|Year : 2020 | Volume
| Issue : 1 | Page : 28-34
Platelet count/splenic diameter ratio: a noninvasive method for diagnosis of esophageal varices in Egyptian cirrhotic patients
Walaa A Elsalakawy1, Ahmed G El-Rab2
1 Clinical Hematology and Bone Marrow Transplant Unit, Department of Internal Medicine, Faculty of Medicine, Ain Shams University, Cairo, Egypt
2 Endoscopy Units, Luxor International Hospital, Luxor, Egypt
|Date of Submission||27-Oct-2019|
|Date of Acceptance||19-Nov-2019|
|Date of Web Publication||10-Sep-2020|
Walaa A Elsalakawy
Department of Internal Medicine, Faculty of Medicine, Ain Shams University, Abbassia, 11241
Source of Support: None, Conflict of Interest: None
Aim/Objectives To assess the utility of of platelet count (PC)/ SD ratio as a noninvasive predictor of EV.
Background Bleeding esophageal varices (EV) remain a leading cause of death in cirrhotic Egyptian patients. Although current guidelines recommend screening for EV, repeated endoscopies cause a significant burden and cost to endoscopy units and may expose patients to unnecessary procedures.
Methods This prospective cross-sectional analytic study included 100 Egyptian patients with liver cirrhosis, where splenic diameter assessment by ultrasound and upper endoscopy for grading of EV were done.
Results We found that the mean platelet count (PL) in patients with varices is 103 300±25 520/ml, whereas in patients with no varices is 212 390±58 980/ml. The mean spleen diameter (SD) in patients with varices is 154.59±10.57 mm, whereas in patients with no varices is 140.14±8.74 mm. The mean PL/SD ratio in patients with varices is 668.22±169.10 whereas in patients with no varices is 1515.56±470.40. The three relations werehighly significant (P≤0.00). At a cutoff value of 909, PL/SD showed sensitivity of 82.81% and specificity of 91.67%, with diagnostic accuracy of 89.8%.
Conclusions In conclusion, the use of PC/SD ratio can be of help to the physicians as a noninvasive predictor of EV to restrict the use of endoscopic screening only to patients presenting a high probability of EV. This is especially useful in clinical settings where resources are limited and endoscopic facilities are not present in all areas.
Keywords: cirrhosis, hepatitis C infection, noninvasive, esophageal varices, platelet count to spleen diameter ratio
|How to cite this article:|
Elsalakawy WA, El-Rab AG. Platelet count/splenic diameter ratio: a noninvasive method for diagnosis of esophageal varices in Egyptian cirrhotic patients. Egypt J Haematol 2020;45:28-34
|How to cite this URL:|
Elsalakawy WA, El-Rab AG. Platelet count/splenic diameter ratio: a noninvasive method for diagnosis of esophageal varices in Egyptian cirrhotic patients. Egypt J Haematol [serial online] 2020 [cited 2020 Dec 3];45:28-34. Available from: http://www.ehj.eg.net/text.asp?2020/45/1/28/294785
| Introduction|| |
Hepatitis C virus (HCV) is a major health problem in Egypt, and liver cirrhosis in Egyptian patients is a common clinical presentation. Esophageal variceal (EV) bleeding remains the leading cause of acute mortality in patients with cirrhosis . EVs are present at diagnosis in ∼50% of cirrhotic patients, being more common in Child–Pugh class C patients .
Screening is mandatory for early detection of EVs before first attack of hematemesis. Endoscopy is the only validated method for diagnosis of EV .
Repeated endoscopies cause a significant burden and cost to endoscopy units and expose patients to unnecessary procedures, as up to 50% of patients may still not have developed EV 10 years after the initial diagnosis .
However, in recent years, several noninvasive methods for detecting EV have been evaluated. These include clinical and biochemical parameters, ultrasonographic findings , transient elastography , computed tomography scanning , and video capsule endoscopy .
