Key points

Introduction

In the past decade, numerous studies have attested to the efficacy of endoscopic therapy in controlling ulcer hemorrhage. In 2 separate meta-analyses, endoscopic therapy has been shown to reduce the rate of recurrent bleeding; the need for surgery; and, most importantly, in-hospital deaths. In the modern literature, need for surgery is defined as a treatment failure in most trials of endoscopic therapy. In most centers, surgery is reserved for patients with failed endoscopic therapy. A recent study from Denmark that included all hospitals managing acute peptic ulcer bleeding analyzed data on 13,498 patients with peptic ulcer bleeding. Despite improvement in all outcome indicators and a decline in patients requiring surgical intervention, 30-day mortality remained at 11%. Surgery maintains an important gatekeeping role despite its diminished role in the management algorithm of bleeding peptic ulcers. The precise role of surgery in the context of endoscopic hemostasis therapy has not been well studied.

Before endoscopic therapy was widely available, controversies existed as to when was best to apply surgery and which type of surgery should be performed when a patient required it. The goal of surgery was not merely the cessation of hemorrhage but as definitive therapy so as to prevent recurrence of peptic ulcer. In the 1980s, proponents of an aggressive, definitive surgical approach to the management of bleeding peptic ulcer reported an admirable overall mortality of 20%. However, there was an increased risk of major adverse events. A minimal surgical approach to achieve primary hemostasis in a significantly shorter surgery time was advocated by other investigators who argued that the minimal surgical approach would result in even lower surgical morbidity and mortality.

In the modern era of endoscopic management for bleeding peptic ulcer, there are challenges and controversies in the timing and type of surgical procedures to salvage ulcer rebleeding. In this article, the surgical approach to the management of bleeding peptic ulcers is reviewed with a focus on the current evidence and modern surgical techniques to achieve surgical hemostasis.

Introduction

In the past decade, numerous studies have attested to the efficacy of endoscopic therapy in controlling ulcer hemorrhage. In 2 separate meta-analyses, endoscopic therapy has been shown to reduce the rate of recurrent bleeding; the need for surgery; and, most importantly, in-hospital deaths. In the modern literature, need for surgery is defined as a treatment failure in most trials of endoscopic therapy. In most centers, surgery is reserved for patients with failed endoscopic therapy. A recent study from Denmark that included all hospitals managing acute peptic ulcer bleeding analyzed data on 13,498 patients with peptic ulcer bleeding. Despite improvement in all outcome indicators and a decline in patients requiring surgical intervention, 30-day mortality remained at 11%. Surgery maintains an important gatekeeping role despite its diminished role in the management algorithm of bleeding peptic ulcers. The precise role of surgery in the context of endoscopic hemostasis therapy has not been well studied.

Before endoscopic therapy was widely available, controversies existed as to when was best to apply surgery and which type of surgery should be performed when a patient required it. The goal of surgery was not merely the cessation of hemorrhage but as definitive therapy so as to prevent recurrence of peptic ulcer. In the 1980s, proponents of an aggressive, definitive surgical approach to the management of bleeding peptic ulcer reported an admirable overall mortality of 20%. However, there was an increased risk of major adverse events. A minimal surgical approach to achieve primary hemostasis in a significantly shorter surgery time was advocated by other investigators who argued that the minimal surgical approach would result in even lower surgical morbidity and mortality.

In the modern era of endoscopic management for bleeding peptic ulcer, there are challenges and controversies in the timing and type of surgical procedures to salvage ulcer rebleeding. In this article, the surgical approach to the management of bleeding peptic ulcers is reviewed with a focus on the current evidence and modern surgical techniques to achieve surgical hemostasis.

