Methods
Study Cohort and Follow-up
We extracted the data from our prospective database of consecutive patients with cirrhosis admitted at Liver Unit of Hospital Clínic, Barcelona with acute esophageal variceal bleeding between June 2007 and February 2011.
Inclusion criteria for the present study were diagnosis of cirrhosis (based on liver biopsy and/or unequivocal clinical data and compatible findings on imaging techniques); acute bleeding from esophageal varices or gastroesophageal varices type 1 confirmed by emergency endoscopy according to Baveno II-V definitions; initiation of secondary prophylaxis between days 5 and 10 of the index bleeding with the combination of EBL + NSBB (± isosorbide-5-mononitrate [ISMN]). Patients with advanced hepatocellular carcinoma (HCC) at baseline (a single nodule of >5 cm or more than 3 nodules >3 cm), portal vein thrombosis, pregnancy, previous liver transplantation, human immunodeficiency virus infection, known advanced malignancy, and those in whom the outcome could not be assessed (lost before 6 months of follow-up) were not considered for the study.
Patients were classified according to whether they were receiving NSBB as primary prophylaxis (NSBB group) or not (non-NSBB group). On admission, all patients were treated according to current recommended therapy, specifically the combination of an intravenous vasoconstrictor (terlipressin or somatostatin) and endoscopic therapy (EBL), performed <12 hours after admission, together with prophylactic antibiotics. Patients were followed according to the standardized protocols of our unit. Patients with controlled bleeding were started (day 5) on oral propranolol (20 mg twice a day) or nadolol (20 mg once a day), which were stepwise increased if tolerated until heart rate had fallen to ~55/min or systolic blood pressure was below 95 mm Hg. ISMN was initiated in some patients by administering 10 mg at nighttime and increased stepwise up to a maximum of 20 mg twice a day or the maximum tolerated dose. Patients not tolerating a dose increment were kept at the previous dose. EBL sessions were scheduled every 2 weeks until variceal eradication (disappearance of varices or being too small to be sucked into the banding device). EBL sessions were performed by using multiband devices; application of the bands was started at the gastroesophageal junction and progressed upward in a helical way for approximately 5–8 cm. Patients received proton pump inhibitors until reaching variceal eradication. After eradication, first surveillance endoscopy was performed at 1–3 months and then every 6–12 months. If varices reappeared, further EBL sessions were initiated.
All patients were followed in the outpatient clinic at 1, 3, and 6 months and every 6 months thereafter. Medical history, physical examination, biochemistry, hematologic tests, abdominal ultrasound, and continued alcohol abuse were recorded every 6 months. Compliance was assessed on the basis of attendance at scheduled visits, anamnesis, and heart rate. It was considered good when these parameters were met in ≥80% of visits, fair when met in 50%–80% of visits, and poor when met in ≤50% of visits. Follow-up data were collected for up to 2 years, death, or liver transplantation.
The main clinical end point was rebleeding from any source. The secondary end point of the study was orthotopic liver transplantation (OLT)-free survival. For survival analysis, patients with an incomplete follow-up were traced by contacting them or their relatives or, when this was not successful, by means of a request to local civil registry to assess the exact date and cause of death.
This study was approved by the local ethics committee of Hospital Clinic-Barcelona.
Statistical Analysis
Statistical analysis was performed with PASW 18 statistical software package (SPSS, Chicago, IL) and R (http://www.r-project.org). Comparisons between patients with and without previous therapy with NSBB were performed with unpaired Student t test, Mann–Whitney test, or Fisher exact test as appropriate. The risk of developing recurrent bleeding was described with the cumulative incidence function, taking into account death or liver transplantation as competing risks. This provides a more accurate estimation of rebleeding rates than censoring patients at the time of death or liver transplantation in a Kaplan–Meier analysis. Comparisons between patients with and without previous NSBBs were performed with the Gray test. Factors associated with rebleeding were analyzed with the modification of the Cox proportional hazards model proposed by Fine and Gray, again accounting for death and OLT as competing risks. Multivariable analysis was performed with backward stepwise regression, maintaining in the model variables explaining a statistically significant proportion of the variance (P < .1). All competing risks analysis was performed with the R package cmprsk. OLT-free survival was assessed by Kaplan–Meier method, and the comparison between patients with and without primary prophylaxis was performed by the log-rank test. Predictors associated with OLT-free survival were analyzed by bivariable and multivariable Cox model. To reduce possible colinearities, redundant variables were not introduced in the final analysis. For the purpose of the analysis, patients with both viral infection and a history of alcohol consumption were considered to have cirrhosis of viral etiology. The Wald test was used to further assess the homogeneity of the results across relevant strata (alcoholic vs nonalcoholic patients, active alcoholism vs nonactive alcoholism, ISMN treatment vs not ISMN treatment). Significance was established at P < .05.