Benefits in NNT
No one was helped (no further bleeding, death, or hemostatic intervention was prevented)
Early versus Delayed Colonoscopy for Acute Lower Gastrointestinal Bleeding
Early colonoscopy within 24 h of presentation does not improve clinical outcomes such as further bleeding or mortality in patients hospitalized with acute LGIB
Harms in NNT
No one was harmed (no difference in harms)
View As:
Source
Shobowale O, Youssef E. Early versus delayed colonoscopy for acute lower gastrointestinal bleeding. Academic Emergency Medicine. Published online May 7, 2025:acem.70061.
Study Population: 4 trials comprising 463 adult patients presenting with LGIB or hematochezia, randomized to early colonoscopy (within 24 h) or elective colonoscopy (beyond 24 h)
Study Population: 4 trials comprising 463 adult patients presenting with LGIB or hematochezia, randomized to early colonoscopy (within 24 h) or elective colonoscopy (beyond 24 h)
Efficacy Endpoints
Rate of further bleeding, mortality, diagnostic yield, endoscopic intervention, or hemostatic intervention
Harm Endpoints
Adverse events
Narrative
Acute lower gastrointestinal bleeding (LGIB) is one of the most common indications for hospital admission due to digestive disorders. Incidence has been increasing relative to upper gastrointestinal bleeding over the past decade, with an estimated annual incidence of about 35–51 per 100,000 persons.1, 2 The current guidelines recommend colonoscopic evaluation for nearly all patients presenting with acute LGIB with the goal of identifying the source of bleeding and achieving hemostasis.2 However, the optimal time to perform this procedure is not clear. Here we summarize the data compiled from a systematic review and meta-analysis detailing if early colonoscopy improves outcomes for patients presenting with LGIB.3
The systematic review and meta-analysis discussed here included four randomized control trials comprising of 463 participants. The studies included adult patients presenting with LGIB, or hematochezia, that were randomized into either early colonoscopy (within 24 h) or elective colonoscopy (beyond 24 h). The study population consisted mainly of participants that presented to the emergency department (ED) and subsequently admitted for inpatient units. The trials were conducted in Japan (one multicenter trial),4 the United States (two trials),5, 6 and the Netherlands (one trial).7 Only one study included an outpatient cohort who presented with moderate to severe hematochezia.4 The trials enrolled both hemodynamically stable and unstable patients. However, one study excluded patients requiring blood product transfusion.4 Hemodynamic instability was reported in 83%,5 64%,6 26% (with ongoing bleeding in 78%),7 and 3% (with 33% having “severe bleeding,” defined as hemoglobin ≤8 g/dL, systolic blood pressure <90 mmHg, or heart rate >100 beats/min).4
The primary outcome of the systematic review was further bleeding, defined as persistent or recurrent bleeding after initial colonoscopy or other initial diagnostics. Secondary outcomes included mortality, diagnostic yield, stigmata of recent hemorrhage, red blood cell transfusion, length of hospitalization, endoscopic intervention, hemostatic intervention, procedural intervention, and adverse events/complications. Endoscopic intervention is defined as any hemostasis performed at time of initial colonoscopy. Primary hemostatic intervention was defined as any hemostasis performed endoscopically, surgically, or by interventional radiology. Procedural intervention (surgery or interventional radiology) was defined as any participants needing further intervention after initial hemostatic intervention. Follow-up periods were at 1 month,4, 7 a mean of 62 and 58 months for the urgent colonoscopy versus standard therapy group,6 and throughout the hospital stay (mean hospital days 5.2 vs. 4.8 in urgent vs. elective colonoscopy group).5
Compared to early colonoscopy, further bleeding was not significantly different among patients receiving elective colonoscopy. There was no significant statistical difference in secondary outcomes of mortality, diagnostic yield, endoscopic intervention, or any primary hemostatic intervention. The data could not be pooled for length of hospital stay and transfusion requirement due to paucity of data. Adverse events were relatively uncommon and similar in the two groups.
