Printed on 3/17/2026
For informational purposes only. This is not medical advice.
The Rockall scoring system predicts rebleeding and mortality after upper gastrointestinal hemorrhage. It includes pre-endoscopy variables (age, shock, comorbidity) and post-endoscopy findings (diagnosis, stigmata of recent hemorrhage). Scores range 0–11 and guide disposition and treatment intensity. Use [Glasgow-Blatchford Score](/tools/glasgow-blatchford) for complementary pre-endoscopy risk stratification. In cirrhotic patients, assess hepatic reserve with [Child-Pugh Score](/tools/child-pugh-score) and [MELD Score](/tools/meld-score) — liver disease comorbidity adds 2 points to Rockall. For ICU-level patients, assess severity with [APACHE II](/tools/apache-ii-score) and [SOFA Score](/tools/sofa-score).
Formula: Age (0–2) + Shock (0–2) + Comorbidity (0–3) + Diagnosis (0–2) + Stigmata (0–2). Total 0–11.
In the emergency department, immediately calculate the pre-endoscopy Rockall using three variables: age (0-2 points), hemodynamic status (0-2 points for no shock, tachycardia, or hypotension), and comorbidity (0-3 points for none, major cardiac/CHF/IHD, or renal/liver failure/malignancy). This pre-endoscopy score (maximum 7 points) helps guide initial decision-making. A score of 0-1 suggests very low risk and may allow outpatient endoscopy or early discharge in select patients. Scores ≥3 indicate moderate-to-high risk requiring admission and urgent endoscopy. Document the pre-endoscopy score while awaiting endoscopy, as it provides early prognostic information.
After endoscopy, add the two endoscopic variables: diagnosis (0-2 points for Mallory-Weiss/no lesion, all other diagnoses, or malignancy) and stigmata of recent hemorrhage (0-2 points for none/dark spot, or active bleeding/clot/visible vessel). The complete Rockall score ranges from 0-11 and provides the most accurate prediction of rebleeding and mortality. Stigmata scoring: 0 points for clean-based ulcer or flat pigmented spot; 2 points for active arterial bleeding, visible vessel (non-bleeding), adherent clot, or oozing without visible vessel. The full score guides post-endoscopy management decisions including ICU vs. floor admission, timing of repeat endoscopy if rebleeding occurs, and safe discharge criteria.
Use the completed Rockall score to stratify rebleeding and mortality risk: Score 0-2 (low risk): <0.2% mortality, <5% rebleeding — consider early discharge (24-48 hours) with PPI and outpatient follow-up. Score 3-4 (moderate risk): 5-7% mortality, 10-15% rebleeding — standard ward admission, continue PPI infusion for 72 hours, monitor hemoglobin, repeat endoscopy if clinically indicated. Score 5-7 (high risk): 12-17% mortality, 20-25% rebleeding — consider ICU monitoring, prolonged PPI therapy, low threshold for repeat endoscopy or interventional radiology consultation. Score ≥8 (very high risk): >35% mortality — ICU admission, aggressive resuscitation, multidisciplinary team involvement (GI, surgery, IR), discussion about goals of care. Document the score and management plan clearly in the medical record.
Emergency physicians, gastroenterologists
Calculate pre-endoscopy Rockall in the ED to determine urgency of endoscopy and level of care. Score 0-1 may allow outpatient endoscopy in carefully selected patients (no active hematemesis, stable vitals, reliable follow-up). Score ≥3 requires admission. This early risk stratification guides resource allocation and helps prioritize patients for urgent vs. semi-urgent endoscopy within the first 24 hours. Most guidelines recommend endoscopy within 24 hours for all upper GI bleeds, but Rockall helps identify those who need it within 6-12 hours (high scores) vs. those who can wait safely.
Gastroenterologists, hospitalists, ICU physicians
After endoscopy, the complete Rockall score determines whether the patient requires ICU-level monitoring or can be managed on a medical floor. Score ≥5-7 often warrants ICU admission for close hemodynamic monitoring, especially if high-risk stigmata (active bleeding, visible vessel) were seen. Lower scores allow floor-level care. The score also guides PPI therapy duration (72-hour infusion for high-risk vs. oral PPI for low-risk) and helps predict which patients will require repeat endoscopy or interventional radiology embolization if rebleeding occurs.
Hospitalists, gastroenterology consult services
Rockall score helps determine when it's safe to discharge a patient after upper GI bleed. Score 0-2 with stable hemoglobin (no drop for 24 hours), tolerating oral intake, and no signs of rebleeding can typically be discharged after 24-48 hours on oral PPI with outpatient GI follow-up in 1-2 weeks. Higher scores require longer observation (typically 3-5 days for score 3-4, or longer for ≥5). The score provides objective data to support discharge decisions and helps standardize practice across providers.
