Measurement of Splenic Stiffness as a Predictor of Oesophageal Varices in Patients with Liver Cirrhosis in Zagazig University Hospitals

laboratory evaluation, BMI, splenic and liver stiffness measurement, upper GIT endoscopy. Splenic stiffness measurement repeated for patients who had varices after 6 months of pharmaco medical treatment. Results: Splenic stiffness was found to be higher in cirrhotic group than control group, splenic stiffness measurement was found to be higher in patients who had varices than no varices in cirrhotic patients, cut off of SSM for the presence of varices ≥39.5 kpa had AUROC 0.999, sensitivity 97.7%, specificity 96.9%, PPV 97.8% and NPV while LSM showed cut off value for presence of varices >22.5 kpa had AUROC 0.641 sensitivity 84.44%, specificity 60.61%, PPV 74.5% and NPV 74.1%. PSR showed cut of ≤657.7 had AUROC 0.855 sensitivity 95.56%, specificity 78.79%, PPV 86% and NPV 92.9%. APRI showed cut off >2.7 had AUROC 0.657 sensitivity 57.78, specificity 93.94%, PPV 92.9% and NPV 62%. There was highly significant difference in median SS in patients with large varices versus small varices (49.6vs 71.58 kpa with p<.0001) . SSM is not a useful tool for follow up of varices after pharmacological treatment with non selective beta-blockers (p=0.014). Conclusion: Fibroscan is a sensitive and reliable method for detection of esophageal varices. Splenic stiffness showed the best performance on detection of eosophageal varices, when compared to other non invasive predictors, PSR came in the 2nd place. Splenic stiffness measurement can differentiate


INTRODUCTION
Portal hypertension (PH) is a frequent complication of cirrhosis, contributing to the development of ascites, esophageal varices (EV) and hepatic encephalopathy.
The best available methodology for the assessment of PH is measurement of the hepatic vein pressure gradient (HVPG).However, the performance of HVPG is limited to highly specialized centers and requires extensive experience and therefore is not used routinely [1].
Accordingly, the introduction of noninvasive methods able to predict the stage of PH (i.e., not clinically significant, significant, and severe) could help to identify patients who are subjected to measurement of HVPG and, ultimately, optimize the diagnostic management of cirrhotic patients.
Several studies had shown that measurement of liver stiffness (LS) by transient elastography (TE) may represent a rapid and noninvasive method for predicting the presence of clinically significant (ie, HVPG ≥10 mm Hg) or severe (ie, HVPG ≥12 mm Hg) PH [2].On the other hand, LS shows a poor correlation with HVPG values ≥12 mm Hg, because of the increased incidence of extrahepatic factors conditioning the progression of PH [3].
Consequently, it is not surprising that LS is not an adequate method for prediction of the presence and grade of EV (and none of the thus far proposed noninvasive methods can be considered equivalent to measurement of HVPG or endoscopy in terms of overall accuracy [4].
Splenomegaly plays an important role in the pathophysiology of PH by increasing splanchnic inflow [5].
However, although splenomegaly represents a common finding in patients with cirrhosis and PH, the relationship between spleen size and PH grading or EV degree is controversial [6].
The possibility of predicting the presence of EV by using clinical parameters related to splenomegaly was initially suggested by the use of the spleen diameter, assessed by ultrasonography (US), in the platelet count/spleen diameter ratio (Plt/Spl) [7].
Recently, a direct correlation between splenic stiffness (SS), assessed by magnetic resonance elastography, and HVPG has been reported in a large animal model of PH [8].Accordingly, the possibility of detecting the presence of EV by the measurement of SS by TE in cirrhotic patients has also been recently proposed Regardless, a precise characterization of the relationship between SS and PH with relative complications, particularly the presence of EV, is still lacking.
This study aimed to determine efficacy of splenic stiffness measurement as a non-invasive tool in predicting the presence of esophageal varices in patients of liver cirrhosis evaluate validity of fibro scan of spleen in follow up degree of esophageal varices in patients of liver cirrhosis, Measure the ability of splenic stiffness measurement to determine grade of esophageal varices.9-Measurement of SS: SS values were obtained using the FibroScan with the same probe used to perform LS after at least 6 hours of fasting and under US assistance.In the absence of guidelines for the measurement of SS by FibroScan, the same guidelines for the measurement of LS were applied (i.e., success rate, IQR, and IQR/M), with some adjustments due to individual spleen anatomic characteristics.

PATIENTS AND METHODS
In particular, with the patient in a supine position with maximal abduction of the left arm, the probe was positioned in an intercostal space where the spleen was correctly visualized by US.Measurement of SS at presentation and after 6 months of treatment for cirrhotic group.

