Differences in bleeding behavior after endoscopic band ligation: a retrospective analysis

Portal hypertension is a hemostatic disorder that is often caused by liver cirrhosis. It is defined as an increase in blood pressure in the portal venous system. Portal hypertension can be the origin of severe complications in cirrhosis. Gastroesophageal varices are a potentially life-threatening complication of portal hypertension. Even though the outcome of bleeding from gastroesophageal varices has improved, it is still associated with a high mortality rate. The mortality rate of acute variceal bleeding has decreased from 42% in 1981 to about 15% to 20% at present [1–4].

Risk factors for bleeding from esophageal varices have been identified in several randomized controlled trials (RCT). Poor Child-Pugh score [3, 5–7], bacterial infection [6–8], elevated aspartate amino transferase levels [3], and a hepatic venous pressure gradient greater than 20 mmHg, measured shortly after admission [5, 6, 8–10], are associated with a higher bleeding rate.

Management of acute variceal bleeding consists of pharmacologic and endoscopic treatment. Endoscopic techniques currently used to stop variceal bleeding are endoscopic sclerotherapy, variceal obturation with glue and endoscopic band ligation (EBL). EBL is the endoscopic treatment of choice because it is an effective treatment for bleeding from esophageal varices. It is generally accepted and recommended for the treatment of acute bleeding events. EBL is also used for the secondary prophylaxis of esophageal variceal hemorrhage [11–15].

Non-selective beta-blockers are the treatment of choice to prevent initial bleeding episodes. However, beta-blockers do not prevent formation of esophageal varices. For primary prophylaxis of esophageal bleeding, EBL should be restricted to patients with contraindications or intolerance to beta-blockers [14, 16–19].

EBL sessions are usually repeated at 1-2 week intervals until complete obliteration of all esophageal varices has been achieved [15]. Complete obliteration is achieved in about 90% of patients after 2-4 sessions. Variceal recurrence occurs frequently, with 20%-75% of patients requiring repeated EBL sessions.

There is numerous data, including randomized controlled trials and meta-analyses, concerning the safety of EBL and its clinical effectiveness in controlling variceal bleeding. However, there are only few studies that have evaluated the risk of bleeding from ligation-induced ulcers. The results of these few studies are contradictory [19–21]. Stiegmann et al. [21] reported an overall complication rate of only 2%. Schepke et al. [19] completed a multicenter trial comparing EBL with beta-blocker therapy for primary prophylaxis of variceal bleeding. This study found the incidence of bleeding from ligation ulcers after EBL to be 6.7%. Another retrospective data analysis of EBL described hemorrhage from ligation ulcers as 5.7%, irrespective of the indication [20].

Meetings were held to discuss clinical trials and propose guidelines for the management of patients with esophageal varices. During the proceedings at the Baveno IV workshop in April 2005, attendees came to an agreement about the management of endoscopic treatment of esophageal varices [14]. However, as far as we know, there is no data or guidelines concerning endoscopic control of ligation ulcers. Furthermore, there is no data or guidelines on whether EBL should be restricted to in-patients. Most gastroenterologists, at least in Germany, do not perform out-patient EBL. Patients are kept under surveillance until all ligation bands are dropped off. This monitoring is based on the theory that the process of dropping off ligation bands is associated with an elevated risk of recurrent bleeding.

EBL is routinely used in nearly all gastrointestinal endoscopy units. Tremendous efforts are invested in the therapy of esophageal varices. Despite this, there is no evidence how to monitor patients after EBL. Gastroenterologists, like all other physicians, are increasingly under economic pressure. German health insurance funds and hospital administration executives try continuously to further minimize the duration of hospitalizations. We therefore conducted a retrospective study of endoscopic band ligation procedures with regard to the primary outcome of bleeding complications after EBL. Furthermore we wanted to evaluate whether ligation ulcers after EBL are a locus of high risk for procedure associated bleeding events.

