Chronology of Histological Changes after Band Ligation of Esophageal Varices in Humans

• Introduction
• Patients and Methods
• Case Histories
• Results
• Discussion
• References

Background and Study Aims: While the histological effects of endoscopic sclerotherapy in humans have been extensively described, the effects of endoscopic ligation have been reported in only two cases. The purpose of this study was to reconstruct the chronological sequence of histological changes after ligation of esophageal varices.

Patients and Methods: Autopsy specimens from six patients who received ligation of varices from nine hours to 22 months ante-mortem were evaluated for gross and microscopic changes.

Results: Early after ligation, the appearance was that of a polyp with its base compressed by the band. Variceal thrombosis was seen on day 2. Varying degrees of ischemic necrosis of the polyp were present on days 0-5. If the bands did not remain in situ for two days (premature loss), necrosis of the polyp and dilated variceal vessels were seen. On day 22, superficial ulcers were observed. After complete healing, fibrosis was seen in the submucosa.

Conclusions: The changes seen in the present study are similar to those described in animals. The delay in ulcer healing, compared with the gross changes reported during follow-up endoscopic examinations, may be related to the severity of the underlying illness and the compromised immune status of patients in the present series.

Introduction

Esophageal variceal band ligation has been available for over a decade as an alternative to injection sclerotherapy for the endoscopic treatment of esophageal variceal bleeding [1]. Controlled trials have shown that ligation is superior to sclerotherapy in measures of efficacy, including improved survival [2], and is associated with fewer complications [2] [3] [4] . Ligation is therefore considered to be the treatment of choice for esophageal variceal bleeding [5] [6] .

The histological changes that occur after sclerotherapy are well described in the literature, and are included in standard gastrointestinal pathology texts [7] [8] . By contrast, knowledge of the histological changes that take place after ligation in humans is limited to two case reports, which described the gross and microscopic appearance of lesions on days 5 and 7 after ligation, respectively [9] [10] . Since the continuum of tissue changes in humans has not been described, the relevance of histological events to potential clinical outcomes has been speculative [11].

The present study analyzes postligation lesions in six autopsy cases. It describes the morphological findings at various time intervals, from 9 hours to 22 months, and provides an opportunity to document the chronology of histological changes.

Patients and Methods

The records for six patients who died at various time intervals after ligation treatment were reviewed. The patients had received ligation with rubber bands using the Stiegmann-Goff “single-shot” ligator (C.R. Bard, Inc., Tewksbury, Massachusetts, USA), with the Six-Shooter multiple ligator (Wilson-Cook Medical Inc., Winston-Salem, North Carolina, USA), or with neoprene bands using the Speed-Band multiple ligator (Boston Scientific Corporation, Natick, Massachusetts, USA).

Specimens of the esophagi obtained during autopsy were examined for gross and microscopic changes. Sections of postligation lesions were stained with hematoxylin and eosin and examined for evidence of varix thrombosis, necrosis, and healing of residual ulcers. If microorganisms were seen, their presence was confirmed by tissue Gram staining and Gomori methenamine-silver (GMS) staining. Trichrome staining was used to confirm fibrosis.

Case Histories

Case 1. The patient was a 42-year-old woman with alcoholic cirrhosis and variceal bleeding. Esophageal varices had been eradicated with ligation treatment 22 months previously. The patient was admitted with upper gastrointestinal bleeding after a drinking binge. Esophagogastroduodenoscopy (EGD) revealed bleeding from a gastric ulcer and nonbleeding gastric varices. She received conservative treatment, but continued to have persistent bleeding. On hospital day 7, she had an episode of major rebleeding; EGD showed that the bleeding was from gastric varices. She underwent rubber-band ligation of five gastric variceal sites and one esophageal site, and the bleeding was controlled. However, due to the patient’s poor overall clinical condition, supportive measures were discontinued, and the patient died nine hours after ligation. At autopsy, a perforated gastric ulcer was found.

Case 2. The patient was a 27-year-old man who had undergone orthotopic liver transplantation (OLT) for fulminant hepatic failure due to an acetaminophen overdose. After OLT, the patient had had multiple episodes of cytomegalovirus pneumonitis and liver allograft rejection. His final hospital stay, 18 months after OLT, was complicated by encephalopathy, hemolytic anemia, aspiration pneumonia, and sepsis. Allograft rejection and evidence of progressive fibrosis were seen on serial liver biopsies. The patient had upper gastrointestinal bleeding. Bleeding esophageal varices were seen at the EGD. The bleeding was controlled with rubber bands, and five sites were ligated. The patient had no repeat bleeding episodes, but died the following day. At autopsy, there was microscopic evidence of hepatic venous outflow obstruction, in addition to bronchopneumonia, with rod-shaped bacteria.

