Evidence-based, cost-effective management of lower gastrointestinal bleeding. An algorithm of the journal of trauma and acute care surgery emergency general surgery algorithms work group

This algorithm was developed by the Journal of Trauma and Acute Care Surgery Emergency General Surgery algorithms working group to provide an evidence-based, practical approach to the initial evaluation and management of lower gastrointestinal (LGI) bleeding in adults presenting in the emergency setting.1 The algorithm (Fig. 1) is intended to serve as a bedside reference for clinicians. It is annotated with letters that link to the corresponding text, which provides the rationale and references to support these recommendations. The algorithm is not a substitute for the clinical judgment and experience of bedside clinicians and should not be considered the “standard of care.” We encourage institutions to use these recommendations to formulate local clinical protocols, recognizing that patient- and institutional-specific factors may necessitate deviations from this algorithm.Figure 1: LGI bleeding algorithm landscape.The reported incidence of LGI bleeding is 33 to 87/100,000 and accounts for 3% of emergency surgical referrals.1–4 The population of patients with LGI bleeding tends to be older with increased number of comorbidities than patients with upper gastrointestinal hemorrhage. While patients with LGI bleeding necessitate emergency surgical consultations, the optimal management of patients who present with hemodynamic instability requires a multimodality team comprised of acute care surgeons, gastroenterologists, and interventional radiologists.5–8 While the textbook definition of LGI bleeding is any bleeding distal to the Ligament of Treitz, the algorithm predominantly describes the approach to patients with LGI bleeding occurring between the ileocecal valve and the anal verge (colon, rectum, and anus). Small bowel hemorrhage is discussed in a separate section later in this manuscript. The most common etiology of LGI is diverticular disease, which accounts for up to 25% to 50% of cases.9 This is followed by anal and rectal pathology (hemorrhoids, fissures, and rectal ulcers), colitis, arteriovenous malformations, cancer, and polyps. Of note, in upwards of one quarter of patients, the source of their LGI bleed is never definitively found. The most common presentation of LGI bleeding is hematochezia or red blood per rectum.6–9 Compared to other forms of gastrointestinal bleeding; patients often seek medical attention urgently. One reason is that patients often present after a bowel movement, and the mixture of even a small quantity of blood with the water in the commode can be quite alarming, making the volume of hemorrhage appear quite large. LGI bleeding, even if severe, is self-limited in ≥80% of patients allowing time for a workup to be undertaken.6–8 LGI bleeding generally occurs in an older population (>70 y) with concomitant comorbidities, frailties, and in the presence of anticoagulation.10,11 While the overall mortality from LGI bleeding is low, ranging between 2% and 4%, the incidence is higher in older patients or those with comorbidities. Death is more commonly associated with the complications of bleeding rather than hemorrhage itself.12 Assessment As with any patient who is bleeding, assessment of the degree of hemorrhage is paramount. Vital signs, shock index, lactate, and base deficit remain crucial for assessing the presence and depth of hypoperfusion.13 It should also be recognized that in the older population, biochemical markers of shock such as lactate and base deficit may be less apparent than in a younger cohort.14 Blood should also be sent for coagulation studies and, if available, viscoelastic testing. As the population of patients with LGI bleeding often has significant cardiovascular disease, eliciting a medication history, especially anticoagulants, are critical. The use of anticoagulants such as warfarin, direct oral anticoagulants, or dual platelet therapy is associated with an increased risk of all GI bleeding.15,16 The decision-making regarding the management of anticoagulation in these patients is complicated and balances the risk of bleeding against the underlying indication for anticoagulation. These considerations will be discussed later in this manuscript. In addition to the standard physical examination, careful attention should be given to inspection and examination of the anus and distal rectum for the presence of hemorrhoids or other ano-rectal pathologies. Unless overt bleeding can be seen, current data does not support the performance of sigmoidoscopy in the unprepared colon.6–8 Hemodynamically Stable Patients The optimal way to categorize patients with LGI bleeding is to separate those who are hypotensive, require transfusion, or have signs of hypoperfusion from patients without signs of shock or significant blood loss. In the latter group, the decision to admit and evaluate or discharge the patient from the ED is significant. The most used and studied risk-adjusted tool is the Oakland Score (Table 1). This score was originally developed in 2,336 patients from 143 hospitals in the United Kingdom.17 The model was designed to identify predictors of safe discharge based on the minimal risk of rebleeding, blood transfusion, therapeutic intervention, 28-day readmission, or death. A score of ≤8 had a 95% chance of predicting a safe discharge. This score was further studied and validated