We observed a significant difference in bleeding events when EBL for treatment of acute variceal bleeding was compared to electively performed ligations. The incidence of hemorrhage after acute intervention was 12.1% (22/182) compared to 3.9% (8/205) after elective treatment (p = 0.004). The bleeding rate from ligation sites in the group that underwent emergency ligation was 7.1% (13/182). This was more than fourteen times higher than the group that underwent elective ligation treatment (0.5% (1/205); p < 0.001). However, bleeding incidence from recurrent varices did not vary significantly. Bleeding from recurrent varices occurred in seven patients (3.4%) in the elective group and in eight patients (4.4%) in the emergency group.

The average number of ligation bands applied per session is shown in Tables 4 and 5. Significantly more ligation bands were applied in interventions that were followed by hemorrhage from ligation ulcers (p = 0.009). We found no significant difference in the number of bands applied when we compared those with bleeding from recurrent varices and those who had no bleeding at all.

An analysis of the time until bleeding in all patients showed a predominance of bleeding events during the first four days after endoscopic treatment (figure 2). Sixty-seven percent of bleeding events occurred in this period of time. The majority of these events were hemorrhage from recurrent varices. There were eight events from ligation sites and eleven events from recurrent varices. In addition, there was one hemorrhage without clear determination of localization. Seventy-five percent (6/8) of all bleeding episodes after elective treatment occurred during the first four days after EBL treatment. Most rebleedings after emergency ligation were also observed within the first 4 days, i.e. sixty-four percent (14/22) (figure 3). The latest report of a bleeding after elective ligation treatment was on the 11th day of the follow-up period.

Irrespective of the indication for EBL or the localization of bleeding, the cumulative risk of bleeding events was 7.8%. The initial bleeding risk of 2.6% in the first four days after the endoscopic procedure was reduced to 0.5% after 11 days (figure 4). The cumulative rebleeding risk within the first four days after emergency ligation is 0.077. After electively performed treatment, the cumulative bleeding risk is 0.029. A comparison using the Mann-Whitney-Test determined this difference to be significant (p = 0.04) (figure 5).

Endoscopic band ligation (EBL) is the gold standard for the treatment of acute bleeding of esophageal varices. It is also effective in secondary prophylaxes and in primary prophylaxis for patients who are insensitive to beta-blocking agents. EBL is mostly performed in the in-patient setting. Patients are often hospitalized until endoscopy proves all applied ligation bands have dropped off. However, there is no proven, scientific rational for this strategy.

This single center, retrospective analysis studied 255 patients who underwent endoscopic ligation therapy. The aim of this study was to propose an algorithm for the surveillance of patients after EBL. We documented an overall bleeding rate after EBL of 7.8%. Since this rate is consistent with those reported in other studies (9% to 19%) [22–28], it emphasizes that our results are comparable with those of others. These studies were performed for primary prophylaxes [22, 23, 25, 26, 28], secondary prophylaxes [24] and control of acute [27] variceal bleeding. The results should be therefore comparable with results obtained in our collective.

One important result of this study was that elective ligation therapy had a significantly lower incidence of procedure related bleeding events than treatment of acute bleeding events. According to our knowledge, there are no other studies relating the bleeding rate after EBL to its indication.

When ligation bands are dropping off, ligation ulcers may form. This process is thought to be associated with an elevated risk of bleeding complications. For this reason, patients are often hospitalized until all applied ligation bands drop off. We found that bleeding from EBL induced ulcers occurred only in 3.6% (14/387) of ligation sessions. In a meta-analysis comparing EBL with sclerotherapy, Laine et al. [29] reported on bleeding from treatment-induced ulcers in seven studies. These studies had patient populations smaller than ours, and ranged from n = 8 to n = 64 for EBL groups. Overall, the documented bleeding rate from treatment-induced ulcers after EBL in these studies was 3.3% (9/274). In later published studies, this value has been reported as 5.4% to 14.2% in patients who underwent the procedure for control of acute variceal hemorrhage [30–32].

