Acute upper gastrointestinal bleeding

Definition — bleeding proximal to the ligament of Treitz, split at first sight

Acute upper gastrointestinal bleeding is bleeding proximal to the ligament of Treitz, presenting with haematemesis, melaena, or — the trap — haematochezia when bleeding is brisk. The first viva decision is the split into nonvariceal disease (peptic ulcer, erosions, Mallory-Weiss, tumour, Dieulafoy) and variceal disease in portal hypertension, because resuscitation overlaps but endoscopic therapy, drugs, and rescue pathways diverge completely.^[4][46]

Peptic ulcer is the commonest single cause, responsible for about 50% of all cases, followed by oesophagitis and erosive disease; in cirrhotic patients varices cause 50–60% of bleeds.^[46] The 2024 Lancet ulcer review frames the disease for the examiner: annual prevalence 0.12–1.5%, usually attributable to Helicobacter pylori, aspirin/NSAIDs, critical illness, or idiopathic disease, with management built on H. pylori eradication, acid suppression — most often proton pump inhibitors — and, for bleeding, early endoscopy with high-dose PPI cover.^[26]

Epidemiology — how common, how lethal, how often it rebleeds

Upper GI bleeding is a common medical emergency with a reported mortality of 2–10%.^[4] The epidemiological surveys add texture: several show a falling incidence of all-cause UGIB even as peptic-ulcer bleeding holds stable; rebleeding occurs in 7–16% despite endoscopic therapy and runs highest in variceal and ulcer bleeding; mortality ranges 3–14%, unchanged over a decade, rising with age and strikingly higher in patients already admitted for comorbidity.^[46]

^{[4] [46] [7]}

Ulcer risk factors the examiner expects named: NSAID use and H. pylori infection — found in about half of ulcer bleeders, to be tested for and eradicated in every ulcer patient — with COX-2 agents trading fewer ulcers for more cardiovascular events, and gastroprotection scandalously underused (only 10% of at-risk NSAID users protected).^[46]

Initial assessment and resuscitation — transfuse restrictively

Resuscitation and assessment proceed simultaneously: intravenous fluids to restore perfusion, then red cells at a restrictive threshold. For patients without cardiovascular disease the International Consensus Group suggests transfusion below 80 g/L with a higher threshold for cardiac disease; the BMJ review gives 70–80 g/L after resuscitation; the ACG suggests 7 g/dL for hospitalised bleeders.^[1][4][3]

The numbers behind the rule come from Villanueva's NEJM trial: 921 severe bleeders randomised to a restrictive threshold of 7 g/dL versus a liberal 9 g/dL — 6-week survival 95% versus 91% (hazard ratio for death 0.55), further bleeding 10% versus 16%, adverse events 40% versus 48%. Benefit held in peptic-ulcer bleeding and in Child-Pugh A/B cirrhosis but not Child-Pugh C, and the portal-pressure gradient rose significantly under the liberal strategy.^[20]

TRIGGER tested the same question pragmatically across six UK hospitals (restrictive 80 g/L versus liberal 100 g/L): the cluster design proved feasible with 96% versus 83% protocol adherence and fewer patients transfused (33% versus 46%), though the feasibility sample showed no significant outcome difference.^[21] The pooled verdict then favoured restriction — lower all-cause mortality (relative risk 0.65) and less rebleeding (relative risk 0.58) with no ischaemic excess across five randomised trials and 1,965 participants.^[22] The 2025 Cochrane update (61 trials, 27,639 participants) confirms the GI-bleeding exception: restrictive strategy lowers 30-day mortality in gastrointestinal bleeding (relative risk 0.63) while other contexts show no mortality difference.^[23]

Risk scores — Glasgow-Blatchford discharges, Rockall stages, AIMS65 warns

Three guidelines converge on the discharge rule: a Glasgow-Blatchford score of 0–1 identifies very-low-risk patients who may be discharged with outpatient follow-up — the Consensus Group, ESGE (≤1: very low risk of rebleeding, 30-day mortality or hospital intervention, manageable as outpatients), and the ACG (0–1 in the emergency department).^[1][2][3]

The Stanley international study (3,012 consecutive bleeders, six hospitals across four continents) is the comparison to quote: Glasgow-Blatchford predicted intervention-or-death best (AUROC 0.86) against full Rockall (0.70), PNED (0.69), admission Rockall (0.66) and AIMS65 (0.68); a score of ≤1 gave 98.6% sensitivity for survival without intervention; ≥7 best predicted endoscopic treatment. PNED and AIMS65 predicted mortality best — but no score usefully predicted rebleeding or length of stay.^[7]

Rockall's original construction still matters because examiners ask what goes into it: age, shock, comorbidity, diagnosis, major stigmata of recent haemorrhage and rebleeding independently predict mortality, while haemoglobin, sex, presentation other than shock, NSAIDs and anticoagulants dropped out. The score identified 15% of cases at presentation (26% after endoscopy) at negligible risk of death — the original early-discharge argument.^[8] The Dutch prospective validation found Glasgow-Blatchford (AUC 0.88) superior to both Rockall scores for predicting need for treatment, with ≤2 as the optimal low-risk cut in that cohort.^[10]