Diagnosing EV by noninvasive methods could reduce the need of endoscopy. This is of major importance especially in patients with a low probability of having EV, thus avoiding unnecessary examinations. Among the noninvasive methods, the platelet count (PL)/spleen diameter (SD) ratio has shown promising performance characteristics.
| Aim|| |
To evaluate accuracy of using the PL/SD ratio in predicting presence and grades of EV in Egyptian cirrhotic patients in comparison with upper gastrointestinal endoscopy findings.
| Patients and methods|| |
This prospective cross-sectional analytic study included 100 Egyptian patients with liver cirrhosis. A written informed consent has been obtained from all the study participants along with the approval of the study by the local Ethics Committee Board. The study conformed to the stipulations of declaration of Helsinki. The study was conducted in accordance with the stipulations of the local ethical and scientific committees of Ain Shams University, and the procedures respected the ethical standards in Helsinki declaration of 1964.
All patients were diagnosed as having liver cirrhosis based on physical findings, laboratory investigations, ultrasonographic findings, or histopathological findings whenever available.
Exclusion criteria included the following:
- Patient with bilharziasis.
- Patients who had a previous attack of variceal bleeding.
- Previous endoscopic sclerotherapy or band ligation of EV.
- Previous surgery for portal hypertension or TIPS.
- Patients with hepatocellular carcinoma or portal vein thrombosis.
- Coinfection hepatitis B virus (HBV) and HCV.
- Patients with gastric fundal varix.
- Basic assessment: all patients were subjected to full history taking and clinical examination, and routine laboratory investigations [complete blood count, liver function test, viral markers (HCVAb, HbsAg, and HBcAb; IgM and IgG), and alpha-fetoprotein].
- Child–Pugh classification: evaluation of the severity of liver disease was done using the Child’s score.
Child–Pugh–Turcotte classification of the severity of cirrhosis .
Child–Pugh–Turcotte classification was as follows:
- Child A: 5–6 points.
- Child B: 7–9 points.
- Child C: 10–15 points.
Abdominal ultrasonography for spleen diameter assessment
Patients were prepared for abdominal ultrasound by fasting for 8 h, and then the maximum spleen bipolar diameter was measured. While the patient is in the right lateral decubitus position with the left arm raised away from the abdomen, the transducer was placed between the ribs at the level of the ninth intercostal space, and then the patient was asked to take a deep breath and hold it. The transducer was manipulated in the coronal plane or the coronal oblique plane until a suitable longitudinal view of the spleen is obtained, then the length of the spleen was measured between the superior and the inferior borders of the spleen. Gel was applied to the upper abdomen before scanning for better resolution.
Upper gastrointestinal endoscopy
Lastly, all patients underwent an upper gastrointestinal endoscopy using a videoscope. All endoscopies were performed by experienced endoscopists, and a grading classification of I–IV was used .
- Grade I: the vein is flush with the wall of the esophagus.
- Grade II: protrusion of the varix not further than half way to the center of the lumen.
- Grade III: protrusion more than half way to the center of the lumen.
- Grade IV: the varices are so large that they meet at midline.
Then, PL/SD ratio for all patients was calculated.
Statistical analysis was performed using SPSS, V17 (Corporate headquarters1 New Orchard Road, Armonk, New York, United States). Quantitative data were represented as mean and SD for parametric data and as median and range in nonparametric data. Qualitative data were represented as number and percentage.
Comparisons of qualitative variables were conducted between groups using the c2 and comparisons of quantitative variables were conducted between groups using the Mann–Whitney for nonparametric data and Student t test for parametric data. However, comparison between more than two groups with parametric distribution was done by using one-way analysis of variance and Kruskal–Wallis for nonparametric distributions.
In addition, correlations between quantitative variables within groups were performed using the Pearson correlation coefficient. P value less than 0.05 and less than 0.001 were set as statistically significant and highly significant, respectively.
| Results|| |
This prospective cross-sectional analytic study included 100 Egyptian patients with liver cirrhosis. The study included 66 (66%) males and 34 (34%) females. The age ranged from 34 to 65 years (mean±SD, 49.7±7.7 years). A total of 88 (88%) patients were infected with HCV, whereas the remainder (12%) were infected with HBV. Initial laboratory investigations are shown in [Table 1]. By Child–Pugh classification, Child A included 24 (24%) patients, Child B included 31 (31%) patients, and Child C included 45 (45%) patients.
|Table 1 Initial laboratory investigation among study groups, including alanine aminotransferase, aspartate aminotransferase, international normalized ratio, albumin, bilirubin, hemoglobin level, and white blood cell count|
Click here to view
Most patients [70% (70 cases)] have no previous attack of hepatic encephalopathy, whereas 30% (30 cases) had at least one attack of hepatic coma. Ascites were classified sonographically into mild, moderate, and marked ascites. Overall, 30 (30%) cases have no ascites, whereas 70 (70%) have ascites; 23 (23%) of them have mild ascites, 28 (28%) have moderate ascites, and 19 (19%) have marked ascites.