These results are consistent with our study. However, to our knowledge, there is no data correlating the indication for EBL to the bleeding rate of ligation ulcers. In our patients, bleeding from ligation ulcers occurred in 7.1% (13/182) of cases after emergency ligation. After electively performed EBL, bleeding at ligation ulcers occurred only in 0.5% (1/205) of cases. Regarding our entire patient cohort, the incidence of bleeding from ligation-induced ulcers was 3.6% (14/387). This was slightly lower than the bleeding rate from recurrent varices, which was 3.9% (15/387). These findings demonstrate that bleeding from ligation ulcers after EBL is not the most important risk factor for bleeding. These data also show that there is a significant difference in bleeding incidence between ligation-induced ulcers after electively performed ligation and after EBL for treatment of acute hemorrhage.

Our results are supported by data concerning bleeding events in recurrent varices. The overall incidence of 3.9% (15/387) in our study is lower than data reported by others (11.9% to 28.6%) [29–32]. However, these studies were designed to compare different endoscopic techniques for the treatment of acute hemorrhage of esophageal varices. Therefore, they did not consider electively performed EBL. Clinical trials have been performed to analyze the prophylactic potential of endoscopic band ligation in elective interventions. In these studies, the incidence of hemorrhage from recurrent varices ranged from 12.5% to 31.7%. Bleeding from ligation-induced ulcers varied from 2.4% to 6.7% [19, 26, 33]. This data originates from various studies with different indications for EBL. Therefore, the studies are difficult to compare. Studies often do not report the different sources of bleeding in their patients. The need for data concerning EBL in all indications is important for everyday care. Our study is the first analysis of bleeding risk from treatment-induced ulcers or recurrent varices after EBL that focuses on the indication for the interventions.

We were able to show a correlation between the number of applied ligation bands and the bleeding events from ligation ulcers. In most trials, the total number of applied bands per patient seems to have no correlation with bleeding events. Harewood et al. compared the number of bands per session and the bleeding incidence [24]. In this small collection of 40 patients, no significant correlation between the median number of bands used in EBL sessions and bleeding episodes was found. In our study, patients with bleeding events at ligation sites have been treated with significantly more ligation bands. Therefore, one can assume that application of a higher number of ligation bands to treat esophageal varices is a risk factor for bleeding events. Different studies examined the impact of endoscopic therapy of esophageal varices on portal venous pressure [34–36]. The results are not very consistent. However, Ramirez and coworkers showed in a prospective study that there is no better outcome for patients when more than six bands are applied compared to patients treated with a maximum of six ligation bands [37].

We are aware that our study is only a retrospective data analysis and prospective studies should be conducted in order to evaluate policy on EBL procedures. Ligation procedures enrolled in this study could be considered as a further limitation. We reported the analysis of all ligation sessions performed during the study period, because the aim of our study was the evaluation of the procedure associated risk of bleeding after EBL. One could raise concern about this evaluation because there are patients who underwent more then one EBL procedure and were included more than once in data analysis. We therefore recalculated all data after exclusion of all patients who underwent more than one EBL procedures from re-entry into data analysis after first ligation session. The reported results remained significant. (See "Additional file 1", for graphs see "Additional file 2" and "Additional file 3")

Our study confirms former reports that have demonstrated the safety and effectiveness of endoscopic band ligation for treatment of esophageal varices. The risk of bleeding from treatment-induced ulceration is lower after elective EBL than after emergency intervention. We, therefore, propose that endoscopists may consider elective EBL as an out-patient procedure. In cases when EBL is performed as an in-patient procedure, one may consider restricting the period of surveillance after elective EBL to four days. Elective EBL should be done until all varices are eradicated. An excessive application of ligation bands should be avoided. However, we propose to keep patients who have undergone endoscopic band ligation due to acute esophageal hemorrhage under medical surveillance for at least 8-11 days.