The systematic review and meta-analysis discussed here included four randomized control trials comprising of 463 participants. The studies included adult patients presenting with LGIB, or hematochezia, that were randomized into either early colonoscopy (within 24 h) or elective colonoscopy (beyond 24 h). The study population consisted mainly of participants that presented to the emergency department (ED) and subsequently admitted for inpatient units. The trials were conducted in Japan (one multicenter trial),4 the United States (two trials),5, 6 and the Netherlands (one trial).7 Only one study included an outpatient cohort who presented with moderate to severe hematochezia.4 The trials enrolled both hemodynamically stable and unstable patients. However, one study excluded patients requiring blood product transfusion.4 Hemodynamic instability was reported in 83%,5 64%,6 26% (with ongoing bleeding in 78%),7 and 3% (with 33% having “severe bleeding,” defined as hemoglobin ≤8 g/dL, systolic blood pressure <90 mmHg, or heart rate >100 beats/min).4
The primary outcome of the systematic review was further bleeding, defined as persistent or recurrent bleeding after initial colonoscopy or other initial diagnostics. Secondary outcomes included mortality, diagnostic yield, stigmata of recent hemorrhage, red blood cell transfusion, length of hospitalization, endoscopic intervention, hemostatic intervention, procedural intervention, and adverse events/complications. Endoscopic intervention is defined as any hemostasis performed at time of initial colonoscopy. Primary hemostatic intervention was defined as any hemostasis performed endoscopically, surgically, or by interventional radiology. Procedural intervention (surgery or interventional radiology) was defined as any participants needing further intervention after initial hemostatic intervention. Follow-up periods were at 1 month,4, 7 a mean of 62 and 58 months for the urgent colonoscopy versus standard therapy group,6 and throughout the hospital stay (mean hospital days 5.2 vs. 4.8 in urgent vs. elective colonoscopy group).5
Compared to early colonoscopy, further bleeding was not significantly different among patients receiving elective colonoscopy. There was no significant statistical difference in secondary outcomes of mortality, diagnostic yield, endoscopic intervention, or any primary hemostatic intervention. The data could not be pooled for length of hospital stay and transfusion requirement due to paucity of data. Adverse events were relatively uncommon and similar in the two groups.
Caveats
This systematic review and meta-analysis has several limitations. First, the low event rates for outcomes led to wide confidence intervals and reduced the precision of effect size estimates. Additionally, because the review included only randomized controlled trials, the number of studies was limited. However, this also contributed to a moderate-to-high quality of evidence. Blinding was not feasible for the intervention comparing early versus elective colonoscopy. Moreover, heterogeneity was observed among the included trials for both the primary outcome and the secondary outcome of endoscopic intervention. This variability was largely influenced by one trial in which most participants in the control group underwent imaging studies before elective colonoscopy.
The included trials had variations in patient populations and clinical settings, which can introduce bias and affect the generalizability of the findings. In one study, patients in the elective colonoscopy would have received endovascular hemostasis prior to colonoscopy if appropriate.6 In another, patients first underwent urgent upper endoscopy with colonoscopy only performed on those who had a negative endoscopy,5 significantly altering the patient characteristics in that study versus the others. Differences in preendoscopy resuscitation, medications administered, and endoscopic techniques across studies could influence outcomes, making it difficult to isolate the effect of early colonoscopy. The systematic review did not adequately explore subgroup effects (e.g., severity of bleeding, comorbid conditions). Therefore, it may have overlooked patient populations that could still benefit from early colonoscopy.
The systematic review also does not provide complete clarity on this aspect of hemodynamic stability. It appears that most patients were hemodynamically stable, but there was still a substantial number of hemodynamically unstable patients (35%) enrolled in the trials. As a result, the findings of this review may not be applicable to high-risk or hemodynamically unstable patients, for whom different management strategies may be required. Further investigation with studies consisting of only hemodynamically unstable or other high-risk features is needed. The incidences of adverse events and mortality were evaluated at 30 days,4, 7 throughout the hospital stay (mean of 5 days),5 and at 60 months.6 While some of these time frames may be relevant for acute outcomes, they may not capture longer-term complications, which could influence the overall benefits or harms.