AIMS65 earns its place as the mortality score: superior to Glasgow-Blatchford and pre-endoscopy Rockall for inpatient mortality (AUROC 0.80 versus 0.76 versus 0.74) and best for ICU admission and length of stay, while Glasgow-Blatchford remains superior for transfusion prediction.^[9] The 2025 nonvariceal meta-analysis (seven studies, 755 participants) refines the comparison: Glasgow-Blatchford beats Rockall for predicting transfusion need and surgical intervention, but both perform poorly for mortality and rebleeding.^[11] The viva line: Blatchford decides disposition, Rockall stages mortality after endoscopy, AIMS65 warns of death — and none replaces judgment.

Pre-endoscopy drugs — PPI considered, erythromycin given, vasoactive started

The pre-endoscopy bundle after resuscitation: consider high-dose intravenous PPI without delaying endoscopy (ESGE 2026 suggestion), give a prokinetic so the endoscopist can see, and — in suspected variceal bleeding — start vasoactive drugs and antibiotics immediately.^[6][3][27]

Erythromycin is the prokinetic with guideline weight: the ACG suggests infusion before endoscopy, and ESGE's variceal guideline specifies 250 mg intravenously 30–120 minutes pre-endoscopy; where erythromycin is unavailable, ESGE 2026 suggests intravenous metoclopramide in severe or ongoing bleeding.^[3][27][6] The BMJ review's bundle adds antibiotics and vasoactive drugs for cirrhotics alongside PPI and erythromycin for all comers.^[4]

For suspected variceal bleeding the vasoactive doctrine is absolute: terlipressin, octreotide or somatostatin at presentation, continued for up to 5 days; AGA best practice says start before endoscopy, continue 2–5 days after hemostasis, and names octreotide the drug of choice on safety.^[27][31] The efficacy nuance the examiner probes: terlipressin versus placebo cuts all-cause mortality (relative risk 0.66, Cochrane), making it the only vasoactive with a mortality signal — yet terlipressin/vasopressin versus octreotide/somatostatin shows identical mortality, bleeding control and rebleeding with significantly more adverse events on terlipressin (relative risk 2.39).^[33][32] The 2012 review's bedside version: start vasoactive as soon as variceal bleeding is suspected, maintain up to 5 days, prefer terlipressin where available for its survival data.^[34]

Antibiotics are not optional in variceal bleeding: ceftriaxone 1 g daily for up to 7 days (or per local resistance) for every patient with advanced chronic liver disease presenting with variceal hemorrhage.^[27]

Endoscopy timing — within 24 hours, not within 6

Every major guideline lands in the same place: after haemodynamic resuscitation, upper endoscopy within 24 hours — the Consensus Group suggests it, ESGE 2021 strongly recommends early (≤24-hour) endoscopy on high-quality evidence, the ACG suggests it, the BMJ states it with earlier scope considered after resuscitation in high-risk instability.^[1][2][3][4] ESGE 2021 goes further and recommends against urgent (≤12-hour) endoscopy since outcomes are not improved over early endoscopy; ESGE 2026 recommends against emergent (≤6-hour) or urgent (≤12-hour) scope unless the patient remains unstable despite adequate resuscitation.^[2][6]

Lau's NEJM trial is the evidence to quote precisely: 516 high-risk bleeders (Glasgow-Blatchford ≥12) randomised to endoscopy within 6 hours versus 6–24 hours after consultation — 30-day mortality 8.9% versus 6.6% (difference 2.3 points, confidence interval −2.3 to 6.9), further bleeding 10.9% versus 7.8%. Urgent scope found more active bleeding and visible vessels (66.4% versus 47.8%) and treated more lesions (60.1% versus 48.4%) — yet saved no lives.^[47]

The confirmatory layers close the argument: an RCT-only meta-analysis (five trials, 926 patients) found no rebleeding, mortality, ICU, transfusion, surgery or stay benefit for very-early endoscopy, only a higher need for haemostatic treatment (relative risk 1.23).^[48] The Hong Kong territory-wide cohort (6,474 patients receiving therapeutic endoscopy within 48 hours) found early 6–24-hour endoscopy superior to both urgent ≤6-hour and late 24–48-hour timing for 30-day mortality, repeat endoscopy and ICU admission — especially in nonvariceal bleeding.^[49] The exam sentence: resuscitate and medicate first, scope within 24 hours, and reserve urgent scope for refractory instability.