RESULTS
This study showed no statistically significant difference between demographic data in cirrhotic patients and apparently healthy control as shown in table (1)., and the American Association for the study of liver disease (AASLD) had determined that every patient diagnosed with cirrhosis should be investigated for presence of EV regardless child class and the cause.The splenomegaly developing in the context of liver cirrhosis is commonly ascribed to blood congestion, but older studies demonstrated that it cannot be considered only as a consequence of increased portal pressure and augmented resistance to splenic vein outflow [16].Surprisingly, no relationship could be found between the spleen size and the degree of esophageal varices [17].Multiple studies demonstrated pooling of blood in the red pulp, intraparenchymal arterial aneurysms, and other multiple histopathologic changes, which evolve towards diffuse fibrosis of the spleen [18].So, in this study, it is only logical to presume that the increase in size should determine changes in the spleen's density as well, which is a physical parameter that may be quantified by elastography.
This study showed significant increase in liver and splenic stiffness values in cirrhotic patients as compared with controls which are consistent with Bureau et al.In this study among cirrhotic group 33 patients (42.3%) had no varices, 45 patients (57.7%) had varices.
This study revealed highly significant increase in portal vein diameter, splenic bipolar diameter, total and direct bilirubin, INR and prothrombin time in patients who had oesophageal varices (EV) between cirrhotic patients with and without varices respectively and these results was in accordance with Schepis et al. [21] who showed that high portal vein diameter serve as a predictor for presence of EV and with Sharma et al. [22] who concluded that increase splenic bipolar diameter in patients with EV.Also there is highly significant decrease in platelet count and albumin in patients who had EV compared to patients who had no varices in cirrhotic group and this is consistent with [23].
Non invasive methods of liver fibrosis detection as liver stiffness measurement (LSM), Splenic stiffness measurement ( SSM), Platelet count/spleen diameter ratio (PSR) and AST to platelet ratio index (APRI) and its relation to portal hypertension and so oesophageal varices prediction was studied by many authors as Saad et al.In this study there is a highly significant difference between patients with EV and those without regarding the spleen diameter, Platelet count/spleen diameter ratio (PSR) AST to platelet ratio index, (APRI), Liver stiffness measurement (LSM) and Splenic stiffness measurement( SSM).These results are consistent with Saad et al. [24].Also, SSM and LSM were evaluated by Calvaruso et al. [25].This study concluded a highly significant difference in mean SSM values between patients with EV and those without (64.5 versus 24.6 kPa respectively; P<0.001).
In this study SSM had a cut of ≥39.5 kpa for the presence of EV, with 97.7%,96.9%sensitivity and specificity respectively and PPV 97.8% , NPV 97%, and AUROC 0.999, while LSM had lower sensitivity and specificity 84.44%,60.61%respectively and low PPV and NPV 74.5%, 74.1%, respectively and AUROC 0.641.So SSM is more sensitive and specific than LSM in the prediction of EV, these results are in agreement with Mohsen et  In their study, LSM could predict EV with cut-off ≥25 with sensitivity 56% and specificity 97%, while SSM could predict EV with cut-off value ≥55 with sensitivity 71% and specificity 95%.According to our study SSM showed better performance than LSM, Also SSM was the most sensitive parameter when compared with APRI, PSR and LSM as regards EV detection, PSR came in the 2 nd place similarly.Giannini et al. [30] proposed PSR of ≤909, as an accurate non-invasive marker for the presence of EV.
The result of this study are the same results of Cherian et al. [31] and González-Ojeda et al.
[32] who found that PSR was significantly lower in patients with EV than in those without.Mangone et al. [33] concluded that PSR is not a useful parameter to avoid unnecessary upper endoscopy in cirrhotic patients.Using the ROC curves, they found that PSR <936.4 for the prediction of presence of EV showed sensitivity 64.5%, specificity 64.3%, PPV 50% and NPV 76.6% (accuracy 0.671).Chawla et al. [34] supported these data in their meta-analysis where they concluded that PSR cut-off level of 909 may not be adequate to completely replace upper GI endoscopy as a non-invasive screening tool for EV.
On the contrary, Abu El Makarem et al. [35] found that PSR had a better diagnostic performance.In their study, PSR in patients with EV was significantly lower than in those without.In an analysis of the receiver operating characteristic curves (ROCs), an optimal cutoff value of 939.7 for this ratio, gave sensitivity 100%, specificity 86.3%, PPV 95.6%, NPV100% and AUROC of 0.94, 96.6% accuracy.
Regarding APRI, our results agreed with Zambam de Mattos et al. [36], that APRI was not a good index for the prediction of EV, because its sensitivity, specificity and predictive values were insufficient.In their cross-sectional study, APRI with a cutoff point of 1.3 demonstrated a sensitivity 64.7%, specificity 72.7%, PPV 86.5% and NPV 43.2%.
Regarding large EV detection, in this study among 45 cirrhotic patients had varices, 20 of them had small varices, 21 of them had large varices and 4 patients had both fundal varix and small varices.There was highly significant difference in median SS in patients with large varices versus small varices (49.6vs 71.58 kpa with P<.00001) respectively.However there is no significant difference in median LSM 26.7 VS 33.4 kpa with p 0.929.Also there is no significant difference in median APRI in patients with large varices versus small varices (2.4 vs 3.1 with p 0.768), Also there is no statistically significant difference in median PSR between both groups.
The results of this study agreed with Sharma et al. [22] who concluded that SS measurement can differentiate between small and large varices (56 kPa vs. 49 kPa, P=0.001), also these results agreed with Hua et al. [37] reported who that LSM couldn't assess EV accurately with no significant difference in LSM value between patients with large EV and those having small EV (31 kPa versus 28.18 kPa).