Gastroenterologists, hospitalists, pharmacists
Rockall score informs PPI therapy intensity. High-risk patients (score ≥5, especially with high-risk stigmata) should receive IV PPI infusion (80 mg bolus then 8 mg/hr × 72 hours) per international guidelines. Lower-risk patients can receive oral PPI twice daily. The score helps avoid over-treatment of low-risk patients (unnecessary IV PPI) and under-treatment of high-risk patients. After discharge, all patients typically continue oral PPI therapy, with duration depending on the underlying diagnosis (e.g., 4-8 weeks for peptic ulcer).
GI quality officers, hospital administrators
Use Rockall score to audit upper GI bleed management, track outcomes, and compare performance against published benchmarks. Analyze whether actual rebleeding and mortality rates match predicted rates for each Rockall category. Identify opportunities for improvement: Are high-risk patients receiving appropriate ICU care? Are low-risk patients being unnecessarily admitted? Are endoscopy timelines meeting guidelines? Rockall provides standardized risk stratification for case-mix adjustment when comparing outcomes across institutions or time periods.
Medical students, residents, fellows
Rockall score teaches systematic approach to GI bleeding risk assessment. It demonstrates how to integrate clinical data (age, vitals, comorbidity) with endoscopic findings (diagnosis, stigmata) to predict outcomes. Understanding Rockall reinforces key concepts: hemodynamic instability predicts worse outcomes, comorbidity matters, endoscopic stigmata guide management, and malignancy portends poor prognosis. Use it on rounds to teach evidence-based prognostication and demonstrate how scoring systems translate into clinical action.
A pre-endoscopy Rockall of 0-1 identifies very-low-risk patients who may be suitable for outpatient endoscopy in carefully selected cases. Criteria: age <60, no shock, no major comorbidity, hemodynamically stable, no active hematemesis, reliable for follow-up. However, most guidelines still recommend admission for all upper GI bleeds, so use this approach only in low-resource settings or when patients strongly prefer outpatient care. Glasgow-Blatchford score of 0 is actually better for identifying patients safe for discharge.
Score shock based on hemodynamic status: 0 points = SBP ≥100 and HR <100. 1 point = tachycardia (HR ≥100) with SBP ≥100. 2 points = hypotension (SBP <100) regardless of heart rate. Use the worst vitals from the ED presentation, not after resuscitation. Document clearly: 'Presented with SBP 85, HR 115 (shock score 2). After 2L crystalloid, SBP 110 (but still score 2 for Rockall).' Pre-resuscitation vitals reflect true physiologic compromise.
Score comorbidity: 0 = none. 2 = CHF, IHD, or any major comorbidity. 3 = renal failure, liver failure, or disseminated malignancy. 'Major comorbidity' (2 points) includes things like COPD on home O2, recent stroke, active malignancy without metastases. The jump to 3 points is reserved for end-organ failure or widespread cancer. If unclear, be consistent: document your rationale ('Scoring 2 for severe COPD with recent exacerbation'). This subjectivity is a known limitation of Rockall.
Score stigmata as 2 points for: active arterial spurting, active oozing, non-bleeding visible vessel, or adherent clot. Score 0 for clean-based ulcer or flat pigmented spot. The 'visible vessel' (bluish bulge in ulcer base) is high-risk even without active bleeding — 50% will rebleed without intervention. Adherent clot that cannot be washed off is scored 2 points; if the clot is washed off revealing a clean base, score 0. Endoscopist documentation is key: 'Duodenal ulcer with visible vessel, thermal therapy applied (stigmata score 2).'
Mallory-Weiss tears (esophageal mucosal lacerations from retching/vomiting) have excellent prognosis regardless of appearance, so diagnosis trumps stigmata. Score diagnosis as 0 for Mallory-Weiss even if it's oozing. This acknowledges that Mallory-Weiss rarely rebleeds and almost never causes death. Similarly, 'no lesion seen' (sometimes after brisk bleed that self-resolved) scores 0 for diagnosis. In contrast, gastric or esophageal cancer scores 2 for diagnosis due to poor prognosis.
Glasgow-Blatchford (GBS) is better for ruling out intervention need (GBS 0 = safe discharge). Rockall is better for predicting rebleeding and mortality after endoscopy. Ideal workflow: calculate GBS in ED → if 0, consider discharge (rare, requires institutional protocol). If GBS ≥1, admit and do endoscopy → then calculate full Rockall post-endoscopy to guide monitoring level, treatment intensity, and discharge timing. Don't rely on one score alone; they provide complementary information.
Rockall was derived and validated for non-variceal upper GI bleeding (peptic ulcers, Mallory-Weiss, erosive gastritis, etc.). Variceal hemorrhage from portal hypertension has different risk factors and management (band ligation, octreotide, antibiotics, TIPS). If varices are found, use Child-Pugh or MELD score for prognostication, not Rockall. Document: 'Esophageal varices with active bleeding. Rockall not applicable. Managing per variceal bleeding protocol with Child-Pugh C cirrhosis.'