For variceal bleeding the window is tighter: endoscopic evaluation within 12 hours once resuscitated.^[27]

Endoscopic hemostasis — dual therapy by Forrest class

The Forrest-anchored prescription never changes. For actively bleeding ulcers (FIa spurting, FIb oozing): epinephrine injection plus a second modality — contact thermal or mechanical therapy. For a nonbleeding visible vessel (FIIa): thermal, mechanical or sclerosant injection, each alone or combined with epinephrine. Thermocoagulation and sclerosant injection are recommended; clips are suggested; haemostatic powder is temporizing, not sole therapy, for actively bleeding ulcers.^[1][2] The ACG's modality list concurs — bipolar electrocoagulation, heater probe and absolute ethanol injection recommended, with clips, argon plasma coagulation and soft monopolar coagulation supported on lower-quality evidence.^[3]

The meta-analytic verdict (28 trials, 2,988 patients): injection alone is inferior to injection-plus-clip and injection-plus-thermal for rebleeding and emergency surgery; clip beats injection for rebleeding; thermal-plus-injection beats thermal alone. The recommendation: clips, or combined injection-plus-thermal — never injection alone.^[15]

ESGE 2026 adds the modern refinements: adherent clot (FIIb) should undergo clot removal with subsequent hemostasis where the endoscopist has the competence to handle conversion to a higher-risk lesion; haemostatic forceps with soft coagulation may serve as monotherapy for FIa/FIb/FIIa; haemostatic agents must not be first-line monotherapy for high-risk stigmata; and no consensus exists for routine Doppler-probe guidance.^[6] The techniques review reminds the candidate that Doppler can objectify the treatment endpoint and AI risk models are coming — but standard dual therapy remains irreplaceable.^[19]

Over-the-scope clips and haemostatic powder — rescue with defined seats

Recurrent ulcer bleeding after initial hemostasis is where over-the-scope clips earned their place. The STING trial (66 patients across nine referral centres): further bleeding 15.2% with OTSC versus 57.6% with standard therapy (through-the-scope clips or thermal-plus-adrenaline) — an absolute difference of 42.4% — with persistent bleeding 6.0% versus 42.4% and no mortality or salvage-therapy difference.^[12] Guidelines now place the cap-mounted clip as the endoscopic answer to clinical rebleeding, with TAE next and surgery after that.^[2][6]

First-line OTSC has since been tested directly. The TOP trial randomised Forrest Ia–IIb ulcers to OTS versus through-the-scope clips: 30-day rebleeding was low in both arms (1.6% versus 3.9%, not significant), but initial hemostasis (98.4% versus 78.4%) and overall clinical success (96.7% versus 74.5%) both favoured OTS clips.^[13] ESGE 2026 consequently suggests OTS clips as first-line monotherapy alternative to combination therapy for FIa/FIb disease given the lower further-bleeding risk — and for FIIa disease as an alternative monotherapy.^[6]

Haemostatic powder's seat is narrower but real. The 2025 Gut trial (341 high-risk NVUGIB patients with initial hemostasis, mostly ulcer disease, two-thirds Forrest I): Nexpowder versus nothing — 72-hour rebleeding 2.9% versus 11.3%, 30-day cumulative rebleeding 7.0% versus 18.8%, with no powder-related adverse events.^[14] Against that result stands the standing doctrine across three guidelines: powder is rescue and temporizing — massive bleeding with poor visualisation, salvage therapy, diffuse malignant bleeding — preferentially rescue rather than primary hemostasis except in malignant or massive bleeding where thermal therapy or clipping is impossible.^[5][1][6]

Acid suppression after hemostasis — the 72-hour course and its challengers

After successful endoscopic therapy of high-risk-stigmata ulcers, prescribe the course every guideline repeats: high-dose PPI — intravenous bolus (e.g. 80 mg) then continuous infusion (e.g. 8 mg/hour) for 72 hours — then oral PPI twice daily through 14 days, then daily for a lesion-dependent total duration. The Consensus Group, ESGE (which accepts twice-daily IV-bolus or oral twice-daily as alternatives), and the ACG (continuous or intermittent for 3 days, then twice-daily oral for 2 weeks) all agree on the shape.^[1][2][3]

Then deliver the viva flourish — the dose controversy. Intermittent PPI is noninferior to bolus-plus-infusion for 7-day rebleeding (risk ratio 0.72, absolute difference −2.64%, well within the 3% noninferiority margin), with every secondary outcome favouring intermittent numerically.^[24] High-dose versus non-high-dose PPI shows no difference in rebleeding, surgical intervention or mortality across seven randomised trials and 1,157 patients, regardless of stigmata severity, route or dose.^[25] Yet every major guideline still prescribes the high-dose 72-hour course — so the exam answer is the guideline regimen, with the controversy as commentary, not substitution.

