Non-invasive Methods For Diagnosing Of Esophagogastric Varices And Predicting The Occurrence Of Esophagogastric Variceal Bleeding In Patients With Cirrhosis

Part I Ratio of platelet count/spleen diameter predicted the presence of esophagogastric varices in patients with schistosomiasis liver fibrosisAim:To examine the platelet count/spleen diameter(PC/SD) ratio in predicting the presence of esophagogastric varices(EV) in patients with schistosomiasis liver fibrosis.Materials and methods:Patients who were diagnosed with schistosomiasis fibrosis at the Gastroenterologic Clinic of the Changshu Affiliated Hospital of Soochow University between May 2012 and March 2014 were enrolled in this study. Criteria for exclusion from the study were:(1) patients who declined the evaluation schedule;(2) patients with evidence of hepatitis B/C virus infection, hematologic disease, myelodysplastic syndrome, primary hypersplenism, chronic ethanol intake, or hepatic malignancy;(3)patients who frequently took medicines such as proton pump inhibitors, propranolol,or antiplatelet agents;(4) patients who received procedures that may influence outcomes(e.g., transjugular intrahepatic portosystemic shunts, sclerotherapy, or band ligation endoscopy);(5) patients with serious complications such as active bleeding, psychiatric disorders, or renal failure. All patients were diagnosed by endoscopy. Demographic, laboratory, and Doppler ultrasound parameters were collected and analyzed. Binary logistic regression analysis was performed to identify independent risk factors associated with EV occurrence. Receiver operating curves were generated to obtain the platelet count/spleen diameter ratio cutoff values for the optimal sensitivity and specificity with regard to EV.Statistical analysis was performed using SPSS 20.0(IBM Corp., Armonk, NY, United States) and Med Calc 12.7.0.0 software(Med Calc Software bvba, Ostend, Belgium). Pearson’s 2 tests were used to analyze categorical data(expressed as percentages),whereas independent samples t tests were used to analyze continuous data(expressed as the mean ± standard deviation) after evaluation with a normality test.Binary logistic regression analysis was performed to identify independent risk factors associated with EV occurrence. Receiver operating curves(ROC) were generated to obtain the cutoff values for the optimal sensitivity and specificity of the variables with regard to EV. Negative- and positive-predictive values with a confidence interval(CI) of 95% were also obtained. Two-sided P values < 0.05 were considered significant.Results : A total of 236(102 male, 134 female) consecutive patients with schistosomiasis liver fibrosis with an average age of 61.4 ± 10.2 years were enrolled in this trial. Of these,95 patients had endoscopic evidence of EV(EV + group), while the remaining 141 patients did not(EV- group), resulting in an EV incidence of 40.3%. Patients with EV were significantly older, had higher model of end liver disease(MELD) scores, portal vein diameter(PVD), and splenic vein diameters(SVD),and lower PC/SD ratios compared to the EV- group(all P < 0.05).A binary logistic regression analysis identified PVD and PC/SD as independently associated with EV occurrence. However, the areas under the ROC curves showed that the PC/SD ratio had a significantly higher diagnostic accuracy for EV as compared to PVD(0.891 [95% CI: 0.844–0.928] vs 0.764 [95% CI: 0.705–0.817]; P <0.01).Using the ROC curves, a PC/SD ratio lower than 1004 had optimal sensitivity and specificity for the judgment of EV(85.3% [95% CI: 76.5–91.7%] and 83.0% [95%CI: 75.7–88.8%], respectively),a 77.1%(95% CI: 67.9–84.8%) positive-predictive value and a 89.3%(95% CI: 82.7–94.0%) negative-predictive value. Using a cutoff of 1004,it was determined that 117/141(83.0%) patients without EV could avoid undergoing unnecessary endoscopy, whereas 14/95(14.7%) patients with EV would be misdiagnosed.Conclusion:As a noninvasive approach with a negative predictive value close to 90%, this