Factors affecting mortality from bleeding peptic ulcers

In a population-based United Kingdom audit, a crude mortality figure of 14% was reported among 4185 patients with bleeding peptic ulcer. Mortality was substantially higher among in-patients who developed bleeding (33%) compared with those patients presenting to the emergency department for bleeding (11%). The salient features of the study were that the overall incidence of bleeding was high (103 per 100,000 adults/y) and most patients were elderly. Rockall and colleagues developed a well-known prediction score for mortality from a prospective multicenter population-based study involving 4185 cases of acute upper gastrointestinal (GI) hemorrhage. The Rockall score was subsequently validated in another cohort of 1625 patients. The Rockall score takes into account age, shock, comorbidity, endoscopic diagnosis, and the presence of major stigmata of recent hemorrhage. It has been shown that the risk of mortality increases when the score increases. Our group studied the predictive factors to bleeding peptic ulcer–related mortality after therapeutic endoscopy among 3220 patients. Ulcer rebleeding and need for surgery were identified as significant risk factors for mortality, together with older age, multiple comorbidities, hemodynamic shock, and in-hospital bleeding. In another study, we investigated the causes of mortality among 10,248 cases of upper GI bleeding. Although most patients died secondary to nonbleeding causes, 18.4% died of bleeding-related causes. The most common bleeding-related causes of death included failure to control bleeding during initial endoscopy (primary hemostasis), death within 48 hours of receiving endoscopic therapy, and surgical adverse events. Death was most common among elderly patients with severe comorbidities. Physicians are now faced with an increasing proportion of elderly patients and a further reduction in mortality secondary to improved management is likely to be marginal.

Peptic ulcer rebleeding after therapeutic endoscopy

The current strategies to prevent ulcer rebleeding after therapeutic endoscopy include adjunctive high-dose proton pump inhibitor (PPI) infusion or performance of scheduled second-look endoscopy. Prospective randomized trials show that both strategies significantly reduce the rate of peptic ulcer rebleeding. With the consideration of patient discomfort, workload, and costs of scheduled second-look endoscopy, PPI infusion is advocated as the preferred approach after therapeutic endoscopy. The rate of rebleeding after therapeutic endoscopy and PPI infusion is now 5% to 10%. Patients who develop rebleeding are those at high risk for surgical procedures because they are usually elderly with multiple comorbidities.

Predicting failure of endoscopic therapy

Because rebleeding and need for surgery impose significant patient risk, predicting failure of endoscopic therapy could facilitate intensive monitoring and earlier surgical intervention among these patients. Several studies have defined predictors of rebleeding after endoscopic therapy. Of the risk factors evaluated in these studies, the most consistent predictor of rebleeding seems to be the location of the ulcer in the posterior duodenal bulb. Other independent predictors of failure of endoscopic therapy include ulcer size, comorbid illnesses, and older age ( Table 1 ). Among 3386 patients who had bleeding peptic ulcers receiving therapeutic endoscopy, our group showed that hypotension, initial hemoglobin level less than 10 g/dL, fresh blood in the stomach, actively bleeding ulcers, and large ulcers were independent factors to predict rebleeding. Parameters identified from these studies are potentially helpful in selecting high-risk patients for planned urgent surgery.

