need for a simplistic approach to insulin sensitivity quantification [18].

Matthews et al. initially developed a homeostasis model assessment (HOMA model) in 1985. It is a tool used to measure tolerance to insulin and beta-cell function from concentrations of fasting levels of both glucose and insulin (or C-peptide). The HOMA model has been a successful clinical and epidemiological method for insulin resistance assessment [19].

QUICKI is a reliable and accurate approach based on a mathematical translation of fasting blood glucose and plasma insulin concentrations produced empirically [20].

The linear association between QUICKI and glucose clamp is much greater. Many studies have found that QUICKI and glucose clamp values have excellent linear correlations [21-23].

Many studies have investigated the validity of insulin resistance for the detection of esophageal varices. The aim of the study was to assess the validity of insulin sensitivity and compare it to insulin resistance markers and other non-invasive markers for the detection of esophageal varices in post chronic hepatitis C virus cirrhotic patients.

This cross-sectional research was conducted on 95 patients who attended Tanta University Hospital's Department of Tropical Medicine.Before the beginning of the research, approval was taken from the Institutional Ethical Committee (approval code number:344447\ 2\ 21).

For all study participants, the objective of the analysis was made transparent, and every patient signed informed consent before getting enrolled.The study data were accessed by all the contributors, and the final text was checked and authorized.

The mean age of all the included participants was 52.65 ± 7.55 years. Eighty-four percent were males, and 15.7% were females. Only 8 patients of all the studied patients (10.53%) had diabetes. The whole cases mean Child-Pugh score was 8.30 ± 1.97. All of the patients had a mean body mass index (BMI) of 24.9± 2.2 kg/m2. In terms of liver function tests, there was no statistically significant difference between the investigated groups (P > 0.05) except for serum albumin (P = 0.003). Group III had lower albumin than group II (P = 0.009), and group III had lower albumin than group I (P = 0.009) (Table 1).

Regarding fasting and 2h postprandial blood glucose levels, there was no statistically significant difference among the three studied groups. The mean values of fasting blood glucose were 97.31±14.97, 100.76 ± 16.90, and 104.76± 24.05 in groups I, II, and III, respectively. Postprandial blood glucose mean values were 147.15±19.81, 163.56±62.52, and 173.6±40.92 in groups I, II, and III, respectively (Table 1).

3.3. Platelet Count/Spleen Diameter Ratio (PC/SD)

It showed a statistically significant difference among the three groups (P<0.001). The Post Hoc test showed a significant difference between group I and group II (P = 0.000) and between group I and group III (P =0.000); this means that the PC/SD ratio has a significant inversely proportional relationship with the grade of EV (Table 1).

In an attempt to find significant predictive factors of esophageal varices, Receiver Operator Characteristic (ROC) curve analysis was carried out (Table 3).

Regarding insulin resistance, a cut-off value of ≥ 3.4 could significantly predict EV (AUROC= 0.834) with 72% sensitivity, 80.0% specificity, 83.7% positive predictive value (PPV), and 57.6% negative predictive value (NPV). This denotes that HOMA-IR is a good diagnostic test for the presence of EV (Fig. 2).

For portal vein diameter, a cut-off value of ≥13.1 mm could predict EV (AUROC= 0.688) with 74.0% sensitivity, 61.5% specificity, 78.7% PPV, and 55.2% NPV (Table 3, Fig. 3).

Regarding spleen diameter, a cut-off value of ≥ 16.6 cm could predict EV (AUROC= 0.723) with 70% sensitivity, 65.4% specificity, 79.5% PPV, and 53.1% NPV (Table 3, Fig. 4).

Table 1. Patients’ characteristic.