Case 3. The patient was a 45-year-old woman with hepatitis C and alcoholic cirrhosis. She had end-stage liver disease and intractable ascites. She was admitted for transjugular intrahepatic portosystemic shunt (TIPS) placement. Upper gastrointestinal bleeding from esophageal varices followed the TIPS procedure. The EGD examination showed bleeding esophageal varices. She received rubber-band ligation of four sites. Continued bleeding required balloon tamponade with a Minnesota tube. The further hospital course was complicated by acute renal failure and disseminated intravascular coagulopathy (DIC). Supportive care was withdrawn, and the patient died two days later.

Case 4. The patient was a 48-year-old woman with polycythemia vera, diagnosed eight years before the final admission. She was admitted with severe epigastric and back pain, followed by massive hematemesis. On EGD, large esophageal varices with red color signs were seen. Five distal esophageal sites received ligation with neoprene bands, and six sites proximal to these received ligation with rubber bands. Further investigations showed a leukemoid reaction and portal vein thrombosis. The patient died two days after ligation, due to massive mesenteric and portal vein thrombosis.

Case 5. The patient was a 62-year-old man with hepatitis C cirrhosis. He was admitted with bleeding esophageal varices, and received rubber band ligation of five sites. The following day, he underwent TIPS. After TIPS, he suffered neurological deterioration, and developed cardiac arrhythmias. He died five days later. The autopsy findings included cirrhosis, multifocal hepatocellular carcinoma with metastases to a mediastinal lymph node and rib, acute myocardial infarction of the septum and anterior free wall, and Staphylococcus aureus and Candida bronchopneumonia with diffuse alveolar damage.

Case 6. The patient was a 47-year-old man who had undergone OLT for hepatitis C and alcoholic cirrhosis. He had recurrent hepatitis C with cirrhosis. Three years after OLT, he was admitted with respiratory and renal failure. He was found to have multiorganism pneumonia, coagulase-negative Staphylococcus septicemia, and encephalopathy. While in hospital, he bled from esophageal varices and received rubber-band ligation of six sites. Subsequently, bleeding recurred, and he received balloon tamponade and TIPS treatment. A second OLT failed intraoperatively due to massive adhesions and hemorrhage. The patient died due to multiorgan failure 22 days later.

Results

Thirty-two variceal sites in the six patients were examined at autopsy; 27 had been treated with rubber bands and five with neoprene bands. The bands were in situ in 13 sites (Table [1]). Retained bands were present in 13 of 16 (81 %) of the “early” sites treated using rubber bands (nine hours to two days after ligation), and no bands were present in the five sites treated with neoprene bands. None of the 11 sites that had been treated five or more days earlier had bands in situ.

The gross appearance of the early ligation lesions that had a band in situ was polypoid, with the base of the polyp compressed by the band, forming a stalk (Figure [1]). The lesions without a band in situ were also polypoid, and consisted of a broad base without a stalk (Figure [2], the five distal sites). Corresponding microscopic changes consisted of a polypoid configuration of the mucosa, with the base consisting of compressed vessels, varices, and muscularis mucosa.

Microscopically, variceal thrombosis was not seen on days 0 or 1 in the lesions studied. On day 2, thrombosis was present in varices with bands in situ. In the varices without bands, there was early ischemic necrosis in the entire mucosa, including the vessels and the dilated varices. While early organization was observed within these vessels, there were no established thrombi. Thrombosis was present in the varices from day 5 after banding.

Ischemic necrosis with epithelial sloughing was seen in the superficial portions of the mucosa in the varices examined on days 0-2. Involvement of the submucosa by coagulative necrosis was evident on day 2. In lesions with rubber bands in situ, the necrosis was prominent in the mucosa compressed by bands. The vascular spaces within the polyp had collapsed and were not filled with blood (Figure [3]). In lesions with no band in situ, mucosal ulceration, submucosal necrosis, and moderate acute inflammation were seen. The dilated, partially necrotic vascular channels of the submucosa were only partially involved by nonocclusive early thrombosis (Figure [4]). Variable degrees of mucosal and submucosal hemorrhage were seen in the early lesions. The day 5 polyp showed markedly dilated, thrombosed vascular channels in the mucosa, replaced by necrotic inflammatory debris (Figure [5]). The feeding vessel outside of the banded area contained a mural thrombus. Nonhealed mucosal and submucosal ulcers were present in the banding sites on day 22 (Figure [6]).