Write clearly in your note: 'Pre-endoscopy Rockall 3 (age 1, shock 0, comorbidity 2) → admitted for urgent endoscopy. Post-endoscopy Rockall 6 (pre-endo 3 + diagnosis 1 [gastric ulcer] + stigmata 2 [visible vessel]). 12-17% mortality risk. 20-25% rebleeding risk. Plan: ICU monitoring, IV PPI infusion 72hr, repeat endoscopy if rebleeds.' This documentation communicates your risk assessment and justifies your management decisions to consultants and follow-up providers.
Age scoring: 0 points for <60, 1 point for 60-79, 2 points for ≥80. While younger patients generally do better, don't ignore other high-risk features in patients <60. A 55-year-old with metastatic cancer and visible vessel (score potentially 5-6) still has significant risk. Age is just one component. Conversely, an 85-year-old with Mallory-Weiss and no stigmata (score 2) may do fine despite age. Don't let age alone drive your entire risk assessment.
If a patient initially stable (shock 0) develops hypotension 12 hours later, or if repeat endoscopy shows new high-risk stigmata, recalculate Rockall. The score is a snapshot, not static. Document: 'Initial Rockall 3. Developed hematemesis and hypotension 18 hours post-admission (new shock score 2). Updated Rockall 5. Escalating to ICU.' Similarly, if a patient improves dramatically (stable vitals, no rebleeding, no stigmata on repeat endoscopy if done), this lowers their risk trajectory even if the original score remains documented.
Your Rockall score estimates the risk of rebleeding and mortality following an upper gastrointestinal hemorrhage. A score of 0–2 indicates low risk, with a predicted mortality of approximately 0.1% and rebleeding risk under 5%. Scores of 3–4 represent moderate risk with mortality around 5–7%. Scores of 5–7 carry a high risk with mortality of 12–17% and significant rebleeding probability. Scores of 8 or above indicate very high risk with mortality exceeding 35%.
The complete Rockall score incorporates both pre-endoscopy variables (age, hemodynamic status, comorbidities) and post-endoscopy findings (diagnosis and stigmata of recent hemorrhage). This makes it most useful after endoscopy for refining risk stratification and guiding decisions about monitoring intensity, repeat endoscopy, and discharge timing.
Use the Rockall score after endoscopy in patients presenting with upper GI hemorrhage to predict the risk of rebleeding and mortality. The pre-endoscopy Rockall (using only age, shock, and comorbidity, maximum 7 points) can be calculated in the emergency department for initial triage — a pre-endoscopy score of 0–1 identifies patients who may be suitable for early discharge or outpatient endoscopy.
The full Rockall score is most valuable in the post-endoscopy setting for determining the level of care required (ICU vs. floor), the need for repeat endoscopy, and the timing of discharge. It is widely used in the United Kingdom and internationally and is endorsed by multiple gastroenterology society guidelines for risk stratification of non-variceal upper GI bleeding.
The Rockall score was derived and validated primarily in populations with non-variceal upper GI bleeding and may not perform as well in patients with variceal hemorrhage, who have different risk profiles and treatment algorithms. The comorbidity scoring is broad (major cardiac disease, renal failure, liver failure, metastatic cancer) and does not account for the severity gradient within each category.
Compared to the Glasgow-Blatchford score, the Rockall score is less effective at identifying very low-risk patients who can be safely discharged without endoscopy. The Glasgow-Blatchford score of 0 has superior negative predictive value for intervention need. Additionally, the endoscopic component introduces inter-observer variability in assessing stigmata of recent hemorrhage, and the score was developed before the widespread use of proton pump inhibitor infusions and modern endoscopic hemostasis techniques, which have changed the natural history of upper GI bleeding.
Disclaimer: This tool is for educational and informational purposes only. It is not a substitute for professional medical advice, diagnosis, or treatment. Always consult a qualified healthcare provider with questions about your health.
Calculate the Glasgow-Blatchford score to identify low-risk upper GI bleed patients. Score 0–1: safe for outpatient endoscopy without admission. Higher scores guide urgency and ICU level of care.
ClinicalCalculate the Child-Pugh score to classify the severity of chronic liver disease and estimate prognosis. Uses bilirubin, albumin, INR, ascites, and encephalopathy.
ClinicalCalculate the MELD and MELD-Na scores to assess liver disease severity and transplant priority. Uses bilirubin, INR, creatinine, and sodium.
EmergencyCalculate the Shock Index (HR/SBP ratio) for rapid hemodynamic assessment. Normal: 0.5–0.7. Score ≥1.0 indicates hemodynamic compromise; ≥1.4 indicates severe shock requiring immediate intervention.