Regarding the presence of varices, 36% (36 cases) have no varices, whereas 64% (64 cases) have varices. Regarding the grading of varices, grades I, II, III, and IV were seen in 8, 23, 5, and 28 patients, respectively.
There is a tendency of high grade of varices in patients with history of encephalopathy compared with patients with no history of encephalopathy as 50% of patients with history of encephalopathy had EV grade IV, whereas 47.1% of patients without history of encephalopathy had no varices ([Table 2], [Figure 1]).
|Table 2 Comparison between patient with history of encephalopathy and those without regarding grades of varices|
Click here to view
|Figure 1 Receiver operating characteristics (ROC) curve diagnostic performance of PL/SD for discrimination of UGIE finding of varices. ROC curve was used to define the best cutoff value of PL/SD, which was 909, with sensitivity of 82.81%, specificity of 91.67%, positive predictive value of 94.6%, negative predictive value of 75%, with diagnostic accuracy of 89.8%. PL, platelet count; SD, spleen diameter; UGIE, upper gastrointestinal endoscopy.|
Click here to view
Moreover, grades of varices showed statistically significant difference between ascitic and nonascitic patients, as in nonascitic patients, 80% of patients had no varices, whereas in ascitic patients, 38.57% had grade IV varices and 31.43% of them had grade II varices, as shown in [Table 3].
|Table 3 Comparison between ascitic and nonascitic patients regarding grades of varices|
Click here to view
In parallel, grades of varices showed statistically significant difference between Child–Pugh classes, as in class A, 91.7% showed no varices, whereas in class B 41.9% showed grade II EV. In contrast, patients in class C showed grade IV in 57.8% of them. This is shown in [Table 3].
In our cohort of patients, PL ranged from 47 000 to 415 000/ml, with a mean of 142 570±66 400/ml. The mean PL in patients with varices was 103 300±25 520/ml, whereas in patients with no varices was 212 390±58 980/ml.
The mean SD in patients with varices was 154.59±10.57 mm, whereas in patients with no varices was 140.14±8.74 mm. The mean PL/SD ratio in patients with varices is 668.22±169.10, whereas in patients with no varices is 1515.56±470.40. There was a statistically significant difference between patients with varices and those with no varices regarding the mean of the three parameters (PL, SD, and PL/SD ratio), as shown in [Table 4].
|Table 4 Comparison of upper gastrointestinal endoscopy grades of varices in different Child–Pugh classes|
Click here to view
Furthermore, by multivariate regression analysis, we found a highly significant correlation between presences of varices (yes/no), PL/1000, and grading of varices (P≤0.001).
Via receiver operating characteristics curve, sensitivity and specificity results for PL/SD ratio in prediction of EV at a cutoff value of 909 were 82.81 and 91.67%, respectively, with a diagnostic accuracy of 89.9% ([Table 5] and [Table 6]).
|Table 5 Correlation between upper gastrointestinal endoscopy finding of varices (yes/no) and both platelet count and spleen diameter|
Click here to view
|Table 6 Diagnostic performance of platelet count/spleen diameter ratio in discrimination of upper gastrointestinal endoscopy finding of varices|
Click here to view
| Discussion|| |
Chronic liver diseases and cirrhosis are now being recognized as an important cause of morbidity and mortality worldwide. In Egypt, liver cirrhosis is a major medical, social, and financial problem especially after hepatitis C infection. Approximately 20% of population are positive for HCV with dominancy for genotype 4. Approximately more than half a million people are newly infected each year .
Acute variceal hemorrhage secondary to cirrhosis is to date the most important cause of mortality in cirrhosis. Upper gastrointestinal endoscopy is usually performed for detection of EV in cirrhotic patients to avoid life-threatening bleeding.
Repeated endoscopies cause a significant burden and cost to endoscopy units and expose patients to unnecessary procedures as up to 50% of patients may still not have developed EV 10 years after the initial diagnosis .
This prospective study was conducted on 100 Egyptian patients with liver cirrhosis to evaluate the possibility of using PL/SD ratio as a noninvasive predictor of EV in cirrhotic patients, for restriction of performance of screening endoscopy.