Early versus delayed surgery for bleeding peptic ulcer

Before the era of therapeutic endoscopy, surgery was the only effective means to stop bleeding in peptic ulcers. The timing for surgery has been a subject of intense debate in the last few decades. Because continued or recurrent bleeding was the single most adverse prognostic factor, surgeons and gastroenterologists aimed to stop the hemorrhage promptly. Patients who sustain rebleeding are often elderly with severe comorbid illnesses who poorly tolerate blood loss. Early surgery in this subgroup of patients theoretically neutralizes the risk of recurrent bleeding and death. A group from Nottingham, United Kingdom, retrospectively analyzed short-term outcomes of 908 patients with GI bleeding during 1975 to 1980. Twenty-seven percent of patients underwent operations. Comparing the years 1975 to 1977 with 1978 to 1980, the operation rate decreased from about 33% to 21%. The reduction in the rate of operation had no appreciable effect on mortality (13.9% vs 12.9% in gastric ulcers, 9.4% vs 9% in duodenal ulcers). Dronfield and colleagues analyzed data from 2 hospitals in Nottingham with different policies in offering surgery. Of 206 patients from the first hospital, 66 (32%) were operated on compared with 44 (46%) of 96 at another hospital ( P = .03). Overall mortality was less in the hospital with a lower rate of operations (12.7% vs 17.7%). The investigators questioned the role of aggressive surgical policy in bleeding peptic ulcer. Rofe and colleagues from Australia reported a series of 86 patients with bleeding gastric and duodenal ulcers in which only 5.1% were operated on. One of 4 patients died after surgery but the overall mortality was only 3.5%. The investigators suggested that a more conservative approach be adopted. Hunt and colleagues advocated aggressive resuscitation in a specialized unit and early surgery in patients with bleeding peptic ulcers. In a decade-long study of 633 patients, 206 patients underwent emergency surgery with an operative mortality of 12%, a decline from the operative mortality of 17.6% that occurred among 142 patients during the previous decade. In a later prospective study over a 6-year period, 376 patients were recruited with bleeding peptic ulcers. Surgery was performed for patients with exsanguinating hemorrhage or shock on admission more than 50 years of age or further bleeding. The operation rate was 21.7% (102 patients). The overall mortality was 7%.

The only prospective randomized study to compare early and delayed surgery in bleeding ulcers came from the Birmingham group. One hundred and four patients were randomized. Criteria for early surgery were 4 units of blood or plasma expander transfused in 24 hours, 1 rebleed, and endoscopic stigmata or 1 previous bleed with 2 years. Criteria for surgery in the delayed group were 8 units of blood or plasma expander in 24 hours, 2 rebleeds, and persistent bleeding requiring 12 units of blood in 48 hours or 16 units in 72 hours. In patients less than 60 years of age, there was no death in either group but the early surgery policy led to an unacceptably high operation rate (52% in the early group and 5% in the delayed group). For patients aged more than 60 years, the operation rate was 62% in the early group and 27% in the delayed group. There were 3 deaths among 48 patients (6%) in the early group compared with 7 deaths among 52 patients (13%) in the delayed group. The difference was not statistically significant on an intention-to-treat analysis. The trial was criticized for the low number of patients in the subgroup analysis and for allowing ongoing exsanguination of patients in the delayed group. In a posttrial 4-year audit from the same group, indications for urgent surgery were modified; patients were operated on if they had exsanguinating hemorrhage or if a spurting vessel was seen at endoscopy. Patients more than 60 years of age were operated on if 1 episode of rebleeding occurred in hospital, and 4 units of blood or colloid for volume replacement or 8 units of blood or colloid were transfused over 48 hours. Of 342 patients, 214 were 60 years of age or older. Fifty-two (24%) of them received surgical hemostasis with an operative mortality of 6%. The evidence for early surgical intervention was inadequate, especially in the modern era in which therapeutic endoscopy has become the primary treatment of bleeding ulcers. However, the trials on timing of surgery clearly show that early intervention was critical in elderly patients. Under a centralized joint medical and surgical unit, early surgical intervention among patients at high risk resulted in a lower overall mortality of 3.7% and surgical mortality of 15.2%.

Planned urgent surgery has been the dogma in the management of bleeding ulcers for many years. Benders and colleagues operated on patients with shock on admission, age greater than 65 years, ulcer size greater than 2 cm, or with stigmata of recent hemorrhage and previous admission for ulcer complication. Sixty-six patients (mean age of 58 years) were included in a 5-year period with no mortality. Mueller and colleagues operated on patients with spurting hemorrhage, nonbleeding visible vessels on posterior duodenal ulcers, blood transfusions greater than 6 units in the first 24 hours, and rebleeding within 48 hours. In a consecutive series of 157 patients, the 30-day mortality was 7%. At present, a planned surgical procedure before ulcer rebleeding is not practically possible with an increasing trend of elderly and high-risk patients, especially when no good selection criteria are available for predicting those who might benefit.