In the patient (case 1) who had also received rubber-band ligation 22 months prior to death, there was a shallow mucosal depression in the distal esophagus. Microscopically, fibrosis was seen in the submucosa. The underlying muscularis propria was normal.

Gram staining and GMS staining confirmed the presence of microorganisms that was suggested by sections stained with hematoxylin and eosin. Extensive colonization of necrotic polyps by budding yeast with pseudohyphae was present in cases 3 and 5 (two and five days after ligation). Focal Gram-negative rods had colonized necrotic tissue in case 2 (one day after ligation). The base of the ulcer in case 6 (22 days after ligation) contained rare Gram-positive cocci.

Discussion

This study documented the sequences of histological changes in lesions observed in patients, ranging from 9 hours to 22 months after ligation treatment. Necrosis and superficial ulceration, previously described in an animal study [12], were also seen in this series. The extent of the necrosis was limited to the mucosa and submucosa, with no involvement of the muscularis propria. This contrasts with the findings after sclerotherapy, where a wide variation in the depth of tissue injury has been observed and deep necrosis was seen in many cases [7]. Differences in the depth of the tissue injury caused by ligation and sclerotherapy are not unexpected. While the maximum volume of tissue that can be ligated is consistent and limited by the volume of the ligating chamber of the ligation device, many (often unpredictable) procedural variables affect sclerotherapy - for example, the type and volume of the sclerosing solution injected, variability in tissue planes, the length of the needle used for the injection, and even if the patient coughs or gags during the injection, pushing the needle even deeper [6].

Early slippage of bands from treated sites is a matter of potential clinical concern in relation to the risk for rebleeding. In-vitro bench studies have shown differences in the resting inner diameters and radial retraction force of bands on different ligation devices. The larger resting inner diameter and a significantly lower retraction force in neoprene bands, compared to rubber bands, predisposed these bands to early slippage [13]. In order to quantify early slippage, a series of 27 patients who had received 63 treatment sessions of esophageal varices with neoprene bands underwent endoscopic examination within 24 hours. In addition to nine bands that were lost to slippage after deployment while the treatment session was still in progress, bands were not found at 93 of 300 treated sites (31 %; range 0 - 100 % per individual patient) during the follow-up endoscopy. The fact that immediate rebleeding had not occurred led the authors to speculate that variceal thrombosis “follows initial trauma of the wall” [11]. The present series shows that this is not the case; thrombosis was seen on day 2 after ligation, but was absent on the day following ligation. In case 4 in this series, there was a unique opportunity to study the gross and histological differences between sites with retained bands and sites with prematurely lost bands, side by side in the same patient. A remarkable difference was obvious between the histological appearances of sites with and without bands (Figures [3] [4] ). Although the patient did not have recurrent bleeding, the necrosis of the mucosa, necrosis of vessels within it, and persistence of dilated varices suggested a high potential for bleeding from the site if the band had been lost early. This might result from the lack of viability of the variceal wall prior to occlusion by a mature thrombus.

Bacterial or fungal colonization of necrotic tissue in varying amounts was seen in four of the six cases. Although similar findings have previously been described in a postligation lesion observed at endoscopy [10], this was not reported in the previous human autopsy case or the animal study [9] [12] . The relatively high frequency of colonization found in the present series may reflect the immunocompromised state and critical medical condition of the patients at the time of, and following, band ligation.

In the present series, nonhealed ulcers were found 22 days after ligation. This finding differs from that of an endoscopic report in which it was estimated that the mean postligation ulcer healing interval was 14 days [14]. In other endoscopic reports, ulcers had healed by three weeks [15] [16] . The placement of a Minnesota tube in the patient (case 6) studied on 22 day may have contributed to the delay in ulcer healing. Moreover, in the previous reports, the patients were neither immunocompromised nor colonized with microorganisms. In contrast to a report of transmural necrosis in postligation ulcers in an immunocompromised patient [17], this complication was not seen in any of the lesions studied in this series.

No obvious lesions were observed at the treated site 22 months after ligation, confirming that complete healing of treated sites occurs over time. The appearance was similar to that seen in the animal study [12], in which only a shallow depression persisted in the mucosa and fibrosis was present in the submucosa 50-60 days after ligation.