	Group I= 26	Group II=25	Group III=25	ANOVA
Variable	Mean + SD	Mean + SD	Mean + SD	F	P	Post-hoc
Age	52.39 ± 6.85	53.92 ± 7.68	51.68 ± 8.22	0.569	0.568
BMI	24.55 ± 2.26	24.26 ± 1.92	25.90 ± 2.16	4.23	0.018*	P1 0.880 P2 0.067 P3 0.022*
Child Pugh score	8.42 ± 2.08	7.32 ± 1.78	9.16 ± 1.65	6.296	0.003*	P1 0.09 P2 0.34 P3 0.002*
RBCs (106/mm3)	3.73 ± 0.64	3.52 ± 0. 82	3.48 ± 0.57	1.005	0.371
Platelets (103/mm3)	121230.77± 34990.64	79800.08± 32055.74	93080.00 ± 25460.62	11.806	0.000*	P1 0.000* P2 0.003* P3 0.434
Albumin (g/dl)	2.98 ± 0.49	2.98 ± 0.43	2.59 ± 0.42	6.168	0.003*	P1 1.000* P2 0.009* P3 0.009*
International normalized ratio INR	1.59 ± 0.51	1.43 ± 0.21	1.66 ± 0.42	2.295	0.108
FBG(mg/dl)	97.31±14.97	100.76 ± 16.90	104.76± 24.05	0.982	0.379
2 hour post prandial blood glucose (mg/dl)	147.15±19.81	163.56±62.52	173.6±40.92	2.313	0.106
AST (U/L)	63± 29.72	60.28 ± 25.21	65.16± 18.27	0.241	0.787
ALT (U/L)	40.12± 18.09	34.36 ± 17.01	44.0± 15.56	2.051	0.136
HOMA IR	2.92 ± 0.55	3.82 ± 1.16	4.42 ± 1.48	11.306	0.000	P1 0.016* P2 0.000* P3 0.161
QUIKI	0.32 ± 0.008	0.316 ± 0.010	0.309 ± 0.015	13.975	0.000	P1 0.008* P2 0.000* P3 0.097
Fasting insulin (Mu/l)	12.25 ± 1.87	15.31 ± 2.96	16.96 ± 3.78	16.613	0.000	P1 0.001* P2 0.000* P3 0.132
Total Bilirubin (mg/dl)	2.45 ± 2.18	2.14 ± 1.96	2.90 ± 1.95	0.885	0.417

Group I = cirrhotic patients without esophageal varices
Group II = cirrhotic patients with small varices.
Group III = cirrhotic patients with large varices; P1--- no varices/ small varices; P2--- no varices/ large varices; P3 ---small varices/ larg varices.
* < 0.05

Fig. (1). ROC curve of QUIKI test as predictor of esophagial varices.

Fig. (2). ROC curve of HOMA-IR as predictor of esophagial varices

Table 2. Basal radiological data.

Variable	Group I= 26	Group II=25	Group III=25	ANOVA
	Mean ± SD	Mean ± SD	Mean ± SD	F	p	Post Hoc
PV diameter (mm)	13.23±2.01	13.80 ± 2.18	15.56 ±2.08	8.525	0.000*	P1 0.588 P2 0.000* P3 0.011*
Spleen diameter (cm)	15.89 ±2.70	16.96 ± 2.99	19.65 ±2.98	11.316	0.000*	P1 0.386 P2 0.000* p3 0.005*
Platelet count/spleen diameter	799.44 ± 266.99	488.04 ±187.10	460.24±138.66	17.195	0.000	P1 0.000* P2 0.000* P3 0.947

* Significant at P < 0.05

Table 3. ROC curve results of different diagnostic tests for the presence of esophageal varices.

ROC curve result	Insulin Resistance HOMAIR	PVD	Spleen Diameter
Area under the curve	0.834	0.688	0.723
Cut off value	≥ 3.4	≥13.1	≥ 16.6
Sensitivity	72%	74%	70%
Specificity	80%	61.5%	65.4%
Postive predictive value	83.7%	78.7%	79.5%
Negative predictive value	57.6%	55.2%	53.1%

Fig. (3). ROC curve of portal vein diameter as predictor of esophagial varices.

Fig. (4). ROC curve of spleen diameter as predictor of esophagial varices.

The initiation of EV is well known to be a critical turning point in the clinical progression of liver disease. As a result, the latest guideline recommendations stated that once cirrhosis is diagnosed, all patients should undergo endoscopic screening for varices and routine periodic assessment [26].

