Non-Invasive Predictors of Esophageal Varices in Alcoholic Chronic Liver Disease

Background Alcohol is widely consumed socially accepted recreational beverage, that is toxic and affects directly or indirectly almost every organ. Spectrum of alcoholic liver disease ranges from fatty liver to cirrhosis. One of the complications of the later spectrum is portal hypertension, around 50% develops varices and bleeding depends on the size of the varices. Predicting varices without endoscopic is difcult but few non-invasive parameters are available. Materials and Methods It was a prospective cross-sectional study done in Nobel Medical College Teaching Hospital, Biratnagar, Nepal from September 2018 to August 2019. Approval was acquired from Institutional Review Committee. Patients with chronic ethanol ingestion and features suggestive of chronic liver disease clinically and investigation wise were enrolled in the study. History, physical examinations along with platelet count, prothrombin time was taken and ultrasonography abdomen and upper gastrointestinal endoscopy was done to see the splenic diameter, and varices. Results Esophageal varices were present in 53%. Mean platelet count with varices was 122566 ± 36024.8 3 /mm , splenic diameter was 133.1 ± 21.3 mm, prothrombintime (PT) time was 19.3 ± 5.0 sec and ratio 3 3 of platelet per spleen diameter was 930.2 ± 259.4 /mm /mm. Platelet count < 163500/mm has sensitivity and specicity 83.0% and 83.0% respectively. Ratio of platelet per splenic diameter ratio cutoff 1293.7 has 88.7% sensitivity and 85.1% specicity for predicting varices. Conclusion In chronic alcoholic liver disease patients low platelet count, increased splenicdiameter, low platelet per splenic diameter ratio are useful in predicting presence of esophageal varices.


Introduction
Alcohol is widely consumed recreational beverage that is toxic to humans. It's a socially accepted widely consumed drug that affects directly or indirectly almost every organ of the human body. Here Alcohol refers to primary alcohol Ethanol (ethyl alcohol). Liver plays a vital role in metabolism of alcohol, except 10% of ingested alcohol eliminated through lungs, kidneys and skin, all remainder has to get oxidized through three different pathways that occur in liver. Chronic alcohol ingestion causes accumulation of toxic metabolites in liver causing liver injury. Alcohol related liver disease encompasses a spectrum of injury, ranging from simple steatosis to frank cirrhosis, depending on various risk factors other than quantity. Spectrum of alcoholic liver disease included fatty liver, alcoholic hepatitis and cirrhosis. Fatty liver, transient and reversible, is seen in any individual consuming a large quantity of alcohol over a long period of time [1]. Only 15-20% of chronic heavy drinkers develop hepatitis or cirrhosis occurring concomitantly or in succession [2]. Amount of alcohol intake at risk for alcoholic hepatitis is not known, but the majorities have a history of heavy alcohol consumption more than 100 g/day for more than two decades [3]. Chronic liver disease is an advance disease process lasting more than six months that involves progressive destruction and regeneration of the liver parenchyma ultimately leading to brosis and cirrhosis. It comprises of a wide range of liver pathologies, which include inammation, liver cirrhosis, and hepatocellular carcinoma. Patients usually present with complications of chronic liver disease including portal hypertension, hepatic encephalopathy, hepatorenal syndrome, hepatopulmonary syndrome, portopulmonary hypertension, cirrhotic cardiomyopathy and malnutrition. Portal hypertension is an important complicating feature of decompensated cirrhosis, which is responsible for the development of ascites, spleenomegaly and gastro-esophageal varices. Portal hypertension is dened as a hepatic venous pressure gradient (HVPG) of ≥ 5 mmHg. Acute haemorrhage from ruptured gastroesophageal varices is a medical emergency that occurs when HVPG exceeds 12mmHg and above 20mmHg failure to control bleeding hence increased mortality [4]. Determining HVPG is an invasive procedure where a catheter passed under radiological guidance into the hepatic vein where free and wedged hepatic pressures are measured, the gradient of wedged and free pressure give HVPG. Moreover direct visualization of gastro-esophageal varices itself is an invasive procedure. Around 50% of patient with portal hypertension develops varices and bleeding varices depends on the size. Mortality from acute variceal bleeding is as high as 20% even with the recent improvement in diagnosis and management [5]. Moreover, it is the second most common cause of death in cirrhotic patients [6,7]. Predicting gastro-esophageal varices without intervention is a tedious task. In case of medical emergency of upper gastrointestinal bleeding, identifying etiology of bleeding to be esophageal varix without prior history and invasive procedure is cumbersome. Just less than 50% of cirrhotic patients do have varices at diagnosis. Management and prognosis of variceal bleeding is entirely different and also re-bleeding is common ~ 30%-50% [5].Few non-invasive measure like splenomegaly, increased portal vein diameters on ographically, decreased platelet count etc. can predict varices (esophageal) but not with precision. This study is conducted in order to see the relationship of these non-invasive measures with the occurrence of varices and its bleeding tendency. Identifying non-invasive measure of varices prediction and bleeding tendency can be advantageous in the country like ours as invasive modalities are limited to urban areas. In our country where most of the variceal bleedings are managed conservatively due to lack of facilities, this would add ease to complexity. Moreover, prophylactic use of non-selective beta-blockers can be done using non-invasive predictors in chronic liver disease patients.