The present study is the largest human series reported to date. Although the heterogeneity of the patients may mean that the local tissue reactions observed may not be wholly typical, the study allowed documentation of the progression of chronological changes after band ligation of esophageal varices in humans. The study showed that, in contrast to sclerotherapy, mural injury after ligation extends to the depth of the layer at which varices are present - i. e., the submucosal layer. Thus, ligation may be regarded as inherently safer than sclerotherapy with regard to complications. On a histological basis, one might suspect that early loss of bands from treated sites - i. e., before thrombosis has occurred - would make rebleeding more likely. Whether or not this occurs has not been specifically reported in the literature.

References

• 1 Stiegmann G V, Cambre T, Sun J H. A new endoscopic elastic band ligating device.  Gastrointest Endosc. 1986;  32 230-233
  PubMedSearch in Google Scholar
• 2 Stiegmann G V, Goff J S, Michaletz-Onody P A, et al. Endoscopic sclerotherapy as compared with endoscopic ligation for bleeding esophageal varices.  N Engl J Med. 1992;  326 1527-1532
  CrossrefPubMedSearch in Google Scholar
• 3 Gimson A ES, Ramage J K, Panos M Z, et al. Randomised trial of variceal banding ligation versus injection sclerotherapy for bleeding oesophageal varices.  Lancet. 1993;  342 391-394
  CrossrefPubMedSearch in Google Scholar
• 4 Laine L, El-Newihi H M, Migikovsky B, et al. Endoscopic ligation compared with sclerotherapy for the treatment of bleeding esophageal varices.  Ann Intern Med. 1993;  119 1-7
  PubMedSearch in Google Scholar
• 5 Laine L. Ligation: endoscopic treatment of choice for patients with bleeding esophageal varices?.  Hepatology. 1995;  22 663-665
  CrossrefPubMedSearch in Google Scholar
• 6 Saeed Z A. Endoscopic therapy of bleeding esophageal varices: ligation is still the best.  Gastroenterology. 1996;  110 635-638
  PubMedSearch in Google Scholar
• 7 Evans D MD, Jones D B, Cleary B K, Smith P M. Oesophageal varices treated by sclerotherapy: a histopathological study.  Gut. 1982;  23 615-620
  PubMedSearch in Google Scholar
• 8 Goldman H, Szabo S. Chemical and physical disorders. In: Ming S, Goldman H, editors. Pathology of the gastrointestinal tract Philadelphia; Saunders 1992: 141-170
  Search in Google Scholar
• 9 Marks R D, Arnold M D, Baron T H. Gross and microscopic findings in the human esophagus after esophageal variceal band ligation: a postmortem analysis.  Am J Gastroenterol. 1993;  88 272-274
  PubMedSearch in Google Scholar
• 10 Jalan R, Piris J, Hayes P C. Variceal band ligation and infection [letter].  Am J Gastroenterol. 1993;  88 2140
  PubMedSearch in Google Scholar
• 11 Hochberger J, Reh H, Hörning F, et al. A new multiband ligator for the endoscopic treatment of esophageal and gastric varices: preliminary results in 63 applications of a prospective ongoing study.  Endoscopy. 1996;  28 S14
  PubMedSearch in Google Scholar
• 12 Stiegmann G V, Sun J H, Hammond W S. Results of experimental endoscopic esophageal varix ligation.  Am Surg. 1988;  54 105-108
  PubMedSearch in Google Scholar
• 13 Reh H, Hochberger J, Hörning F, et al. Comparison of ring elasticity and ligation volume of different variceal single and multiple band ligators using a multicylindrical stretching apparatus in a porcine esophagus model in vitro [abstract].  Endoscopy. 1996;  28 S36
  PubMedSearch in Google Scholar
• 14 Young M F, Sanowski R A, Rasche R. Comparison and characterization of ulceration induced by endoscopic ligation of esophageal varices versus endoscopic sclerotherapy.  Gastrointest Endosc. 1993;  39 119-122
  PubMedSearch in Google Scholar
• 15 Nijhawan S, Rai R R, Nepalia S, et al. Natural history of postligation ulcers.  Am J Gastroenterol. 1994;  89 2281-2282
  PubMedSearch in Google Scholar
• 16 Nijhawan S, Rai R R. Does post-ligation oesophageal ulcer healing require treatment?.  Lancet. 1994;  343 116-117
  PubMedSearch in Google Scholar
• 17 Schoonbroodt D, Zipf A, Jung M. Local necrosis and fatal perforation of oesophagus after endoscopic ligation [letter].  Lancet. 1994;  344 1365
  PubMedSearch in Google Scholar