Vonoprazan offers a genuine alternative: 20 mg twice daily for 3 days then daily for 28 days proved noninferior to IV pantoprazole infusion for 30-day rebleeding (7.1% versus 10.4%, within the 10% margin) with comparable 3-day, 7-day, mortality, rescue-therapy, transfusion and stay outcomes.^[18] ESGE 2026 reaches no consensus for or against routine potassium-competitive blockers — so vonoprazan is defensible, not yet directive.^[6]

Second-look endoscopy — scheduled no, on-demand yes

The 2022 high-risk RCT tested scheduled second-look directly: 157 high-risk ulcer patients (9-point score ≥5) randomised to scheduled second-look versus observation — 30-day clinical bleeding 11.8% versus 18.2% (difference 6.4%, confidence interval −5.0 to 17.8, not significant), with no transfusion, stay or mortality difference, though 30.4% of second-looks needed further treatment. The trial stopped at half enrolment.^[16]

The 2012 meta-analysis (938 patients) explains when second-look ever helped: routine second-look cut rebleeding (odds ratio 0.55) and surgery (odds ratio 0.43) but not mortality — except the single high-dose-PPI trial showed no benefit, and removing the two highest-risk trials erased the effect.^[17] The modern answer: repeat endoscopy for recurrent bleeding (ACG), not routine second-look in the high-dose-PPI era.^[3]

Refractory nonvariceal bleeding — embolize first, operate last

The salvage ladder is identical across guidelines: clinical rebleed gets repeat endoscopy (preferably OTSC); if that fails, transcatheter angiographic embolization; surgery only when TAE is unavailable or has failed.^[2][6] The ACG's version is the single sentence to memorise: if endoscopic therapy fails, transcatheter embolization is suggested.^[3]

The TAE-versus-surgery meta-analysis (13 studies, 1,077 patients, all non-randomised) quantifies the trade: mortality trends toward TAE (odds ratio 0.77, not significant) with significant heterogeneity; rebleeding is significantly higher after TAE (odds ratio 2.44); complications are sharply lower after TAE (odds ratio 0.45); further intervention favours surgery (odds ratio 2.13 for TAE needing more). TAE stands as viable first-line salvage — higher rebleed accepted for lower mortality and morbidity — with the authors calling for randomised trials.^[39] Ripoll's direct comparison supports it: 70 refractory ulcer bleeders, embolotherapy versus surgery — no difference in recurrent bleeding (29% versus 23.1%), rescue surgery or death — despite embolotherapy patients being older with more cardiac disease and anticoagulation.^[40]

Empiric embolization extends the indication: when angiography is negative but endoscopy or imaging localised the source, empiric embolization achieves 74.7% clinical success with 80.9% survival, statistically indistinguishable from targeted embolization for rebleeding, mortality and rescue surgery, with ~2% embolization-specific complications.^[38] On prophylactic embolization after successful hemostasis the guidelines split by recency: AGA discourages it, while ESGE 2026 suggests considering it in selected high-risk ulcers — haemodynamic instability at presentation, posterior duodenal location, size above 2 cm, or uncertain durable hemostasis.^[5][6]

When the surgeon operates — EAST indications and the oversew

Peptic ulcer disease is now medically managed in the majority; surgery is strictly reserved for life-threatening complications — free perforation, refractory bleeding and gastric outlet obstruction — for which the acute-care surgeon must stay facile in bleeding control, stricture bypass, vagotomy, resection and reconstruction.^[41]

The bleeding operation the examiner will demand described stepwise: duodenotomy with oversewing of the bleeding duodenal ulcer and ligation of the gastroduodenal artery — infrequent in the endoscopy/IR era, yet lifesaving when embolization fails or is unavailable. Graham-patch repair (open or laparoscopic by surgeon comfort) covers perforation; giant ulcers pose the exclusion-versus-resection dilemma; vagotomy with antrectomy and Roux-en-Y reconstruction is the least-common complex option.^[41]

The sequencing that keeps the surgeon relevant: failed dual therapy, failed OTSC or powder rescue, failed or unavailable TAE — then theatre, on a resuscitated, restrictively transfused, PPI-loaded patient. Never apologise for operating on the posterior duodenal ulcer above 2 cm with uncertain hemostasis; ESGE 2026 names exactly that lesion for prophylactic-TAE consideration, which is the same lesion that bleeds through embolization onto your table.^[6][39]

Variceal bleeding — bundle, band, glue, then TIPS

The variceal bundle, recited end-to-end: restrictive transfusion (≤70 g/L, target 70–90), Child-Pugh/MELD plus active-versus-inactive bleeding stratification, vasoactive at presentation for up to 5 days, ceftriaxone 1 g daily up to 7 days, erythromycin 250 mg 30–120 minutes pre-endoscopy, and endoscopy within 12 hours once resuscitated.^[27] The older review's identical frame — fluids, 7–8 g/dL transfusion target, antibiotics, immediate vasoactive — credits bundled care with cutting bleeding-episode mortality from 40% to 15–20%.^[34] APASL's 2025 update adds the regional overlay: 6-week mortality 10–20%, with emphasis on safe transport, defined rebleeding timeframes, early-intervention optimisation, salvage TIPS and metal stents, and special populations.^[28]