approach will still aid gastroenterologists in evaluating the activity of schistosomiasis liver fibrosis. Owing to factors such as underdeveloped economy or low education level,schistosomiasis liver fibrosis patients in schistosomiasis epidemic areas, e.g., tropical and subtropical areas, commonly refuse to receive screening endoscopy. This findings indicates the great potential value of PC/SD ratio as a noninvasive,low-cost diagnostic factor of EV for these patients.Part II Modificated Model For End-Stage Liver Disease Systems and US-doppler For Predicting The Occurrence Of Esophagogastric Variceal Bleeding In Patients With CirrhosisAims:Screening endoscopy for assessing the risk of esophagogastric variceal bleeding(EVB) can be invasive,uncomfortable and expensive(to people in developing countries) and this can limit the frequency of examination.This study aimed to find a safe, noninvasive and cost-effective evaluation system to predict the occurrence of cirrhotic complications of variceal haemorrhage.Materials and methods:This study was approved by the Ethical Committee of the Su Zhou University affiliate Chang Shu Hospital, and all patients gave informed consent before enrollment.Patients who declined the schedule of evaluation, who harbored hepatocellular carcinomaor, who frequently taken medicines such as proton pump inhibitor, propranolol or antivirals were excluded from this trial. Patients who had a splenectomy,Transjugular intrahepatic portosystemic shunt(TIPS),or undergone endoscopic treatments such as sclerotherapy and band ligation before or after enrollment were also excluded from this trial.All patients enrolled were confirmed the presence of esophageal varix at entry to study with endoscopy.A total of 486 consecutive decompensated cirrhotic patients[238 males,248 females,mean(standard deviation) 63.1(11.2) years] were included in this study between May 2012 and March 2014 at the gastroenterologic clinics of the Su Zhou University affiliate Chang Shu Hospital.The diagnosis of decompensated cirrhosis was based on the combination of physical, laboratory, radiologic examination results. Variceal haemorrhage was confirmed with the endoscopic findings of esophageal varices with stigmata for recent bleeding(such as a fibrin plug on varix, adhered blood clot, etc.) or active bleeding(such as blood oozing or spurting).Every 3 months, patients enrolled in this trial were regularly subjected to Doppler US and MELD evaluation and the database was updated according to the new records.The end point of follow-up was the occurrence of variceal haemorrhage.Those patients who death for other reasons or lost follow up were excluded from this trial.MELD scores were calculated as follows, according to the United Network of Organ Sharing(UNOS) database:MELD score=9.57×loge(creatinine mg/d L)+3.78×loge(bilirubin mg/d L)+11.2×loge(INR)+6.43(constant for liver disease etiology). Minimal values are set to 1.0 for calculation purposes.The maximal serum creatinine level considered within the MELD score equation is 4.0 mg/d L.△MELD= the change in the MELD score over a period of 3 months.Statistical analyses were performed by using SPSS software(SPSS 20.0 for Windows; SPSS, Chicago, III) and Medcalc 12.7.0.0. software. A χ2 test or Fisher’s exact test(two tailed) was used for categorical data, and an independent samples T-test was used for continuous data. A normality test was performed for continuous data to assure the T-test was the most appropriate to use.Patient characteristics were expressed as mean values±standard deviations and as percentages as appropriate. Multivariate logistic regression with the Wald test for MELD score, △MELD and US parameters were performed to find the independent factors predictive of the EVB. Variables in the univariate analysis with p<0.20 were entered into multivariate logistic regression analysis to search for independent predictive factors.Receiver operating characteristic(ROC) curves were constructed for the independent factors. The area under