Non-invasive methods for predicting the existence of esophageal varices would give a tool for identifying individuals who might benefit from endoscopy, thus reducing unnecessary endoscopy. This would make patient care more cost-effective, especially in areas where endoscopy is unavailable and liver cirrhosis is common. If non-invasive tests employed for first-line selection are sufficiently accurate to distinguish endoscopic necessity, selective screening endoscopy would be more cost-effective than universal screening endoscopy [27].

To the best of our knowledge, this is the first study to evaluate the insulin sensitivity assessed by QUIKI in the prediction of EV. QUIKI showed a statistically significant difference among studied groups (p-value = 0.008 and 0.000), respectively, denoting its ability to predict the presence of EV.

In our study, there was a statistically significant difference among the three groups regarding fasting insulin and HOMA IR, as the patients with an advanced grade of EV had higher insulin resistance (higher HOMA IR and lower QUIKI). HOMA-IR is a good diagnostic test for the prediction of EV. A cut-off value of ≥ 3.4 of HOMA IR had 72% sensitivity and 80% specificity.

Our findings were consistent with those of Camma et al. who found that HOMA-IR of greater than 3.5, independent of diabetes, substantially predicts the occurrence of EV in patients [27].

Moreover, our results were in agreement with the findings of studies conducted by Eslam et al., Wasfy et al., and Abu El Makarem et al., who found that the cut-off value for the HOMA-IR score of greater than 3 was the optimal value for accurate prediction of EVs with different sensitivity and specificity [28-30].

The diverse ethnic groups of patients may cause disparities in the results of the studies. Patients with various etiological causes for cirrhosis, such as alcohol, HCV, HBV, and others, might contribute to this discrepancy, according to the study of Eslam et al. [28], which was conducted in two centers (in Spain and Egypt). This positive association of insulin resistance in the prediction of EVs can be attributed to insulin stimulation of endothelial lining malfunction, which is a significant early event in vascular diseases.

Insulin can change the endometrial nitric oxide and endothelial structure [30]. It can also promote the synthesis of tumor necrosis factor-alpha and connective tissue growth factor and activate hepatic stellate cells (HSC), which play an important role in the development of liver cirrhosis and PH [31].

In our study, a cut-off value of ≥13.1 mm of the portal vein diameter could predict EV (AUROC= 0.688) with 74.0% sensitivity, 61.5% specificity, 78.7% PPV, and 55.2% NPV.

A cut-off value of ≥13 mm of the portal vein was supported by multiple studies with high specificity and sensitivity [29-33].

On the other hand, Prihatini et al. and Sarwar et al. considered a portal vein diameter of 11.5 mm a predictive factor for esophageal varices in liver cirrhosis [34, 35].

In our study, a cutoff value of ≥ 16.6 cm of splenic size could predict EV (AUROC= 0.723) with 70% sensitivity, 65.4% specificity, 79.5% PPV and 53.1% NPV. According to Wasfy et al., a cut-off value of 15.65 could significantly predict EV.

In agreement with our results, other studies also postulated that a spleen size larger than 13 cm predicts the development of gastroesophageal varices in cirrhotic patients [33, 36].

Since platelet count alone may be misleading as it can not be solely attributed to portal hypertension, the platelet count/spleen diameter ratio is a parameter linking splenomegaly implicated in thrombocytopenia of cirrhosis with spleen size being inversely proportional to platelet count [37].

The platelet count/spleen diameter ratio was also used by Giannini et al. to predict esophageal varices. This ratio links thrombocytopenia to splenomegaly to introduce a variable that considers that thrombocytopenia is mainly due to hypersplenism secondary to portal hypertension [38, 39].

The platelet count/spleen diameter ratio (PC/SD) was shown to be statistically significant among the three groups, indicating the manifestation of EV in our study.

The role of PC/SD in the prediction of EV was also supported by Camma et al. in 2009, Esmat et al. in 2012, Abu El Makarem et al. in 2011, and Badawi et al. in 2020 [27, 30, 38, 40].

The study was a single-center study with a limited number of patients and a small sample size. More studies on a larger number of patients are needed.

Lower insulin sensitivity (assessed by QUIKI) and higher insulin resistance (assessed by HOMA IR) are good non-invasive predictors of EV. In addition, PV diameter, spleen diameter, and PC/SD are observed as predictors of esophageal varices.

All the authors participated in the study and approved the final version of the manuscript.