Endoscopic therapy is lesion-specific: band ligation for acute oesophageal variceal hemorrhage; cyanoacrylate injection for cardiofundal gastric varices (GOV2/IGV1); glue or banding for GOV1-specific bleeding; urgent rescue TIPS or BRTO when endoscopic hemostasis fails or bleeding recurs early. Survivors enter secondary prophylaxis: eradication banding at 1–4-weekly intervals plus nonselective beta-blockade (propranolol or carvedilol).^[27] Primary prophylaxis for completeness: compensated advanced chronic liver disease with clinically significant portal hypertension gets carvedilol-preferred beta-blockade; high-risk varices in beta-blocker-intolerant patients get ligation every 2–4 weeks to eradication, then 3–6-monthly surveillance.^[27]

Gastric varices deserve their own viva minute because the evidence is weakest where the stakes are highest: no definitive natural-history studies, only biased case series, restricted cohorts and small randomised trials — yet the AGA update still frames glue obturation as the endoscopic core with TIPS, BRTO and shunt surgery as backstops.^[29] The JAMA surgical review maps the East-West split: Western radiologists prefer TIPS decompression, Eastern radiologists prefer direct BRTO targeting — and BRTO, with falling procedure risk over 20 years, is gaining Western acceptance as a safe, effective gastric-variceal option demanding training, consensus and multidisciplinary selection.^[30]

TIPS — pre-emptive selection, rescue hierarchy, survival benefit

Pre-emptive TIPS is the highest-yield variceal fact in the topic. Garcia-Pagan's NEJM trial randomised high-risk bleeders (Child-Pugh C, or Child-Pugh B with persistent bleeding at endoscopy) within 24 hours of admission to covered-stent TIPS within 72 hours versus vasoactive-plus-banding: treatment failure or rebleeding in 1 versus 14 patients, 1-year freedom from failure 97% versus 50%, deaths 4 versus 12, 1-year survival 86% versus 61%, with less ICU and hospital time and no serious-adverse-event excess.^[35]

The Chinese RCT in Child-Pugh B/C bleeders confirms it: early TIPS within 72 hours versus standard care — transplantation-free survival hazard ratio 0.50, 6-week survival 99% versus 84%, 1-year 86% versus 73%, with no excess of encephalopathy, hydrothorax, peritonitis, hepatorenal syndrome or carcinoma.^[37] The pooled meta-analysis (three RCTs plus six observational studies, 2,878 participants): early TIPS cuts all-cause mortality (relative risk 0.64), failure to control bleeding (relative risk 0.15) and rebleeding (relative risk 0.40) without increasing encephalopathy — benefit confined to Child-Pugh B with active bleeding and Child-Pugh C, absent in low-risk patients.^[36]

The selection rule to recite: consider pre-emptive TIPS within 72 hours (preferably 24) for Child-Pugh C ≤13 or Child-Pugh B above 7 with active bleeding at endoscopy despite vasoactive drugs (or HVPG above 20 mmHg); urgent rescue TIPS for persistent bleeding despite drugs plus endoscopy; balloon tamponade only as a bridge to TIPS; and in mild early rebleeding a second endoscopic attempt is permitted before TIPS.^[27][34]

Antithrombotics — resume early, accept the rebleed

Aspirin for secondary cardiovascular prophylaxis is not interrupted; if stopped for any reason, restart as soon as possible, preferably within 3–5 days. Anticoagulation after ulcer hemorrhage resumes once bleeding is controlled, preferably within or soon after 7 days, weighted by thromboembolic risk and remembering the rapid onset of direct oral anticoagulants versus vitamin K antagonists.^[2] ESGE 2026 restates it as resumption as soon as clinically indicated by thromboembolic risk.^[6] The BMJ review's bedside version: in patients who need antithrombotics, outcomes look better when the drugs go back early.^[4]

The 2026 meta-analysis (21 studies) quantifies the trade the examiner wants weighed aloud: antiplatelet resumption cuts major vascular events (hazard ratio 0.70) and mortality (hazard ratio 0.44) while raising rebleeding (hazard ratio 1.42); anticoagulant resumption cuts vascular events (hazard ratio 0.45) and mortality (hazard ratio 0.53) while raising rebleeding (hazard ratio 1.59). Early ≤7-day resumption of either cuts vascular events (hazard ratio 0.39) with more rebleeding but no mortality difference — a favourable benefit-risk balance.^[43] The 871-patient cohort (mean age 78.9, median follow-up 24.9 months) concurs: resumption meant more rebleeding but fewer ischaemic events and deaths, with identical patterns for upper and lower bleeding — the benefits of early reinstitution outweigh the gastrointestinal risks.^[42]

A final pair of post-hemostasis orders from ESGE 2026: start iron therapy before discharge in deficiency or anemia, and start oral nutrition within 24 hours of durable hemostasis.^[6]

Tranexamic acid — the answer is no

Tranexamic acid is the examiner's trap: a plausible drug with a seductive early signal that the guidelines never adopted. The 2014 Cochrane review (eight trials, 1973–2011): mortality 42/851 versus 71/850 (relative risk 0.60) — but the analysis collapsed when missing outcomes were counted as failures or when restricted to low-attrition trials; rebleeding trended down (relative risk 0.80) without significance; surgery and transfusion signals were inconsistent; and the authors deferred everything to the pending 8,000-patient HALT-IT trial.^[44]

The 2025 meta-analysis (23 studies, over two million participants including HALT-IT-era data) claims reduced rebleeding overall, mortality reduction with combined oral-plus-intravenous routes and in UGIB specifically, and reduced surgical need in low-bias studies — with no definitive thrombotic signal.^[45] Yet no major UGIB guideline recommends tranexamic acid for upper bleeding. The exam answer stays not recommended — quote both reviews, then decline the drug.