Among our cohort of cirrhotic patients, the percentage of HCV was higher than HBV (88 vs. 12%). This of course is related to endemicity of HCV in Egypt. Yet, many other reporters, who worked in different countries, reported different percentages, such as Giannini et al. , who performed a study on 266 patients in Italy, where they found the percentage of HCV infection was 53.7%, alcohol abuse 16.5%, whereas HBV and other causes were 29.8%.
In our study, 30% of patients were not ascetic, whereas 70% were ascitic. Moreover, most patients (70%) had no previous attack of hepatic coma, whereas 30% had an attack or more of coma.
These two components share in the assessment of Child–Pugh classification which revealed 24% for Child A, 31% for Child B, and 45% for Child C, with the high percentage for the Child C. Most studies done in Egypt shared nearly same results, such as the study conducted by Abu El Makarem et al.  on 175 Egyptian patients with liver cirrhosis caused by HCV, and they found that 26.3% of patients were Child A, 33.7% were Child B, and 40% of patients were Child C.
In our study, regarding the presence of EV, 36% of patients included did not have EV and 64% had EV. Mattos et al.  conducted a study in 2010 in which they found that 73.2% of their patients had varices. However, Adarsh et al.  conducted a study on 101 patients in which alcohol was the commonest etiology, and they found that 95% of his patients had varices. This may be because he did not exclude patients with history of upper gastrointestinal tract bleeding as 72% of patients had history of recent bleeding, whereas in our study, we excluded patients with history of gastrointestinal tract bleeding. Moreover, Barrera et al.  found that 85% of included patients had varices.
In our study, there is a tendency of high grade of varices in patients with history of encephalopathy compared with patients with no history of encephalopathy, as 50% of patients with history of encephalopathy had EV grade IV, whereas 47.1% of patients without history of encephalopathy had no varices.
Moreover, the grades of varices show statistically significant difference between ascitic and nonascitic patients (P<0.001), as in nonascitic patients, 80% of patients had no varices, whereas in ascitic patients, 38.57% had grade IV varices and 31.4% of them had grade II varices. This in agreement with Barrera et al.  who found that 43.8% of ascitic patients had high-risk varices (grades III and IV).
In our study, regarding grades of EV in different Child classes, patients with high Child score had increasing possibility to have EV than others with lower child classes. Moreover, grades of varices show statistically significant difference (P<0.001) between Child–Pugh classes, as 91.7% of patients with Child A had no varices, 41.9% of patients with Child B had grade II varices, and 57.8% of Child C had grade IV varices. Abu El Makarem et al.  in their study found that 72.7% of patients with Child A had no varices, 52% of patients with Child B had grade II varices, and 49.6% of Child C had grade IV varices. This is in accordance with us.
This is also in agreement with Madhotra et al.  who found a significant relation between the presence of varices and increased Child score. Thus, the more advanced the liver disease (according to Child classification), the more likely the presence of varices.
Our study shows a statistically significant difference between patients with varices and those with no varices regarding the mean of the three parameters (PL, SD, and PL/SD ratio), as the mean PL in patients with varices was 103.3±25.5/109/l, whereas in patients with no varices was 212.3±58.9/109/l. The mean SD in patients with varices was 154.6±10.6 mm, whereas in patients with no varices was 140.1±8.7 mm, and lastly, the mean PL/SD ratio in patients with varices was 668.2±169.1, whereas in patients with no varices was 1515.5±470.4, with P value less than 0.001 (highly significant).
This is in agreement with Abu El Makarem et al. , who conducted their study on 175 Egyptian patients with liver cirrhosis, and their results showed a statistically significant difference (P<0.001) in between patients with varices (mean, 119.8±38.7/109/l) and those with no varices (mean, 213±69./109/l) regarding the mean PL. The mean SD in patients with varices was 159.4±24.4 mm, whereas in patients with no varices was 140.5±20.7 mm, and the mean PL/SD ratio in patients with varices was 747.6±197.6, whereas in patients with no varices was 1588.8±744.9.
This also is in agreement with Giannini et al. ,,,,,,,, who conducted their study on 266 patients with liver cirrhosis. Their results showed a statistically significant difference (P<0.001) between patients with varices and those with no varices regarding PL. The mean of PL in patients with varices was 112.3±53.3/109/l, whereas in patients with no varices was 221±126.8/109/l; the mean of SD in patients with varices was 159.7±29.9 mm, whereas in patients with no varices was 132.6±21.9 mm; and the mean of PL/SD ratio in patients with varices was 731.1±365.1, whereas in patients with no varices was 1778.6±1228.3.