in 38,067 patients across 140 hospitals in the United States and has been generally accepted as a standard for recommending discharge of patients with LGI bleeding from the ED with plans for further investigation as an outpatient.18,19 TABLE 1 - Oakland Score Variable Value Age group (y) ≤39 0 40–69 1 ≥70 2 Sex Female 0 Male 1 Previous hospitalization for LGI bleeding No 0 Yes 1 Initial heart rate (beats/min) ≤69 0 70–89 1 90–109 2 ≥110 3 Initial systolic blood pressure (mm Hg) 50–89 5 90–110 4 120–129 3 130–159 2 ≥160 0 Initial hemoglobin (g/dL) 3.6–6.9 22 7.0–8.9 17 9.0–10.9 13 11.0–12.9 8 13.0–15.9 4 ≥16.0 0 Total score is the sum of the six component values. An Oakland score ≤8 predicted safe hospital discharge, which was defined as the absence of all of the following: rebleeding, defined as additional blood transfusions requirements or a further decrease in hematocrit of 20% or more after 24 hours of clinical stability, red blood cell transfusion, therapeutic intervention to control bleeding, defined as endoscopic, radiologic, or surgical hemostasis, in-hospital death, and readmission with further LGI bleeding within 28 days. These patients are considered candidates for discharge from the ED. The ACG Guideline also recommends the use of another risk stratification tool, the SHA2PE score (Table 2), which predicts low risk of hospital-based intervention.20 It should be recognized that no prospective multicenter studies have validated the use of either the Oakland or SHA2PE scores in the ED. While these scores can help guide admission decisions, clinical judgment regarding a specific patient’s frailty, comorbidities, and distance from a health care facility able to care for them might influence the decision to pursue a workup during that index admission. TABLE 2 - SHA2PE Score for Prediction of Hospital-based Interventions Variable Score Systolic blood pressure <100 mm Hg 1 Hemoglobin (g/dL) <10.5 2 10.5–12.0 1 Antiplatelet therapy 1 Anticoagulant therapy 1 Pulse rate >100 bpm 1 Emergency Department bleeding 1 SHA2PE Score ≤1 indicates that hospital-based intervention for LGI bleeding is unlikely. Hospital-based intervention is defined as requirement for red blood cell transfusion, endoscopic hemostatic therapy, arterial embolization, or surgery. Hemodynamically Unstable Patients As with trauma patients, those presenting with overt hypotension, tachycardia, or other signs of hypoperfusion should undergo hemostatic resuscitation with whole blood and/or 1:1:1 resuscitation.21 Unsurprisingly, patients who present with hemodynamic instability from gastrointestinal bleeding have higher mortality and more complications (rebleeding, longer length of stay) than those who do not. In a recent meta-analysis, while most of the included studies focused on upper gastrointestinal (UGI) hemorrhage, similar findings were observed in studies that examined LGI bleeding.22 Tranexamic acid (TXA) is currently recommended as part of the resuscitation schema in trauma patients with active hemorrhage. In contrast, the efficacy of TXA for patients with GI hemorrhage is much less certain. A multicenter randomized placebo-controlled study investigated the use of TXA in patients with all causes of acute GI bleeding.23,24 There was no difference in death due to bleeding between the two groups, but the TXA group had a higher rate of venous thromboembolic events. A limitation of the study was that only 11% of the patients had LGI bleeding, and about half had bleeding due to varices. In addition, efficacy of TXA for cessation of LGI bleed was not assessed in this study. Another small prospective study found that TXA had no effect on blood loss.25 A recent meta-analysis found that TXA was associated with increased mortality in patients with LGI bleeding.26 These data suggest that TXA has no positive effects and has possible negative effects in patients with LGI bleeding.6–8,27 A common presentation of patients with massive UGI bleeding is hematochezia. Thus, in patients with active bleeding per rectum, after examining the anus and rectum, one must consider an upper GI source. If an upper source is identified, the recent Journal of Trauma and Acute Care Surgery EGS upper GI bleeding algorithm should be consulted.28 Nasogastric tube lavage is no longer routinely recommended due to its low sensitivity and low negative predictive value for detecting UGI bleeding.29 However, patients with lower GI bleeding who undergo this diagnostic procedure may have a shorter time to complete their workup. The presence of a nasogastric tube may also facilitate the administration of bowel preparation for colonoscopy. CT Angiography Following resuscitation, patients should undergo emergent CT angiography (CTA) of the abdomen and pelvis. The study should be ordered specifically to include an arterial-phase CT angiogram. The advantages of CT angiography are that it requires no bowel preparation and can be performed very rapidly with modern multidetector CT in almost all institutions. It can also identify an upper or lower gastrointestinal bleeding source in patients with active hemorrhage, further streamlining management. In retrospective clinical studies of LGI bleeding, CTA had a reported sensitivity of 79% to 95% and a specificity of 95% to 100%.30,31 The American College of Gastroenterology and Society of Abdominal Radiology 2024 Guidelines recommend that CTA should be performed as the first diagnostic study in hemodynamically unstable patients or those