Revision summary

• Ulcer causes ~50% of UGIB; mortality 2–10%, rebleed 7–16%; NSAIDs and H. pylori are the ulcer risks.^[46][4][26]
• Transfuse restrictively: 7 g/dL (Villanueva survival 95% vs 91%, HR 0.55); variceal ≤70 g/L targeting 70–90.^{[20][27][22][23]}
• GBS 0–1 discharges (Stanley sens 98.6%); Rockall stages mortality; AIMS65 warns of death.^[1][7][8][9]
• Pre-endoscopy: PPI considered without delay, erythromycin 250 mg, vasoactive plus ceftriaxone 1 g/day if variceal.^[6][27][31]
• Scope within 24 h; urgent ≤6 h saves nothing (Lau mortality 8.9% vs 6.6%) — resuscitate first.^{[47][2][49][48]}
• Dual therapy for FIa/FIb/FIIa; never adrenaline alone; FIIb clot removal if competent.^[2][15][6]
• OTSC for rebleed (STING further bleeding 15.2% vs 57.6%) and as first-line alternative; powder as rescue (Nexpowder 72-h 2.9% vs 11.3%).^{[12][13][14][5]}
• High-dose PPI 80 mg + 8 mg/h for 72 h, then twice-daily oral 14 days; vonoprazan noninferior (30-day 7.1% vs 10.4%).^{[1][18][24][25]}
• No routine second-look in the high-dose-PPI era; re-scope for rebleed.^[16][17][3]
• Refractory: repeat endoscopy, then TAE, then surgery; empiric embolization works (74.7% success).^[38][39][40]
• Surgeon operates for free perforation, refractory bleeding, outlet obstruction — duodenotomy, ulcer oversew, GDA ligation.^[41]
• Variceal: band oesophageal, glue gastric (GOV2/IGV1); pre-emptive TIPS within 72 h for Child B-active/C disease (survival 86% vs 61%).^{[27][35][37][36]}
• Resume aspirin (3–5 days) and anticoagulation (within ~7 days); early resumption favours survival despite rebleed.^[2][43][42]
• No tranexamic acid for UGIB; iron before discharge; feed within 24 h.^[44][45][6]