the ROC curve(AUC) was determined and for the best discriminating probability threshold, sensitivity, specificity and predictive values were calculated and compared for diagnostic accuracy. All statistical tests were two-tailed, and results were considered statistically significant at P<0.05.Results:During follow-up, 70 of these patients were excluded from this trial(43 patients declined to have Doppler US or MELD evaluation during the first 3 months follow up,12 patients taken forbidden medicines frequently such as proton pump inhibitor, propranolol or antivirals,8 patients had a splenectomy and 7 patients undergone endoscopic treatments after enrolment).Finally,416 patients fulfilled a period of 31.6 months(range 12 to 47 months) follow-up and the latest follow-up was finished in March 2012.51 patients(27 males and 24 females) experienced variceal hemorrhage and 365 patients(201 males and 164 females) did not,the incidence of EVB was 12.3%。 We divided patients into two groups according to the occurrence of EVB.EVB(+) group comprised the 51 patients with EVB during the follow-up period while EVB(-) group comprised the others. Compared with EVB(-) group,EVB(+) group had a significantly Greater age(67.5±13.4 vs 62.3 ± 11.8,P=0.0085), a significantly higher MELD(26.5 ± 9.8 vs 18.9 ± 10.3, P=0.0000)and △MELD score(1.89±1.23 vs 0.66±0.47,P=0.0001), a significantly greater diameter of the SV(1.23±0.45 cm vs 1.01±0.23 cm, P=0.0006)and gastric coronary vein(GCV)(0.73±0.33 cm vs 0.61±0.21 cm, P=0.0115), a significantly blood flow of the spleen vein(SV)(1124 ± 412ml/min vs 896 ± 331ml/min,P=0.0002), a significantly higher proportion of patients with hepatofugal blood flow in GCV(49/51 vs 73/365,P=0.0000), a significantly higher hepatofugal flow velocity of gastric coronary vein(GCVV)(-21.1 ± 8.1cm/sec vs 14.1 ± 6.9cm/sec, P= 0.0002) and a significantly lower velocity of portal vein(PV)( 16.9 ± 10.1cm/sec vs 19.9 ±8.4cm/sec,P=0.0428). Factors in the univariate analysis with P<0.2 were entered into multivariate logistic regression analysis to search for independent factors predictive of the occurrence of EVB in 3 months. △ MELD(OR=6.195,95%CI: 1.193~22.624), GCVV(OR=7.129,95%CI: 2.725~24.256) and GCV blood flow direction(GCVBFD)(OR=23.306,95%CI: 2.376~337.186) were independently associated with the occurrence of EVB(Table 2).The logistic regression equation is:Logit(P)=1.667*△MELD+2.096*GCVV-3.245*GCVBFD-1.697.When GCV blood flow direction(GCVBFD) is hepatopetal,then GCVBFD =1 and GCVV is negative; When blood flow direction is hepatofugue,then GCVBFD =-1 and GCVV is positive.As the logistic regression model indicated, the predicted probability of the occurrence of EVB in patients with cirrhosis was a function of increased △MELD, GCVV and decreased hepatopetal blood flow in GCV. To strengthen the diagnostic accuracy, we proposed an index,the MMELD, that combined three independent factors,obtained with the following formula: MMELD= Logit( P) =1.667* △MELD+2.096*GCVV-3.245*GCVBFD-1.697.Then we compared the MMELD with △MELD, GCVV, GCVBFD alone for diagnostic accuracy according to ROC analysis. AUCs for predicting the occurrence of EVB were significantly higher for MMELD than for △MELD(0.858[95%CI: 0.774 to 0.920] vs 0.734[95%CI: 0.636 to 0.817], P=0.0403), GCVV(0.858[95%CI: 0.774 to 0.920]vs 0.679[95%CI: 0.578 to 0.769], P=0.0044), GCVBFD(0.858[95%CI: 0.774 to 0.920] vs 0.726[95%CI: 0.627 to 0.810], P=0.0257) alone(Figure 1).When MMELD was set at 46, it had the highest diagnostic accuracy(85.8%%),with specificity 80%[95%CI:65.4 to 90.4] and sensitivity 87.27%[95%CI:75.5 to 94.7].Conclusion: MMELD is a noninvasive,low-cost and convenient index to assess the risk of EVB in decompensated cirrhosis patients,it is suitable recommended to those patients who decline regular screening endoscopy. Whether this index can reduce the need for screening endoscopy is still questionable.and further prospective studies are encouraged to further evaluate the clinical usefulness of this index.