These results are also in agreement with Esmat et al.  who conducted their study on 100 patients with HCV-related liver cirrhosis, and their results show a statistically significant difference (P<0.001) between patients with varices and those with no varices. Regarding the mean of PL, it was 955.6±41.5/109/l in patients with varices, whereas in patients with no varices was 215±69.7/109/l; the mean SD in patients with varices was 157.4±19.7 mm, whereas in patients with no varices was 121.2±13.5 mm; and the mean of PL/SD ratio in patients with varices was 624.8±301.4, whereas in patients with no varices was 1838.4±707.2.
In our study, receiver operating characteristics curve reading shows that PL/SD ratio is an accurate screening test for diagnosis of esophageal varicose, as area under the curve was 89.8%; is a good specific screening test for diagnosis of EV, as specificity was 91.67% with cutoff value of 909; and lastly, is a good sensitive screening test for diagnosis of EV as sensitivity was 82.81%.In the current study, we found that PL/SD with cutoff value 909 correlated significantly with the presence and grading of EV with P value less than 0.001. This is in agreement with Giannini et al. ,,,,,,, who found that PL/SD correlated significantly with the presence and grading of EV using the same cutoff value, and the sensitivity was 100% and the specificity was 93%.
Moreover, this is in agreement with Esmat et al.  who found that PL and SD ratio correlated significantly with the presence and grades of EV, but they used a cut-off value of 1326.58, resulting in 96.34% sensitivity, 83.33% specificity, and 94% accuracy.
This is also in agreement with Abu El Makarem et al.  who found that PL and SD correlated significantly with the presence and grades of EV, but they used a cutoff value of 939.7, resulting in 100% sensitivity, 86.3%specificity, and 96.5% accuracy.
In parallel, González-Ojeda et al.  studied a total of 91 patients. They found that the PL/SD ratio to detect EV independent of the grade using a cutoff value of less than or equal to 884.3 had 84% sensitivity, 70% specificity, and positive and negative predictive values of 94 and 40%, respectively.
However, this is not in agreement with Barrera et al.  who found that platelet count (PC)/SD ratio under than 830.8 could predict high-risk EV only, not all varices, with 76.9% sensitivity and74.2% specificity. This may be owing to different population used in their study, such as cirrhosis caused by alcohol in 26.9% of patients, viral hepatitis 7.5%, autoimmune hepatitis 11.9%, primary biliary cirrhosis 14.9%, NASH 14.9%, and cause undetermined in 26.9% of patients. However, in our study, cirrhosis was caused by viral hepatitis in all patients, the fact that made our study sample more homogenous. Moreover, the different ethnic groups may result in different results in the two studies.
Schwarzenberger et al.  found that among their 137 American cirrhotic patients, PL/SD ratio, with a cut-off value similar to us (909), yielded a negative predictive value of only 73% and a positive predictive value of 74%. They concluded that the PL/SD ratio with a cut-off value of 909 may not be sufficiently accurate in predicting the presence of EV.
The obvious discrepancy in the results of all the mentioned studies reflects the variability in the studied ethnic groups, variable etiologies of cirrhosis, and variable inclusion criteria for patient selection (regarding their previous encephalopathy or bleeding varices history). Yet, Pl/SD, no doubt, deserves further larger studies as a promising noninvasive method for detecting EV.
| Conclusion|| |
In conclusion, the use of PC/SD ratio can help physicians as a promising noninvasive predictor of EV to restrict the use of endoscopic screening only to patients presenting a high probability of EV. This is especially useful in clinical settings where resources are limited and endoscopic facilities are not present in all areas.
Financial support and sponsorship
Conflicts of interest
There are no conflicts of interest.
| References|| |
Chawla S, Katz A, Attar BM, Gupta A, Sandhu DS, Agarwal R. Platelet count/spleen diameter ratio to predict the presence of esophageal varices in patients with cirrhosis: a systematic review. Eur J Gastroenterol Hepatol
Merli M, Nicolini G, Angeloni S, Rinaldi V, De Santis A, Merkel C et al.
