- Pick the mode. Clinical (pre-endoscopy) Rockall uses three items at admission (age, shock, comorbidity). Full Rockall adds two endoscopic items (diagnosis, stigmata of recent hemorrhage) for a complete rebleed and mortality estimate.
- Choose the shock category from the patient's admission vitals: no shock (systolic BP ≥ 100 and pulse < 100), tachycardia (SBP ≥ 100 with pulse ≥ 100), or hypotension (SBP < 100).
- Select the worst comorbidity present.
- In full-Rockall mode, pick the endoscopic diagnosis and the most severe stigma of recent hemorrhage.
- Read the total score and its rebleed and mortality estimates.
All computation runs in your browser; no values are stored or transmitted.
When to Use
Use the Rockall Score in an adult with an acute upper gastrointestinal (GI) bleed to estimate the probability of rebleed and in-hospital mortality. The pre-endoscopy (clinical) version supports the initial triage decision; the full version, calculated after endoscopy, refines the disposition — including the decision to discharge a low-score patient early or to monitor a high-score patient in a step-down setting.
Appropriate population
Adults with acute upper GI bleeding (hematemesis, melena, coffee-ground emesis). The clinical Rockall can be used at admission from the age, admission vitals, and comorbidities alone; the full Rockall is used once the endoscopy report is available and the source of bleeding has been seen. Score 0 (clinical) or ≤ 2 (full) marks a low-risk group that may be considered for early discharge.
When NOT to rely on it
The pre-endoscopy Rockall is less accurate than the Glasgow–Blatchford Score at identifying very-low-risk patients at first contact — do not use it as the sole basis for early discharge before endoscopy. Rockall is also not validated in lower GI bleeding, in children, or in rebleed scored against the same endoscopy; it was derived for the first episode in an unselected emergency cohort.
Pearls & Pitfalls
Clinical and full Rockall answer different questions
The clinical (pre-endoscopy) Rockall is a pre-procedure triage tool — it uses admission vitals only. The full Rockall is a post-procedure prognostic tool — it adds what endoscopy saw. A clinical Rockall 0 roughly identifies patients who will not rebleed; a full Rockall ≤ 2 identifies patients safe for early discharge after endoscopy.
Pair with the Glasgow–Blatchford at first contact
Modern UGIB pathways use the Glasgow–Blatchford Score at the door (better at picking out very-low-risk patients) and the full Rockall after endoscopy (better at quantifying rebleed and mortality risk once the source is known). The two scores answer different questions and complement each other.
Pitfalls
(1) Shock category depends on honest vitals: a patient on a beta-blocker or vasopressor can look non-shocked despite major blood loss. (2) Comorbidity is cumulative in the derivation: the worst band present scores; do not add across diseases. (3) The "all other diagnoses" diagnosis band is a catch-all — a benign erosion scores the same +1 as a bleeding ulcer; the stigmata item is where therapeutic risk is captured. (4) Rockall was derived and validated for the first presentation; it is not a rebleed score for a second endoscopy.
Why Use It
Acute upper GI bleed carries meaningful short-term mortality, driven mostly by the patient (age, comorbidity, shock) rather than by the lesion seen at endoscopy. Rockall and colleagues quantified that split: three admission items predict most of the rebleed and death signal; adding endoscopic findings sharpens the estimate. The practical use is prognostic triage — who needs a monitored bed, who can be stepped down, and who can be safely discharged after endoscopy. The score remains in international guidelines and is taught alongside the Glasgow–Blatchford Score as the pair of tools that frame admission and post-endoscopy decisions in upper GI bleeding.
Rockall Score — Pre-Endoscopy and Full Rockall
Pick the mode. The Clinical (pre-endoscopy) Rockall uses three admission items (max 7). The Full Rockall adds two endoscopic items (max 11) and estimates rebleed and in-hospital mortality from the original derivation.
⚕ Rockall TA et al. Gut. 1996;38(3):316–321. The score estimates rebleed and in-hospital mortality after a first episode of acute upper gastrointestinal (GI) hemorrhage. Confidence intervals widen at the extremes; the pre-endoscopy Rockall under-identifies very-low-risk patients compared with the Glasgow–Blatchford Score. For licensed clinicians; not a substitute for individualized assessment.
Next Steps
Match the Rockall band to the admission trajectory and the endoscopic findings.
- Clinical Rockall 0 — low pre-endoscopy risk. Admit for endoscopy; if the full Rockall after endoscopy remains ≤ 2, consider early discharge.
- Full Rockall ≤ 2 — low risk (rebleed ~ 4%, mortality ~ 0–0.2%). Early discharge with outpatient follow-up is reasonable when hemoglobin is stable and the lesion has been treated.
- Full Rockall 3–5 — intermediate risk (rebleed 11–24%, mortality 3–11%). Monitor on a general ward, keep the endoscopy team informed, and keep IV PPI going.
- Full Rockall ≥ 6 — high risk (rebleed 33–44%, mortality 17–41% in derivation). Step-down or ICU-level monitoring, early reassessment, ready for repeat endoscopy or surgical or radiological rescue.
- Pair with the Glasgow–Blatchford Score at first contact and the Shock Index at the bedside; use the Charlson Comorbidity Index to add longer-term context.
Evidence & References
Scoring
| Item | Category | Points |
|---|---|---|
| Age | < 60 | 0 |
| 60–79 | +1 | |
| ≥ 80 | +2 | |
| Shock | SBP ≥ 100, pulse < 100 | 0 |
| SBP ≥ 100, pulse ≥ 100 (tachycardia) | +1 | |
| SBP < 100 (hypotension) | +2 | |
| Comorbidity (worst) | None | 0 |
| CHF / IHD / major comorbidity | +2 | |
| Renal failure, liver failure, metastatic cancer | +3 | |
| Diagnosis (full only) | Mallory–Weiss / no lesion / no stigmata | 0 |
| All other diagnoses | +1 | |
| Upper GI malignancy | +2 | |
| Major stigmata (full only) | None or dark spot only | 0 |
| Blood in upper GI tract, adherent clot, visible or spurting vessel | +2 |
Full Rockall — rebleed and in-hospital mortality (derivation)
| Score | Rebleed | In-hospital mortality |
|---|---|---|
| 0 | 0% | 0% |
| 1 | 3.4% | 0% |
| 2 | 5.3% | 0.2% |
| 3 | 11.2% | 2.9% |
| 4 | 14.1% | 5.3% |
| 5 | 24.1% | 10.8% |
| 6 | 32.9% | 17.3% |
| 7 | 43.8% | 27.0% |
| ≥ 8 | 41.8% | 41.1% |
Derivation cohort percentages; confidence intervals widen at the extremes because the top-of-scale groups were small. Later validations reproduced the broad trajectory but recommend pairing with the Glasgow–Blatchford Score at first contact.
References
- Rockall, T. A., Logan, R. F., Devlin, H. B., & Northfield, T. C. (1996). Risk assessment after acute upper gastrointestinal haemorrhage. Gut, 38(3), 316–321. https://doi.org/10.1136/gut.38.3.316
- Vreeburg, E. M., Terwee, C. B., Snel, P., Rauws, E. A., Bartelsman, J. F., Meulen, J. H., & Tytgat, G. N. (1999). Validation of the Rockall risk scoring system in upper gastrointestinal bleeding. Gut, 44(3), 331–335. https://doi.org/10.1136/gut.44.3.331