On contrary to this study Mohsen et al. [26]
showed APRI was the best for detecting large varices median 1.38, followed by SS measurement median 72.1kpa for large varices.
As regards bleeding varices, 12 Patients among 45 patients presented with attack of haematemesis and melena, this study showed SSM had moderate performance with cut-off ≥66 with sensitivity and specificity 83.3% and 83.3 % respectively, PPV and NPV47.6%96.5% respectively, which is consistent with Sharma et al. [22] who showed that SS measurement useful to differentiate bleeding vs non bleeding with cut off value 58 kpa.
PSR showed the best performance regarding bleeding varices with cut off value ≤468.6 had AUROC 0.913 sensitivity and specificity 91.67% and 74.24% respectively, PPV 39.3 % and NPV 98%.which is consistent with Sharma et al. [38] showed that PSR useful to differentiate bleeding vs non bleeding with cut off value ≤ 777.
LSM showed cut off value >23.4 kpa had AUROC 0.777 sensitivity, specificity91.67% and 45.45% respectively, PPV 23.4% and NPV 96.8 %.while APRI showed cut off value >2.2 had AUROC 0.793 sensitivity 75 %, specificity 54.55 %, PPV 23.1 % and NPV 92.3%.So this study showed moderate performance of SS and superiority of PSR which may be explained by hemodynamic changes at time of attack may affect platelet count, also SS not measured at the same time.
In this study all patients with small oesophageal varices underwent pharmacological treatment and follow up of splenic stiffness after 6 months.There was no statistically significant difference between initial SSM measurement and SSM after 6 months with p value 0.004, This may be attributed to short period of follow up, to be evaluated by further studies.While patients with large EV and gastric varices who underwent endoscopic band ligation and endoscopic injection respectively there was statistically significant increase of SSM after 6 months with p value <0.001 which may be explained by closure of collateral channels in the form of oesophageal and gastric varices and this reflected as increased splenic congestion and fibrosis and so increased SSM.
From this study and its results we concluded that Spleen stiffness measurement by Fibroscan is a sensitive and reliable method for detection of esophageal varices.Splenic stiffness showed the best performance on detection of oesophageal varices, when compared to other non invasive predictors, PSR came in the 2nd place.Splenic stiffness measurement can differentiate small and large varices.Splenic stiffness measurement can not be used as a tool for follow up of patients with oesophageal varices, who under went either pharmacological or endoscopic treatment.
This possibility is truly intriguing because splenomegaly in cirrhosis is characterized by enlargement and hyper activation of the splenic lymphoid tissue, as well as increased angiogenesis and fibrogenesis, in addition to passive congestion due to PH [10].

Table ( 3
): Correlations between certain studied parameters and SSM values in the whole population Table (1): Comparison between demographic data in cirrhotic patients and apparently healthy control

Table ( 5 ) :
Comparison between SSM, LSM, PSR and APRI Score values in cirrhotic patients reading presence of varices

Table ( 6): Comparison
between SSM, LSM, PSR and APRI Score values in cirrhotic patients regarding small and large varices Varices

Table ( 7
): Area under the ROC curve of SSM as a predictor for detection of presence of Cirrhosis, and development of attack, and varices, in Cirrhotic patients

Table ( 8
): Validity of SSM as a predictor for detection of presence of Cirrhosis, and development of attack, varices, OV Grade 3-4 and OV Grade 1-2 in Cirrhotic patients

Table ( 9
): Comparison between Initial SSM values in cirrhotic patients with small oesophageal varices and large oesophageal & gastric varices and on follow up after 6 Months