References49ShowHide

1. [1]Barkun AN, Almadi M, Kuipers EJ, Laine L, et al. Management of Nonvariceal Upper Gastrointestinal Bleeding: Guideline Recommendations From the International Consensus Group. Ann Intern Med, 2019.PMID 31634917
2. [2]Gralnek IM, Stanley AJ, Morris AJ, Camus M, et al. Endoscopic diagnosis and management of nonvariceal upper gastrointestinal hemorrhage (NVUGIH): European Society of Gastrointestinal Endoscopy (ESGE) Guideline - Update 2021. Endoscopy, 2021.PMID 33567467
3. [3]Laine L, Barkun AN, Saltzman JR, Martel M, et al. ACG Clinical Guideline: Upper Gastrointestinal and Ulcer Bleeding. Am J Gastroenterol, 2021.PMID 33929377
4. [4]Stanley AJ, Laine L Management of acute upper gastrointestinal bleeding. BMJ, 2019.PMID 30910853
5. [5]Mullady DK, Wang AY, Waschke KA AGA Clinical Practice Update on Endoscopic Therapies for Non-Variceal Upper Gastrointestinal Bleeding: Expert Review. Gastroenterology, 2020.PMID 32574620
6. [6]Gralnek IM, Morris J, Laursen SB, Camus M, et al. Endoscopic diagnosis and management of peptic ulcer bleeding: European Society of Gastrointestinal Endoscopy (ESGE) Guideline - Update 2026. Endoscopy, 2026.PMID 42127996
7. [7]Stanley AJ, Laine L, Dalton HR, Ngu JH, et al. Comparison of risk scoring systems for patients presenting with upper gastrointestinal bleeding: international multicentre prospective study. BMJ, 2017.PMID 28053181
8. [8]Rockall TA, Logan RF, Devlin HB, Northfield TC Risk assessment after acute upper gastrointestinal haemorrhage. Gut, 1996.PMID 8675081
9. [9]Robertson M, Majumdar A, Boyapati R, Chung W, et al. Risk stratification in acute upper GI bleeding: comparison of the AIMS65 score with the Glasgow-Blatchford and Rockall scoring systems. Gastrointest Endosc, 2016.PMID 26515955
10. [10]Aquarius M, Smeets FG, Konijn HW, Stassen PM, et al. Prospective multicenter validation of the Glasgow Blatchford bleeding score in the management of patients with upper gastrointestinal hemorrhage presenting at an emergency department. Eur J Gastroenterol Hepatol, 2015.PMID 26049709
11. [11]Kozai L, Tan A, Nebrejas K, Nishimura Y Comparative diagnostic utility of Rockall and Glasgow-Blatchford scores in non-variceal upper gastrointestinal bleeding: a systematic review and meta-analysis. Eur J Gastroenterol Hepatol, 2025.PMID 39400553
12. [12]Schmidt A, Gölder S, Goetz M, Meining A, et al. Over-the-Scope Clips Are More Effective Than Standard Endoscopic Therapy for Patients With Recurrent Bleeding of Peptic Ulcers. Gastroenterology, 2018.PMID 29803838
13. [13]Soriani P, Biancheri P, Bonura GF, Gabbani T, et al. Over-the-scope clip as first-line treatment of peptic ulcer bleeding: a multicenter randomized controlled trial (TOP Study). Endoscopy, 2024.PMID 38599622
14. [14]Shin J, Cha B, Hong J, Kwon KS, et al. Prevention of rebleeding after primary haemostasis using haemostatic powder in non-variceal upper gastrointestinal bleeding: a multicentre randomised controlled trial. Gut, 2025.PMID 40360231
15. [15]Baracat F, Moura E, Bernardo W, Pu LZ, et al. Endoscopic hemostasis for peptic ulcer bleeding: systematic review and meta-analyses of randomized controlled trials. Surg Endosc, 2016.PMID 26487199
16. [16]Pittayanon R, Suen BY, Kongtub N, Tse YK, et al. Scheduled second look endoscopy after endoscopic hemostasis to patients with high risk bleeding peptic ulcers: a Randomized Controlled Trial. Surg Endosc, 2022.PMID 35020056
17. [17]El Ouali S, Barkun AN, Wyse J, Romagnuolo J, et al. Is routine second-look endoscopy effective after endoscopic hemostasis in acute peptic ulcer bleeding? A meta-analysis. Gastrointest Endosc, 2012.PMID 22695209
18. [18]Geeratragool T, Kaosombatwattana U, Boonchote A, Chatthammanat S, et al. Comparison of Vonoprazan Versus Intravenous Proton Pump Inhibitor for Prevention of High-Risk Peptic Ulcers Rebleeding After Successful Endoscopic Hemostasis: A Multicenter Randomized Noninferiority Trial. Gastroenterology, 2024.PMID 38582271
19. [19]Lau LHS, Sung JJY Treatment of upper gastrointestinal bleeding in 2020: New techniques and outcomes. Dig Endosc, 2021.PMID 32216134
20. [20]Villanueva C, Colomo A, Bosch A, Concepción M, et al. Transfusion strategies for acute upper gastrointestinal bleeding. N Engl J Med, 2013.PMID 23281973
21. [21]Jairath V, Kahan BC, Gray A, Doré CJ, et al. Restrictive versus liberal blood transfusion for acute upper gastrointestinal bleeding (TRIGGER): a pragmatic, open-label, cluster randomised feasibility trial. Lancet, 2015.PMID 25956718
22. [22]Odutayo A, Desborough MJ, Trivella M, Stanley AJ, et al. Restrictive versus liberal blood transfusion for gastrointestinal bleeding: a systematic review and meta-analysis of randomised controlled trials. Lancet Gastroenterol Hepatol, 2017.PMID 28397699
23. [23]Carson JL, Stanworth SJ, Dennis JA, Fergusson DA, et al. Transfusion thresholds and other strategies for guiding red blood cell transfusion. Cochrane Database Syst Rev, 2025.PMID 41114449