patients where there is a high degree of suspicion for ongoing hemorrhage.8 CTA is not indicated as a first-line test in hemodynamically stable patients in whom bleeding has subsided since it is of low yield. In this group, colonoscopy is recommended instead (section F). The disadvantage of CTA is that it exposes the patient to radiation and requires an active hemorrhage estimated at ≥0.3 mL/min to localize the source. In a porcine experimental model, sensitivity and specificity were 50%, with bleeding rates of 0.1 mL/min, which increased to 95% with rates ≥0.5 mL/min.32 In a meta-analysis of 14 studies, the sensitivity and specificity of CTA for diagnosing LGI bleeding were 90% and 92%, respectively.33 As with many meta-analyses, the studies included were heterogeneous and varied as to whether adjunctive studies such as colonoscopy were performed. Nonetheless, the overall conclusion from this study, as well as others, is that in patients with suspected active hemorrhage, CTA should be performed.6–8 It should be recognized that in the small subset of patients in hemorrhagic shock requiring active transfusion to maintain blood pressure that CTA is not therapeutic and will delay visceral angiography and the time to hemorrhage control. It will also result in a second contrast load. This concept is analogous to those trauma patients with active hemorrhage where any intervention that is not aimed at hemorrhage control (eg, CT scanning) will increase mortality. As most LGI bleeding is both intermittent and self-limited, even if patients require transfusion or have had initial signs of hypoperfusion, the overall rate of positive CTA can be low. In a retrospective study of patients undergoing CTA for LGI bleeding, nearly 80% of patients with an initial negative CTA had no further clinical or radiologic evidence of rebleeding and never required endoscopic or radiologic intervention.34 Another retrospective study of 854 patients presenting with LGI bleeding found that only 20% of CTA scans were positive. Factors associated with a positive CTA included recent bowel resection or endoscopic intervention, transfusion of more than 3 units of packed red blood cells per day, use of antiplatelet agents, and hemodynamic instability.35 In a large, multicenter Japanese cohort of patients admitted with LGI bleeding, a CTA ≤1 hour from presentation was performed in 97.5%, with 22.0% demonstrating contrast extravasation on CT.36 Visceral Angiography Patients with active hemorrhage identified on CTA should undergo catheter-based visceral angiography.6–8,37 Due to the intermittent nature of LGI bleeding in general, and especially diverticular bleeding, patients with a positive CTA often have a short window where interventional radiology and angioembolization are beneficial. In a single-center retrospective study, patients who underwent angiography within 90 minutes after a positive CTA were 8.5 times more likely to have extravasation detected.38 Thus, like the trauma patient, to achieve a diagnosis and hemostasis, the time from consult to needle stick should be short, ideally <1 hour. Any delays, such as insistence on performing a nuclear scintigraphy scan or CTA, decrease the efficacy of the angiography to identify the source of the hemorrhage.39 A systematic review found that selective angiographic embolization achieved immediate hemostasis in 40% to 100% of cases of diverticular bleeding, with rebleeding rates ranging from 0% to 50%.40 The most significant complication of angioembolization is bowel ischemia with a reported incidence of 1%–4%. This is commonly attributed to the failure to achieve super-selective catheterization and embolization.41,42 In patients who are hemodynamic stable without evidence of ongoing bleeding and who cannot meet these time constraints, the next best step would be urgent colonoscopy. Colonoscopy Patients in whom CTA has failed to identify a source have most likely stopped actively bleeding and have or will soon attain hemodynamic stability. These patients should be admitted to an appropriately monitored setting, and it is this group of patients that colonoscopy has been recommended as the first-line procedure.6–8,12,30 The advantages of colonoscopy include the potential to make a diagnosis, the application of therapeutic modalities to control hemorrhage, the ability to obtain pathologic specimens, and the ability to tattoo areas of pathology for subsequent surgical resection. The reported accuracy of colonoscopy for LGI bleeding varies widely from 42% to 100% and is likely due to the heterogeneous nature of the patient populations studied.7,8,26,43 Of these patients, hemostatic therapy is performed in 10% to 63% of patients.6–8 The optimal timing for performing a colonoscopy for acute LGI has been investigated. Only one randomized prospective study has compared urgent (<12 h) versus elective (36–60 h) colonoscopy in this group of patients.44 There were no outcome differences between the groups. However, the study had serious limitations due to the small sample size. A propensity-matched series from Japan found that the early colonoscopy group had a higher rate of therapeutic interventions, fewer transfusions, and a shorter hospital stay. Mortality and rebleeding rates were unchanged.36 Additional studies have also reached similar conclusions regarding mortality.12,36,45,46 Navaneethan and colleagues, using the 2010 National Inpatient Sample dataset analyzed 58,296 patients admitted with LGI bleeding in