Incidence and natural history of small esophageal varices in cirrhotic patients. J Hepatol
Rye K, Scott R, Mortimore G, Lawson A, Austin A, Freeman J. Towards non invasive detection of esophageal varices. Int J Hepatol
Berzigotti A, Gilabert R, Abraldes JG, Nicolau C, Bru C, Bosch J, García-Pagan JC. Non invasive prediction of clinically significant portal hypertension and esophageal varices in patients with compensated liver cirrhosis. Am J Gastroenterol
Giannini EG, Zaman A, Kreil A, Floreani A, Dulbecco P, Testa E et al.
Platelet count/spleen diameter ratio for the noninvasive diagnosis of esophageal varices: results of a multicenter, prospective, validation study. Am J Gastroenterol
Vizzutti F, Arena U, Romanelli RG, Rega L, Foschi M, Colagrande S et al.
Liver stiffness measurement predicts severe portal hypertension in patients with HCV-related cirrhosis. Hepatology
Perri RE, Chiorean MV, Fidler JL, Fletcher JG, Talwalkar JA, Stadheim L et al.
A prospective evaluation of computerized tomographic (CT) scanning as a screening modality for esophageal varices. Hepatology
De Franchis R. Non-invasive (and minimally invasive) diagnosis of esophageal varices. J Hepatol
Lucey MR, Brown KA, Everson GT, Fung JJ, Gish R, Keeffe EB et al.
Minimal criteria for placement of adults on the liver transplant waiting list: a report of a national conference organized by the American Society of Transplant Physicians and the American Association for the Study of Liver Diseases. Liver Transpl Surg
Westaby D, McDougall B, Saunders J. A study of risk factors in patients with cirrhosis and variceal bleeding. In: Westaby D, McDougall B, Williams R, eds. Variceal bleeding
. London: Pitman 1982. 21.
Lavanchy D, McMahon B. Worldwide prevalence and prevention of hepatitis C. In Liang TJ, Hoofnagle JH eds. Hepatitis C
. San Diego, CA: Academic Press 2014. 185–202
Giannini E, Botta F, Borro P, Risso D, Romagnoli P, Fasoli A et al.
Platelet count/spleen diameter ratio: proposal and validation of a non-invasive parameter to predict the presence of esophageal varices in patients with liver cirrhosis. Gut
Abu El Makarem MA, Shatat ME, Shaker Y, Abdel Aleem AA, El Sherif AM, Moaty MA et al.
Platelet count/bipolar spleen diameter ratio for the prediction of esophageal varices: the special Egyptian situation: noninvasive prediction of esophageal varices. Hepat Mon
Mattos AZ, Mattos AA, vianna FF, Musskopf MI, Pereira-Lima JC, Maciel AC. Platelet count/spleen diameter ration: analysis of its capacity as a predictor of the existence of esophageal varices. Arq Gastroenterol
Adarsh Rajendran P, Anuish P, Krishnadas D, George KC, Jayaprakash, Devi S. Platelet count to spleen size ratio: a simple non invasive tool to identify small varices from large varices in patients having cirrhosis and portal hypertension. J Clin Exp Hepatol
Barrera F, Riqueime A, Soza A, Contreras A, Barrios G, Padilla O et al.
Platelet count/bipolar spleen diameter ratio for non invasive prediction of high risk esophageal varices in cirrhotic patients. Ann Hepatol
Madhotra R, Mulcahy HE, Willner I, Reuben A. prediction of esophageal varices in patients with cirrhosis. J Clin Gastroenterol
Esmat S, Omarn D, Rashid L. Can we consider the right hepatic lobe size/albumin ratio a noninvasive predictor of oesophageal varices in hepatitis C virus-related liver cirrhotic Egyptian patients? Eur J Intern Med
González-Ojeda A, Cervantes-Guevara G, Chávez-Sánchez M, Dávalos-Cobián C, Ornelas-Cázares S, Macías-Amezcua MD. Platelet count/spleen diameter ratio to predict esophageal varices in Mexican patients with hepatic cirrhosis. World J Gastroenterol
Schwarzenberger E, Meyer T, Golla V, Sahdala NP, Min AD. Utilization of platelet count spleen diameter ratio in predicting the presence of esophageal varices in patients with cirrhosis. J Clin Gastroenterol
[Table 1], [Table 2], [Table 3], [Table 4], [Table 5], [Table 6]