24. [24]Sachar H, Vaidya K, Laine L Intermittent vs continuous proton pump inhibitor therapy for high-risk bleeding ulcers: a systematic review and meta-analysis. JAMA Intern Med, 2014.PMID 25201154
25. [25]Wang CH, Ma MH, Chou HC, Yen ZS, et al. High-dose vs non-high-dose proton pump inhibitors after endoscopic treatment in patients with bleeding peptic ulcer: a systematic review and meta-analysis of randomized controlled trials. Arch Intern Med, 2010.PMID 20458081
26. [26]Almadi MA, Lu Y, Alali AA, Barkun AN Peptic ulcer disease. Lancet, 2024.PMID 38885678
27. [27]Gralnek IM, Camus Duboc M, Garcia-Pagan JC, Fuccio L, et al. Endoscopic diagnosis and management of esophagogastric variceal hemorrhage: European Society of Gastrointestinal Endoscopy (ESGE) Guideline. Endoscopy, 2022.PMID 36174643
28. [28]Lesmana CRA, Shukla A, Kumar A, Shalimar, et al. Management of acute variceal bleeding: updated APASL guidelines. Hepatol Int, 2025.PMID 40886248
29. [29]Henry Z, Patel K, Patton H, Saad W AGA Clinical Practice Update on Management of Bleeding Gastric Varices: Expert Review. Clin Gastroenterol Hepatol, 2021.PMID 33493693
30. [30]Lee EW, Shahrouki P, Alanis L, Ding P, et al. Management Options for Gastric Variceal Hemorrhage. JAMA Surg, 2019.PMID 30942880
31. [31]Garcia-Tsao G, Abraldes JG, Rich NE, Wong VW AGA Clinical Practice Update on the Use of Vasoactive Drugs and Intravenous Albumin in Cirrhosis: Expert Review. Gastroenterology, 2024.PMID 37978969
32. [32]Huaringa-Marcelo J, Huaman MR, Brañez-Condorena A, Villacorta-Landeo P, et al. Vasoactive Agents for the Management of Acute Variceal Bleeding: A Systematic Review and Meta-analysis. J Gastrointestin Liver Dis, 2021.PMID 33723542
33. [33]Ioannou G, Doust J, Rockey DC Terlipressin for acute esophageal variceal hemorrhage. Cochrane Database Syst Rev, 2003.PMID 12535432
34. [34]García-Pagán JC, Reverter E, Abraldes JG, Bosch J Acute variceal bleeding. Semin Respir Crit Care Med, 2012.PMID 22447260
35. [35]García-Pagán JC, Caca K, Bureau C, Laleman W, et al. Early use of TIPS in patients with cirrhosis and variceal bleeding. N Engl J Med, 2010.PMID 20573925
36. [36]Zhou GP, Jiang YZ, Sun LY, Zhu ZJ Early transjugular intrahepatic portosystemic shunt for acute variceal bleeding: a systematic review and meta-analysis. Eur Radiol, 2021.PMID 33409783
37. [37]Lv Y, Yang Z, Liu L, Li K, et al. Early TIPS with covered stents versus standard treatment for acute variceal bleeding in patients with advanced cirrhosis: a randomised controlled trial. Lancet Gastroenterol Hepatol, 2019.PMID 31153882
38. [38]Yu Q, Funaki B, Navuluri R, Zangan S, et al. Empiric Transcatheter Embolization for Acute Arterial Upper Gastrointestinal Bleeding: A Meta-Analysis. AJR Am J Roentgenol, 2021.PMID 33566631
39. [39]Tarasconi A, Baiocchi GL, Pattonieri V, Perrone G, et al. Transcatheter arterial embolization versus surgery for refractory non-variceal upper gastrointestinal bleeding: a meta-analysis. World J Emerg Surg, 2019.PMID 30733822
40. [40]Ripoll C, Bañares R, Beceiro I, Menchén P, et al. Comparison of transcatheter arterial embolization and surgery for treatment of bleeding peptic ulcer after endoscopic treatment failure. J Vasc Interv Radiol, 2004.PMID 15126653
41. [41]Hudnall A, Bardes JM, Coleman K, Stout C, et al. The surgical management of complicated peptic ulcer disease: An EAST video presentation. J Trauma Acute Care Surg, 2022.PMID 35358158
42. [42]Sostres C, Marcén B, Laredo V, Alfaro E, et al. Risk of rebleeding, vascular events and death after gastrointestinal bleeding in anticoagulant and/or antiplatelet users. Aliment Pharmacol Ther, 2019.PMID 31486121
43. [43]Liu Y, Hu Y, Song S, Goh WWB, et al. Timing of Antiplatelet and Anticoagulant Therapy Resumption Following Gastrointestinal Bleeding: A Systematic Review and Meta-Analysis. Clin Gastroenterol Hepatol, 2026.PMID 42431497
44. [44]Bennett C, Klingenberg SL, Langholz E, Gluud LL Tranexamic acid for upper gastrointestinal bleeding. Cochrane Database Syst Rev, 2014.PMID 25414987
45. [45]Calderon Martinez E, Briceño Silva GD, Sanchez Cruz C, Woldehana NA, et al. Tranexamic acid as treatment for acute gastrointestinal bleeding: A comprehensive systematic review and meta-analysis. Indian J Gastroenterol, 2025.PMID 40029534
46. [46]van Leerdam ME Epidemiology of acute upper gastrointestinal bleeding. Best Pract Res Clin Gastroenterol, 2008.PMID 18346679
47. [47]Lau JYW, Yu Y, Tang RSY, Chan HCH, et al. Timing of Endoscopy for Acute Upper Gastrointestinal Bleeding. N Engl J Med, 2020.PMID 32242355
48. [48]Merola E, Michielan A, de Pretis G Optimal timing of endoscopy for acute upper gastrointestinal bleeding: a systematic review and meta-analysis. Intern Emerg Med, 2021.PMID 33570742
49. [49]Guo CLT, Wong SH, Lau LHS, Lui RNS, et al. Timing of endoscopy for acute upper gastrointestinal bleeding: a territory-wide cohort study. Gut, 2022.PMID 34548338