Materials and Methods
This study is a prospective cross-sectional study done is Nobel Medical College and Teaching Hospital, Biratnagar from September 2018 to August 2019. This study was started after acquiring approval from the Institutional Review Committee of Nobel Medical College. Written consent was acquired after the patient or patient party was explained about the study, its advantages, procedures and disadvantages. Patients presented to Nobel Hospital within the time frame with history of chronic ethanol ingestion and features suggestive of chronic liver disease clinically and investigation wise were enrolled in the study. Patient falling under the spectrum of chronic hepatitis and cirrhosis were taken. In the study done by Bhattarai S [12], platelet count < 1,44,000/cumm as non-invasive  Table 1.whereas, main clinical and biochemical characteristics of the study population is shown in Table 2.      3 was taken to less than 163500/mm then sensitivity and specicity was83.0% and 83.0% respectively. Another ratio of platelet count per splenic diameter per prothrombin time was also 3 calculated. (Platelet/splenic diameter)/PT /mm / mm/seccutoff value of83 has specicity and sensitivity of 96.2% and 83.0% respectively in predicting the varices.

Discussion
Variceal hemorrhage is a devastating complication of chronic liver disease. The mortality of rst episode of acute variceal bleeding was 30%, which increased on subsequent bleeding episode before widespread use of current therapies, and only onethird of patients survived for one year [8,9]. UGI Endoscopy is regarded as the best screening modality for diagnosing varices and the presence of large varices, cherry red spots etc. on endoscopy are high risk signs associated with bleeding [10,11]. Endoscopic surveillance for varices is recommended for cirrhotic patient repeatedly. UGI endoscopy is an invasive modality that is not accepted by patient at their ease. Moreover availability of this invasive test in rural areas of developing country like ours where chronic alcoholism and chronic liver disease is a common condition is scarce. Many non-invasive predictors of oesophageal varices are studied in different study of which few reliably available modalities are studied in this study.
Of 100 patients studied in the study 76 (76%) were male, similar male predominance of 77%, 69.3% and 86.1% was observed in studies carried out by Bhattarai S et al [12] Mandal L et al [13] and Sharma SK et al [14] respectively. Median age of study population was 52 years which reported similar 54 years as in the study carried out by Bhattrai et al [12] whereas the age was higher in the other above studies 40 and 45 years respectively. Oesophageal varices were detected in 53% of the patients, in contrary higher number of varices were detected in other studies carried out by Bhattrai et al [12] and Mandal L et al [13] 70% and 75.6% respectively. The difference was mainly due to the study population, which not only included cirrhotic but chronic alcolohic liver disease as a whole.  [12]. Another study by Mandal et al [13] also had lower platelet count with varices111000± 2840/cumm whereas platelet count was 215000 ± 5500/cumm without varices, which was higher then above studies. Taking cutoff limit of platelet count <150000/cumm for presence of varices, the sensitivity and specicity was 71.7% and 83% respectively. In other studies, Shanker et al [15] reported platelet count of <120000/cumm to be 90% sensitive and 50% specic in predicting oesophagealvarices. Similarly, Thomopoulos et al [16] mentioned platelet count of <118000/cumm to be a good indicator for presence of varices with sensitivity of 95% and specicity of 73%. In Bhattrai et al [12] study sensitivity of 87.9 % and specicity of 41.7% for cutoff platelets count of <144000/cumm was observed. Though lower platelet count had higher sensitivity in predicting varices but lacked specicity.  [17] and Zimbwa et al [18]. But the cutoff value of the ratio was taken 909 in both the studies and the later study showed 100% specicity as well.
Another ratio calculated in this study was platelet count per splenic diameter per prothrombin time ratio whose cutoff value 83 (according to ROC curve) was 96.2% sensitive and 83.0% specic in predicting varices. AUC as per ROC curve was 0.949 with p value <0.001 shows the ratio is an excellent test to predict occurrence of varices. Variceal bleeding was seen higher in patients with low platelet count (mean 119843.8 ± 34062.7), high PT (mean 19.6 ± 5.6), low platelet per spleen diameter (mean 878.5 ± 253.6) and grade 3 varices (100%) with p value <0.001.

Conclusion
In chronic alcoholic liver disease patient low platelet count, increased spleen diameter, low platelet per spleen diameter ratio and low platelet per spleen diameter per PT ratio are useful in predicting presence of oesophageal varices. Amongst these both the ratios platelet per splenic diameter and platelet per splenic diameter per PT ratio are excellent tools to predict varices. Thus non-invasive predictors mentioned above can be greater tool in differentiating patient with or without varices where invasive modalities are not available.