the United States.46 Multivariate analysis demonstrated that early (<24 h) colonoscopy reduced length of stay, transfusion requirements, and costs, but had no effect on the percentage of patients undergoing endoscopic therapy or mortality. A recent meta-analysis did not demonstrate any significant difference between early and delayed colonoscopy on rebleeding rates or transfusion, but colonoscopy performed within 24 hours was found to significantly increase the diagnostic yield, the percentage of patients undergoing therapeutic interventions, and reduce the hospital length of stay.47 Colonoscopy does require time and preparation, which can result in significant fluid shifts in an elderly population or those who have been recently resuscitated. The optimal colonoscopy bowel preparation has yet to be identified. A retrospective study with or in patients undergoing colonoscopy for acute LGI bleeding found higher diagnostic and reduced the for colonoscopy in the A retrospective study examining the complication rate of bowel preparation in acute versus elective colonoscopy found no significant There is a recent using lower to patient direct are Thus, from the data and current it is recommended that patients admitted with acute LGI bleeding undergo bowel preparation and urgent colonoscopy within 24 h) Surgery with in both endoscopic and interventional radiologic has a very in the acute management of LGI This is especially if no bleeding source has been identified. As almost all LGI bleeding is and with only a of patients massive and hemorrhage. Patients who underwent visceral angiography without the of a source of bleeding should be for urgent colonoscopy. patients who after colonoscopy should be for visceral angiography or a In a retrospective study from the only of patients required surgical In a retrospective cohort of patients at hospitals in admitted for acute only underwent for Of that group, only 90 of the surgical were for A small retrospective study compared patients with a positive CT with versus surgical The study was further by significant differences in patient for and less than those by this patients undergoing surgical intervention had increased complication rates compared with those undergoing In a small number of patients who are to control of hemorrhage, should be to identify the of the bleeding to the of surgical resection. There is no current data on the of surgical to identify and control However, given that these patients are likely to have is likely to be a If no source be or careful attention should be given to the small as it may be the of the hemorrhage. In these can be performed to both and the without requiring It should also be that in cases of a it is not to blood that has in the distal small and the should not be that it is from a small bowel source. If the rectum has been definitively an emergent with is the procedure of Small bowel hemorrhage accounts for to 10% of all gastrointestinal The pathology of small bowel hemorrhage from that of the rectum, and The etiology of bleeding varies with patient In patients and bowel are and in patients it is are observed at all for to 40% of bleeding with the most These may also be by In the making the diagnosis of bleeding has been a both patients and clinicians. In the it was often gastrointestinal bleeding due to the to make a In patients presenting with shock or hypoperfusion and active hemorrhage, visceral angiography or CTA will often identify the source of bleeding from the small red blood cell studies may be more the small bowel as an etiology of the bleeding in this subset as it requires less active hemorrhage the by the Society of and the American College of Gastroenterology is to consider a small bowel source in any patient upper and lower endoscopic studies were Patients should undergo The rate in bleeding from to Compared with has been in prospective studies to be significantly to The also recommends in patients with suspected bleeding as soon as possible after the bleeding ideally within to the diagnostic and subsequent therapeutic yield. identified by can be by or the the and cardiovascular comorbidities in many patients presenting with LGI bleeding, and antiplatelet use is While a of this is the of this it would be not to include recommendations and As with all patients presenting with bleeding on anticoagulants, there is a between the indication for the and the degree of hemorrhage. In most patients, an urgent should be to help identify the optimal management As most patients presenting with LGI bleeding have stopped by the time present to the emergency the decision regarding the management of anticoagulation can be in a more with its is a and has been associated with If coagulation therapeutic or patients should be with a and with in those at high risk (eg, should not be to 14 after the hemorrhage has been or there is no further evidence of oral anticoagulants have a short and most do not require specific should be considered in a small of patients with hemorrhage. require a more and their use to be against the and of the anticoagulation in used for should be patients on dual antiplatelet therapy, the current is to and the other Patients who LGI bleeding while dual antiplatelet therapy have a increase in This must be against the risk of to death or which is a A cohort study compared patients who after admission with LGI bleeding versus those who had it The latter group had fewer rebleeding but significantly more cardiovascular and