Most existing expert society guidelines agree with the findings of this meta-analysis. The American Academy of Gastroenterology recommends that patients presenting with LGIB undergo a nonurgent colonoscopy.2 The British Society of Gastroenterology recommends that patients with major LGIB bleeding should be admitted to hospital for colonoscopy on the next availability but do not specify an exact time frame due to the data that exists showing no harm in early colonoscopy.8
Based on the available data, early colonoscopy does not appear to provide a significant difference over delayed colonoscopy in patients presenting with acute LGIB. Therefore, we have assigned an NNT color recommendation of red (benefits and harms may be equal or equivocal) to this intervention. Prolonged hospital stays solely for the purpose of awaiting colonoscopy should be avoided. However, this recommendation applies only to hemodynamically stable and low-risk patients. In contrast, for patients with hemodynamic instability, urgent endoscopy may still be warranted. While colonoscopy does not occur in the setting of the ED, the findings of this evidence-based medicine summary may affect management. Given that there was no difference in outcomes between early and delayed colonoscopy, emergency medicine providers may feel reassured that they do not need to gear practice toward preparing patients for possible emergent colonoscopy. No emergent changes need to be made toward diet, NPO status, adjustments to diabetes medications, or bowel preparation as this has been proven to not be an emergent procedure and can be left to the discretion of the admission team.
The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.
The included trials had variations in patient populations and clinical settings, which can introduce bias and affect the generalizability of the findings. In one study, patients in the elective colonoscopy would have received endovascular hemostasis prior to colonoscopy if appropriate.6 In another, patients first underwent urgent upper endoscopy with colonoscopy only performed on those who had a negative endoscopy,5 significantly altering the patient characteristics in that study versus the others. Differences in preendoscopy resuscitation, medications administered, and endoscopic techniques across studies could influence outcomes, making it difficult to isolate the effect of early colonoscopy. The systematic review did not adequately explore subgroup effects (e.g., severity of bleeding, comorbid conditions). Therefore, it may have overlooked patient populations that could still benefit from early colonoscopy.
The systematic review also does not provide complete clarity on this aspect of hemodynamic stability. It appears that most patients were hemodynamically stable, but there was still a substantial number of hemodynamically unstable patients (35%) enrolled in the trials. As a result, the findings of this review may not be applicable to high-risk or hemodynamically unstable patients, for whom different management strategies may be required. Further investigation with studies consisting of only hemodynamically unstable or other high-risk features is needed. The incidences of adverse events and mortality were evaluated at 30 days,4, 7 throughout the hospital stay (mean of 5 days),5 and at 60 months.6 While some of these time frames may be relevant for acute outcomes, they may not capture longer-term complications, which could influence the overall benefits or harms.
Most existing expert society guidelines agree with the findings of this meta-analysis. The American Academy of Gastroenterology recommends that patients presenting with LGIB undergo a nonurgent colonoscopy.2 The British Society of Gastroenterology recommends that patients with major LGIB bleeding should be admitted to hospital for colonoscopy on the next availability but do not specify an exact time frame due to the data that exists showing no harm in early colonoscopy.8
Based on the available data, early colonoscopy does not appear to provide a significant difference over delayed colonoscopy in patients presenting with acute LGIB. Therefore, we have assigned an NNT color recommendation of red (benefits and harms may be equal or equivocal) to this intervention. Prolonged hospital stays solely for the purpose of awaiting colonoscopy should be avoided. However, this recommendation applies only to hemodynamically stable and low-risk patients. In contrast, for patients with hemodynamic instability, urgent endoscopy may still be warranted. While colonoscopy does not occur in the setting of the ED, the findings of this evidence-based medicine summary may affect management. Given that there was no difference in outcomes between early and delayed colonoscopy, emergency medicine providers may feel reassured that they do not need to gear practice toward preparing patients for possible emergent colonoscopy. No emergent changes need to be made toward diet, NPO status, adjustments to diabetes medications, or bowel preparation as this has been proven to not be an emergent procedure and can be left to the discretion of the admission team.
The original manuscript was published in Academic Emergency Medicine as part of the partnership between TheNNT.com and AEM.
Published/Updated
August 19, 2025