Z. A. Saeed, M.D.

Division of Gastroenterology
Saint Louis University HSC

3635 Vista Avenue
St Louis
MO 63110
USA


Fax: Fax:+ 1-314-577-8125

Email: E-mail:saeedza@slu.edu

References

• 1 Stiegmann G V, Cambre T, Sun J H. A new endoscopic elastic band ligating device.  Gastrointest Endosc. 1986;  32 230-233
  PubMedSearch in Google Scholar
• 2 Stiegmann G V, Goff J S, Michaletz-Onody P A, et al. Endoscopic sclerotherapy as compared with endoscopic ligation for bleeding esophageal varices.  N Engl J Med. 1992;  326 1527-1532
  CrossrefPubMedSearch in Google Scholar
• 3 Gimson A ES, Ramage J K, Panos M Z, et al. Randomised trial of variceal banding ligation versus injection sclerotherapy for bleeding oesophageal varices.  Lancet. 1993;  342 391-394
  CrossrefPubMedSearch in Google Scholar
• 4 Laine L, El-Newihi H M, Migikovsky B, et al. Endoscopic ligation compared with sclerotherapy for the treatment of bleeding esophageal varices.  Ann Intern Med. 1993;  119 1-7
  PubMedSearch in Google Scholar
• 5 Laine L. Ligation: endoscopic treatment of choice for patients with bleeding esophageal varices?.  Hepatology. 1995;  22 663-665
  CrossrefPubMedSearch in Google Scholar
• 6 Saeed Z A. Endoscopic therapy of bleeding esophageal varices: ligation is still the best.  Gastroenterology. 1996;  110 635-638
  PubMedSearch in Google Scholar
• 7 Evans D MD, Jones D B, Cleary B K, Smith P M. Oesophageal varices treated by sclerotherapy: a histopathological study.  Gut. 1982;  23 615-620
  PubMedSearch in Google Scholar
• 8 Goldman H, Szabo S. Chemical and physical disorders. In: Ming S, Goldman H, editors. Pathology of the gastrointestinal tract Philadelphia; Saunders 1992: 141-170
  Search in Google Scholar
• 9 Marks R D, Arnold M D, Baron T H. Gross and microscopic findings in the human esophagus after esophageal variceal band ligation: a postmortem analysis.  Am J Gastroenterol. 1993;  88 272-274
  PubMedSearch in Google Scholar
• 10 Jalan R, Piris J, Hayes P C. Variceal band ligation and infection [letter].  Am J Gastroenterol. 1993;  88 2140
  PubMedSearch in Google Scholar
• 11 Hochberger J, Reh H, Hörning F, et al. A new multiband ligator for the endoscopic treatment of esophageal and gastric varices: preliminary results in 63 applications of a prospective ongoing study.  Endoscopy. 1996;  28 S14
  PubMedSearch in Google Scholar
• 12 Stiegmann G V, Sun J H, Hammond W S. Results of experimental endoscopic esophageal varix ligation.  Am Surg. 1988;  54 105-108
  PubMedSearch in Google Scholar
• 13 Reh H, Hochberger J, Hörning F, et al. Comparison of ring elasticity and ligation volume of different variceal single and multiple band ligators using a multicylindrical stretching apparatus in a porcine esophagus model in vitro [abstract].  Endoscopy. 1996;  28 S36
  PubMedSearch in Google Scholar
• 14 Young M F, Sanowski R A, Rasche R. Comparison and characterization of ulceration induced by endoscopic ligation of esophageal varices versus endoscopic sclerotherapy.  Gastrointest Endosc. 1993;  39 119-122
  PubMedSearch in Google Scholar
• 15 Nijhawan S, Rai R R, Nepalia S, et al. Natural history of postligation ulcers.  Am J Gastroenterol. 1994;  89 2281-2282
  PubMedSearch in Google Scholar
• 16 Nijhawan S, Rai R R. Does post-ligation oesophageal ulcer healing require treatment?.  Lancet. 1994;  343 116-117
  PubMedSearch in Google Scholar
• 17 Schoonbroodt D, Zipf A, Jung M. Local necrosis and fatal perforation of oesophagus after endoscopic ligation [letter].  Lancet. 1994;  344 1365
  PubMedSearch in Google Scholar

Z. A. Saeed, M.D.

Division of Gastroenterology
Saint Louis University HSC

3635 Vista Avenue
St Louis
MO 63110
USA


Fax: Fax:+ 1-314-577-8125

Email: E-mail:saeedza@slu.edu